<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE article PUBLIC "-//NLM//DTD Journal Publishing with OASIS Tables v3.0 20080202//EN" "https://jats.nlm.nih.gov/nlm-dtd/publishing/3.0/journalpub-oasis3.dtd">
<article xmlns:xlink="http://www.w3.org/1999/xlink" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:oasis="http://docs.oasis-open.org/ns/oasis-exchange/table" xml:lang="en" dtd-version="3.0" article-type="research-article">
  <front>
    <journal-meta><journal-id journal-id-type="publisher">ACP</journal-id><journal-title-group>
    <journal-title>Atmospheric Chemistry and Physics</journal-title>
    <abbrev-journal-title abbrev-type="publisher">ACP</abbrev-journal-title><abbrev-journal-title abbrev-type="nlm-ta">Atmos. Chem. Phys.</abbrev-journal-title>
  </journal-title-group><issn pub-type="epub">1680-7324</issn><publisher>
    <publisher-name>Copernicus Publications</publisher-name>
    <publisher-loc>Göttingen, Germany</publisher-loc>
  </publisher></journal-meta>
    <article-meta>
      <article-id pub-id-type="doi">10.5194/acp-26-16261-2026</article-id><title-group><article-title>Constraining the hydrogen soil sink and photochemical source: insights from atmospheric H<sub>2</sub> inversions (2003–2023)</article-title><alt-title>Insights from atmospheric H<sub>2</sub> inversions (2003–2023)</alt-title>
      </title-group>
      <contrib-group>
        <contrib contrib-type="author" corresp="yes" rid="aff1">
          <name><surname>Stroo</surname><given-names>Firmin T.</given-names></name>
          <email>f.t.stroo@rug.nl</email>
        <ext-link>https://orcid.org/0000-0001-9005-6822</ext-link></contrib>
        <contrib contrib-type="author" corresp="no" rid="aff2">
          <name><surname>Hooghiem</surname><given-names>Joram J. D.</given-names></name>
          
        <ext-link>https://orcid.org/0000-0002-1850-4130</ext-link></contrib>
        <contrib contrib-type="author" corresp="no" rid="aff2 aff3">
          <name><surname>Krol</surname><given-names>Maarten C.</given-names></name>
          
        <ext-link>https://orcid.org/0000-0002-3506-2477</ext-link></contrib>
        <contrib contrib-type="author" corresp="no" rid="aff2">
          <name><surname>Peters</surname><given-names>Wouter</given-names></name>
          
        <ext-link>https://orcid.org/0000-0001-8166-2070</ext-link></contrib>
        <aff id="aff1"><label>1</label><institution>Centre for Isotope Research (CIO), Energy and Sustainability Research Institute Groningen (ESRIG), University of Groningen, Groningen, the Netherlands</institution>
        </aff>
        <aff id="aff2"><label>2</label><institution>Department of Meteorology and Air Quality, Wageningen University &amp; Research, Wageningen, the Netherlands</institution>
        </aff>
        <aff id="aff3"><label>3</label><institution>Institute for Marine and Atmospheric Research Utrecht, Utrecht University, Utrecht, the Netherlands</institution>
        </aff>
      </contrib-group>
      <author-notes><corresp id="corr1">Firmin T. Stroo (f.t.stroo@rug.nl)</corresp></author-notes><pub-date><day>16</day><month>November</month><year>2026</year></pub-date>
      
      <volume>26</volume>
      <issue>22</issue>
      <fpage>16261</fpage><lpage>16294</lpage>
      <history>
        <date date-type="received"><day>28</day><month>May</month><year>2026</year></date>
           <date date-type="rev-request"><day>15</day><month>June</month><year>2026</year></date>
           <date date-type="rev-recd"><day>28</day><month>October</month><year>2026</year></date>
           <date date-type="accepted"><day>3</day><month>November</month><year>2026</year></date>
      </history>
      <permissions>
        <copyright-statement>Copyright: © 2026 Firmin T. Stroo et al.</copyright-statement>
        <copyright-year>2026</copyright-year>
      <license license-type="open-access"><license-p>This work is licensed under the Creative Commons Attribution 4.0 International License. To view a copy of this licence, visit <ext-link ext-link-type="uri" xlink:href="https://creativecommons.org/licenses/by/4.0/">https://creativecommons.org/licenses/by/4.0/</ext-link></license-p></license></permissions><self-uri xlink:href="https://acp.copernicus.org/articles/26/16261/2026/acp-26-16261-2026.html">This article is available from https://acp.copernicus.org/articles/26/16261/2026/acp-26-16261-2026.html</self-uri><self-uri xlink:href="https://acp.copernicus.org/articles/26/16261/2026/acp-26-16261-2026.pdf">The full text article is available as a PDF file from https://acp.copernicus.org/articles/26/16261/2026/acp-26-16261-2026.pdf</self-uri>
      <abstract><title>Abstract</title>

      <p id="d2e142">Hydrogen (<inline-formula><mml:math id="M3" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>) is expected to become an increasingly important energy carrier during the energy transition, likely leading to higher atmospheric <inline-formula><mml:math id="M4" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> levels due to losses during production, transport, storage and usage of hydrogen. Multiple studies have shown this could impact atmospheric composition through interactions with the hydroxyl radical. However, the magnitude of this impact remains uncertain due to large uncertainties in the global <inline-formula><mml:math id="M5" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> budget, particularly in the soil sink and photochemical source. To address this, we present a spatiotemporally resolved <inline-formula><mml:math id="M6" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> budget derived using atmospheric inversions with the TM5 chemical transport model. With this approach, we infer a global mean soil sink of 52.8 [47.8–56.7] <inline-formula><mml:math id="M7" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> and a photochemical source of 34.6 [29.2–38.2] <inline-formula><mml:math id="M8" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> over 2003–2023. Relative to <xref ref-type="bibr" rid="bib1.bibx41" id="text.1"/>, we estimate a soil sink that is 45 % and 35 % weaker in the Middle East and Oceania, and 45 % and 70 % stronger in South America and Russia, respectively. Our results further suggest that variability in the observed <inline-formula><mml:math id="M9" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> growth rate between 2003–2023 was primarily driven by changes in the photochemical source from <inline-formula><mml:math id="M10" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">4</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> oxidation, together with a declining global soil sink at a mean rate of 0.23 <inline-formula><mml:math id="M11" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>. Finally, we infer a sensitivity of the soil sink to the El Niño–Southern Oscillation, strongest over diffusion-limited soils in tropical South America, with increased uptake during drier El Niño conditions and reduced uptake during wetter La Niña conditions.</p>
  </abstract>
    </article-meta>
  </front>
<body>
      

<sec id="Ch1.S1" sec-type="intro">
  <label>1</label><title>Introduction</title>
      <p id="d2e275">Global hydrogen (<inline-formula><mml:math id="M12" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>) demand has roughly doubled over the last two decades and is projected to keep increasing during the ongoing energy transition <xref ref-type="bibr" rid="bib1.bibx26 bib1.bibx27" id="paren.2"/>. As a result, atmospheric <inline-formula><mml:math id="M13" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> levels may increase due to unavoidable losses across production, transport, storage, and usage of hydrogen <xref ref-type="bibr" rid="bib1.bibx40 bib1.bibx14" id="paren.3"/>. This could impact tropospheric methane (<inline-formula><mml:math id="M14" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">4</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>) and ozone (<inline-formula><mml:math id="M15" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">O</mml:mi><mml:mn mathvariant="normal">3</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>), as well as stratospheric water vapour levels, through interactions with the hydroxyl radical (<inline-formula><mml:math id="M16" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">OH</mml:mi></mml:mrow></mml:math></inline-formula>). Together, these effects yield a 100-year global warming potential (GWP<sub>100</sub>) for hydrogen of 11.6 <inline-formula><mml:math id="M18" display="inline"><mml:mo>±</mml:mo></mml:math></inline-formula> 2.8 <xref ref-type="bibr" rid="bib1.bibx57" id="paren.4"/>, meaning that 1 kg of <inline-formula><mml:math id="M19" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> emitted has a climate impact equivalent to 11.6 <inline-formula><mml:math id="M20" display="inline"><mml:mo>±</mml:mo></mml:math></inline-formula> 2.8 kg of <inline-formula><mml:math id="M21" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CO</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> emitted over a 100-year horizon. The uncertainty in the GWP<sub>100</sub> estimate arises primarily from a limited understanding of the global hydrogen budget. Particularly, the photochemical <inline-formula><mml:math id="M23" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source and the soil sink, representing the two largest fluxes in the budget, are highly uncertain (see Table <xref ref-type="table" rid="T2"/>). According to the recent comprehensive assessment of the global <inline-formula><mml:math id="M24" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> budget by <xref ref-type="bibr" rid="bib1.bibx41" id="text.5"/>, the remaining terms in the budget, including biomass burning, fossil fuel combustion, ocean emissions, land emissions from <inline-formula><mml:math id="M25" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">N</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> fixing bacteria, and the <inline-formula><mml:math id="M26" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">OH</mml:mi></mml:mrow></mml:math></inline-formula> sink, are comparatively well constrained in magnitude. This study provides the most detailed uncertainty assessment of the global hydrogen budget to date, based on a diverse ensemble of bottom-up approaches, and will therefore be referred to frequently throughout the remainder of this paper.</p>
      <p id="d2e442">The global photochemical <inline-formula><mml:math id="M27" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source from formaldehyde (<inline-formula><mml:math id="M28" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula>) photolysis is the most uncertain source in the <inline-formula><mml:math id="M29" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> budget. It arises from multiple pathways reflecting the diversity of <inline-formula><mml:math id="M30" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula> precursors and is typically quantified using global chemical transport model simulations. The contribution from <inline-formula><mml:math id="M31" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">4</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> oxidation, the dominant precursor pathway via <inline-formula><mml:math id="M32" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula>, is relatively well constrained. This is because <inline-formula><mml:math id="M33" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">4</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> is a long-lived greenhouse gas with well-characterised global atmospheric abundances, constrained by extensive surface and satellite observations. In contrast, the contribution from oxidation of anthropogenic and biogenic Volatile Organic Compounds (VOCs) is more uncertain, reflecting shorter atmospheric lifetimes and less well-constrained emission inventories of these precursors.</p>
      <p id="d2e529">The global <inline-formula><mml:math id="M34" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> soil sink is the largest and most uncertain term in the <inline-formula><mml:math id="M35" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> budget <xref ref-type="bibr" rid="bib1.bibx58" id="paren.6"/>, as shown in Table <xref ref-type="table" rid="T2"/>. Its uncertainty arises from a complex dependence on multiple interacting environmental drivers, including soil moisture, soil temperature, and soil organic carbon content. It also exhibits strong spatial variability, with field measurements showing that fluxes can differ substantially between sites only a few kilometres apart <xref ref-type="bibr" rid="bib1.bibx13" id="paren.7"/>. Estimates of the global <inline-formula><mml:math id="M36" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> soil sink rely on parametrisations predominantly derived from field measurements at mid-latitude Northern Hemisphere sites <xref ref-type="bibr" rid="bib1.bibx15" id="paren.8"/>. This introduces a representativeness bias, as tropical regions, which are expected to show high soil <inline-formula><mml:math id="M37" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> uptake <xref ref-type="bibr" rid="bib1.bibx66" id="paren.9"/>, are not represented in the underlying observations. In addition, parametrisations are limited by the lack of accurate, high-resolution forcing data for key drivers such as soil moisture and soil temperature, further introducing uncertainty in global flux estimates. Finally, a key limitation of most bottom-up studies is that their fluxes are not propagated through three-dimensional transport models, making it impossible to verify their consistency with long-term global <inline-formula><mml:math id="M38" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> observations.</p>
      <p id="d2e602">To overcome some of these challenges, a top-down approach can be used. Here, bottom-up fluxes serve as a first guess and are subsequently adjusted using <inline-formula><mml:math id="M39" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> mole fraction measurements. These atmospheric observations integrate information from sources and sinks over large spatial scales, thereby reducing the influence of local variability and allowing fluxes to be constrained without detailed knowledge of all underlying environmental drivers. Previous top-down studies with two-dimensional box models by <xref ref-type="bibr" rid="bib1.bibx54" id="text.10"/> and <xref ref-type="bibr" rid="bib1.bibx80" id="text.11"/> suggested that the soil sink and photochemical source were substantially larger than found in bottom-up studies (see Table <xref ref-type="table" rid="T2"/>). Subsequent studies employing three-dimensional chemical transport models yielded budgets more consistent with bottom-up literature, but relied on shorter observational records. For example, <xref ref-type="bibr" rid="bib1.bibx48" id="text.12"/> used observations from a single year (2004), <xref ref-type="bibr" rid="bib1.bibx84" id="text.13"/> relied on relatively few stations, mostly in Europe, and <xref ref-type="bibr" rid="bib1.bibx7" id="text.14"/> analysed observations from 1991–2004. Since then, the observational record has expanded considerably, providing a clear opportunity to revisit this problem. Furthermore, none of these studies included observations from South America or tropical Africa, regions expected to exhibit some of the largest photochemical source and soil sink fluxes globally.</p>
      <p id="d2e635">In this study, we present an updated, comprehensive regional-to-global <inline-formula><mml:math id="M40" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> budget for 2003–2023, derived from atmospheric <inline-formula><mml:math id="M41" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> inversions using the TM5 chemical transport model. We build on the work of <xref ref-type="bibr" rid="bib1.bibx47" id="text.15"/>, incorporating several updates and improvements, including updated emission inventories, a more sophisticated soil sink scheme, improved <inline-formula><mml:math id="M42" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">OH</mml:mi></mml:mrow></mml:math></inline-formula> fields, and higher horizontal and vertical resolution. We also exploit observations from an expanded network of <inline-formula><mml:math id="M43" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> monitoring sites, providing stronger constraints on regional fluxes. Finally, we evaluate our simulated <inline-formula><mml:math id="M44" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> mole fractions against independent observations from four aircraft campaigns.</p>
</sec>
<sec id="Ch1.S2">
  <label>2</label><title>Methods</title>
      <p id="d2e702">The following sections describe the observations (Sect. <xref ref-type="sec" rid="Ch1.S2.SS1"/>), modelling framework (Sect. <xref ref-type="sec" rid="Ch1.S2.SS2"/>), and implementation of hydrogen sources and sinks (Sect. <xref ref-type="sec" rid="Ch1.S2.SS3"/>) used in this study. We then describe the setup of the forward simulations used to pre-balance global mean sources and sinks (Sect. <xref ref-type="sec" rid="Ch1.S2.SS4"/>), followed by the setup of the detailed inversion experiments (Sect. <xref ref-type="sec" rid="Ch1.S2.SS5"/>).</p>

      <fig id="F1" specific-use="star"><label>Figure 1</label><caption><p id="d2e717">Global distribution of <inline-formula><mml:math id="M45" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> monitoring sites (<inline-formula><mml:math id="M46" display="inline"><mml:mrow><mml:mi>N</mml:mi><mml:mo>=</mml:mo><mml:mn mathvariant="normal">96</mml:mn></mml:mrow></mml:math></inline-formula>) from the NOAA-GML, CSIRO-GASLAB, MPI-BGC, AGAGE, ICOS, and EUROHYDROS networks. Only sites with measurements at any time during 2003–2023 are included, with individual sites covering different sub-periods. Key sites are labelled for later reference.</p></caption>
        <graphic xlink:href="https://acp.copernicus.org/articles/26/16261/2026/acp-26-16261-2026-f01.png"/>

      </fig>

<sec id="Ch1.S2.SS1">
  <label>2.1</label><title>Observations</title>
      <p id="d2e756">We use <inline-formula><mml:math id="M47" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> mole fraction observations from multiple surface flask networks, including EUROHYDROS (EUROpean network for atmospheric HYDRogen Observations and Studies; <xref ref-type="bibr" rid="bib1.bibx5" id="altparen.16"/>), NOAA-GML (National Oceanic and Atmospheric Administration Global Monitoring Laboratory; <xref ref-type="bibr" rid="bib1.bibx52" id="altparen.17"/>), CSIRO-GASLAB (Commonwealth Scientific and Industrial Research Organisation Global Atmospheric Sampling Laboratory; <xref ref-type="bibr" rid="bib1.bibx32" id="altparen.18"/>), MPI-BGC (Max Planck Institute for Biogeochemistry; <xref ref-type="bibr" rid="bib1.bibx30" id="altparen.19"/>), and ICOS (Integrated Carbon Observation System; <xref ref-type="bibr" rid="bib1.bibx1" id="altparen.20"/>). We also use in situ measurements from AGAGE (Advanced Global Atmospheric Gases Experiment; <xref ref-type="bibr" rid="bib1.bibx51" id="altparen.21"/>). All data are reported on the MPI-2009 calibration scale <xref ref-type="bibr" rid="bib1.bibx29" id="paren.22"/>. Observations span 96 monitoring sites, with some locations sampled by multiple networks (Fig. <xref ref-type="fig" rid="F1"/>). Most sites do not provide continuous coverage over the full analysis period (2003–2023); details on measurement periods and sampling frequency for each network are provided in Table <xref ref-type="table" rid="TF1"/> of the Appendix. All data are filtered to retain observations representative of background conditions using available quality control and representativity flags. For NOAA-GML, only measurements analysed with a helium pulse discharge detector are used. For independent evaluation of our optimised <inline-formula><mml:math id="M48" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> mole fractions, we use aircraft measurements from NOAA-GML <xref ref-type="bibr" rid="bib1.bibx65 bib1.bibx34" id="paren.23"/>. These measurements include continental aircraft profiles over the United States, as well as the HIPPO (HIAPER Pole to Pole Observations; <xref ref-type="bibr" rid="bib1.bibx79 bib1.bibx77" id="altparen.24"/>) and ATom (Atmospheric Tomography Mission; <xref ref-type="bibr" rid="bib1.bibx35 bib1.bibx67 bib1.bibx78" id="altparen.25"/>) campaigns, which provide pole-to-pole sampling of background atmospheric conditions across multiple altitudes. Additionally, we also use aircraft measurements from IAGOS-CARIBIC (In service Aircraft for a Global Observing System – Civil Aircraft for the Regular Investigation of the Atmosphere Based on an Instrument Container; <xref ref-type="bibr" rid="bib1.bibx4" id="altparen.26"/>) for independent evaluation.</p>
</sec>
<sec id="Ch1.S2.SS2">
  <label>2.2</label><title>Modelling framework</title>
      <p id="d2e829">We use the TM5 three-dimensional chemical transport model <xref ref-type="bibr" rid="bib1.bibx31" id="paren.27"/>, driven by meteorological fields from ERA5 (ECMWF ReAnalysis v5; <xref ref-type="bibr" rid="bib1.bibx24" id="altparen.28"/>), to simulate atmospheric <inline-formula><mml:math id="M49" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> mole fractions. To estimate the magnitude and spatial distribution of monthly global <inline-formula><mml:math id="M50" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> fluxes, TM5 is coupled with the CarbonTracker data assimilation system <xref ref-type="bibr" rid="bib1.bibx71" id="paren.29"/>, which employs an ensemble Kalman filter.</p>
</sec>
<sec id="Ch1.S2.SS3">
  <label>2.3</label><title>Implementation of hydrogen sources and sinks</title>
      <p id="d2e871">Our modelling framework builds on <xref ref-type="bibr" rid="bib1.bibx47" id="text.30"/>, with several updates and improvements. We include <inline-formula><mml:math id="M51" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> surface emissions as monthly fluxes at <inline-formula><mml:math id="M52" display="inline"><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mi mathvariant="italic">°</mml:mi><mml:mo>×</mml:mo><mml:mn mathvariant="normal">1</mml:mn><mml:mi mathvariant="italic">°</mml:mi></mml:mrow></mml:math></inline-formula> resolution. Fossil fuel emissions are calculated by scaling sector-specific <inline-formula><mml:math id="M53" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">CO</mml:mi></mml:mrow></mml:math></inline-formula> emissions from CAMS v6.2 (Copernicus Atmosphere Monitoring Service; <xref ref-type="bibr" rid="bib1.bibx62" id="altparen.31"/>) to <inline-formula><mml:math id="M54" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> using corresponding emission factors from CEDS (Community Emissions Data System; <xref ref-type="bibr" rid="bib1.bibx25" id="altparen.32"/>). This approach exploits the shared origin of <inline-formula><mml:math id="M55" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M56" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">CO</mml:mi></mml:mrow></mml:math></inline-formula> in fossil fuel combustion and related processes such as the water–gas shift reaction, while leveraging the relatively well-constrained global <inline-formula><mml:math id="M57" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">CO</mml:mi></mml:mrow></mml:math></inline-formula> emissions inferred from satellite observations. Oceanic <inline-formula><mml:math id="M58" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> emissions are prescribed from an annually repeating monthly climatology based on the oceanic <inline-formula><mml:math id="M59" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">CO</mml:mi></mml:mrow></mml:math></inline-formula> flux distribution of <xref ref-type="bibr" rid="bib1.bibx12" id="text.33"/>, scaled to a global annual emission of 5 <inline-formula><mml:math id="M60" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> <xref ref-type="bibr" rid="bib1.bibx50" id="paren.34"/>. This approach is based on the shared biological origin of <inline-formula><mml:math id="M61" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> <xref ref-type="bibr" rid="bib1.bibx74" id="paren.35"/> and <inline-formula><mml:math id="M62" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">CO</mml:mi></mml:mrow></mml:math></inline-formula> <xref ref-type="bibr" rid="bib1.bibx19" id="paren.36"/> in the ocean, and uses <inline-formula><mml:math id="M63" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">CO</mml:mi></mml:mrow></mml:math></inline-formula> fluxes as a proxy for biological activity. Similarly, land <inline-formula><mml:math id="M64" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> emissions are prescribed from an annually repeating monthly climatology based on biogenic <inline-formula><mml:math id="M65" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">CO</mml:mi></mml:mrow></mml:math></inline-formula> land fluxes from MEGAN (Model of Emissions of Gases and Aerosols from Nature; <xref ref-type="bibr" rid="bib1.bibx20" id="altparen.37"/>) for 2010, scaled to a global annual emission of 3 <inline-formula><mml:math id="M66" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> <xref ref-type="bibr" rid="bib1.bibx11 bib1.bibx10 bib1.bibx46" id="paren.38"/>. Finally, biomass burning emissions are taken from GFED5 (Global Fire Emissions Database v5; <xref ref-type="bibr" rid="bib1.bibx72" id="altparen.39"/>).</p>
      <p id="d2e1080">We pre-calculated three-dimensional photochemical source fields from <inline-formula><mml:math id="M67" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula> photolysis to use in our system, by running a full-chemistry version of TM5-MP <xref ref-type="bibr" rid="bib1.bibx76" id="paren.40"/> with the CB05 chemical mechanism (Carbon Bond v5; <xref ref-type="bibr" rid="bib1.bibx82" id="altparen.41"/>). The full-chemistry model included 51 species and 156 reactions, covering the oxidation chains of <inline-formula><mml:math id="M68" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">4</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> and other key VOCs. Since <inline-formula><mml:math id="M69" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> was originally treated as a fixed species in CB05, we added <inline-formula><mml:math id="M70" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> production from <inline-formula><mml:math id="M71" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula> photolysis (Reaction <xref ref-type="disp-formula" rid="Ch1.R1"/>) to the chemical scheme using the photolysis rate constant from <xref ref-type="bibr" rid="bib1.bibx59" id="text.42"/>. The remaining <inline-formula><mml:math id="M72" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula> loss pathways were already included in CB05 (Reactions <xref ref-type="disp-formula" rid="Ch1.R2"/> and <xref ref-type="disp-formula" rid="Ch1.R3"/>). 

                <disp-formula specific-use="gather" content-type="numbered reaction"><mml:math id="M73" display="block"><mml:mtable displaystyle="true"><mml:mlabeledtr id="Ch1.R1"><mml:mtd><mml:mtext>R1</mml:mtext></mml:mtd><mml:mtd><mml:mrow><mml:mstyle class="stylechange" displaystyle="true"/><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow><mml:mo>+</mml:mo><mml:mi>h</mml:mi><mml:mi mathvariant="italic">ν</mml:mi><mml:mo>→</mml:mo><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow><mml:mo>+</mml:mo><mml:mrow class="chem"><mml:mi mathvariant="normal">CO</mml:mi></mml:mrow></mml:mrow></mml:mtd></mml:mlabeledtr><mml:mlabeledtr id="Ch1.R2"><mml:mtd><mml:mtext>R2</mml:mtext></mml:mtd><mml:mtd><mml:mrow><mml:mstyle class="stylechange" displaystyle="true"/><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow><mml:mo>+</mml:mo><mml:mi>h</mml:mi><mml:mi mathvariant="italic">ν</mml:mi><mml:mo>→</mml:mo><mml:mrow class="chem"><mml:mi mathvariant="normal">H</mml:mi></mml:mrow><mml:mo>+</mml:mo><mml:mrow class="chem"><mml:mi mathvariant="normal">HCO</mml:mi></mml:mrow></mml:mrow></mml:mtd></mml:mlabeledtr><mml:mlabeledtr id="Ch1.R3"><mml:mtd><mml:mtext>R3</mml:mtext></mml:mtd><mml:mtd><mml:mrow><mml:mstyle class="stylechange" displaystyle="true"/><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow><mml:mo>+</mml:mo><mml:mrow class="chem"><mml:mi mathvariant="normal">OH</mml:mi></mml:mrow><mml:mo>→</mml:mo><mml:mrow class="chem"><mml:mi mathvariant="normal">HCO</mml:mi></mml:mrow><mml:mo>+</mml:mo><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:mrow></mml:mtd></mml:mlabeledtr></mml:mtable></mml:math></disp-formula>

          We performed the full-chemistry simulations at <inline-formula><mml:math id="M74" display="inline"><mml:mrow><mml:mn mathvariant="normal">3</mml:mn><mml:mi mathvariant="italic">°</mml:mi><mml:mo>×</mml:mo><mml:mn mathvariant="normal">2</mml:mn><mml:mi mathvariant="italic">°</mml:mi></mml:mrow></mml:math></inline-formula> horizontal resolution with 34 vertical levels over 2003–2023. Chemistry was driven by the <inline-formula><mml:math id="M75" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> fluxes described in this section, together with <inline-formula><mml:math id="M76" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">4</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> fields nudged towards surface observations and CMIP6 (Coupled Model Intercomparison Project Phase 6; <xref ref-type="bibr" rid="bib1.bibx17" id="altparen.43"/>) emissions for almost all remaining species. We derived separate photochemical <inline-formula><mml:math id="M77" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source fields for production from isoprene oxidation (hereafter the isoprene source) and from oxidation of all remaining VOCs, dominated by <inline-formula><mml:math id="M78" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">4</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> oxidation (hereafter the non-isoprene source). This allows for separate specification of the flux uncertainties, with higher confidence in the non-isoprene <inline-formula><mml:math id="M79" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source due to the well-characterised global abundance of <inline-formula><mml:math id="M80" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">4</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>. We derived the source partitioning from five sensitivity simulations in which isoprene emissions were scaled by a factor 1.0 to 0.6 (in steps of 0.1), yielding an essentially linear response in the photochemical <inline-formula><mml:math id="M81" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source (<inline-formula><mml:math id="M82" display="inline"><mml:mrow><mml:msup><mml:mi>R</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msup><mml:mo>=</mml:mo><mml:mn mathvariant="normal">0.999</mml:mn></mml:mrow></mml:math></inline-formula>). We then estimated the non-isoprene <inline-formula><mml:math id="M83" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source by extrapolating this relationship to zero isoprene emissions, with the isoprene <inline-formula><mml:math id="M84" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source calculated as the residual relative to the base simulation (with a factor of 1.0). We have relatively high confidence in the spatial distribution of our combined photochemical <inline-formula><mml:math id="M85" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source fields, based on a comparison of simulated <inline-formula><mml:math id="M86" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula> mole fractions with <inline-formula><mml:math id="M87" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula> aircraft observations from the ATom campaign, as discussed in Sect. <xref ref-type="sec" rid="App1.Ch1.S1"/> of the Appendix.</p>

      <fig id="F2" specific-use="star"><label>Figure 2</label><caption><p id="d2e1457"><inline-formula><mml:math id="M88" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> annual mean soil sink fluxes (2003–2013) from simulations with the Sanderson <bold>(a)</bold> and Bertagni <bold>(b)</bold> scheme. These simulations were performed using a photochemical source of 38.2 <inline-formula><mml:math id="M89" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> (scaled by a factor of 0.7 for consistency with the literature) and no additional scaling of the soil sink (factor 1.0). Panel <bold>(c)</bold> shows the corresponding latitudinal <inline-formula><mml:math id="M90" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> flux distributions.</p></caption>
          <graphic xlink:href="https://acp.copernicus.org/articles/26/16261/2026/acp-26-16261-2026-f02.png"/>

        </fig>

      <p id="d2e1514">We calculate soil sink fluxes at <inline-formula><mml:math id="M91" display="inline"><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mi mathvariant="italic">°</mml:mi><mml:mo>×</mml:mo><mml:mn mathvariant="normal">1</mml:mn><mml:mi mathvariant="italic">°</mml:mi></mml:mrow></mml:math></inline-formula> resolution following scenario S3c of <xref ref-type="bibr" rid="bib1.bibx47" id="text.44"/>. This approach uses the TM5 dry deposition scheme, which applies a series-resistance formulation including time-varying aerodynamic and soil resistances to compute the <inline-formula><mml:math id="M92" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> deposition velocity. The soil sink flux is then calculated using the deposition velocity and the near-surface <inline-formula><mml:math id="M93" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> mass mixing ratio in the lowest model layer at each 3 h time step of the simulation. In the standard configuration, the soil resistance component follows the <xref ref-type="bibr" rid="bib1.bibx60" id="text.45"/> scheme. This scheme defines seven ecosystem regions representing broad ecosystem categories rather than explicit soil types, and is driven by soil moisture and soil temperature fields from ERA5 <xref ref-type="bibr" rid="bib1.bibx24" id="paren.46"/>. We also implemented an alternative deposition scheme in which the soil and snow resistance formulations are replaced by those of <xref ref-type="bibr" rid="bib1.bibx6" id="text.47"/>. Here, the soil resistance equations include nonlinear soil moisture dynamics and depend on soil moisture, soil temperature, and porosity across eleven soil texture classes. We obtained the required soil property fields from the Global Land Data Assimilation System (GLDAS; <xref ref-type="bibr" rid="bib1.bibx55" id="altparen.48"/>). Figure <xref ref-type="fig" rid="F2"/> highlights the main differences between the two schemes. On average, the Sanderson scheme yields a 40 % larger global soil sink than the Bertagni scheme for 2003–2013. It also yields stronger uptake in subtropical regions such as northern India and South Africa, and weaker uptake in arid regions such as the Sahara and the Tibetan Plateau relative to the Bertagni scheme. Throughout the remainder of this paper, we will refer to simulations using the Sanderson and Bertagni schemes as “Sanderson” and “Bertagni”, respectively (e.g. Sanderson inversion or Bertagni inversion).</p>
      <p id="d2e1573">We calculate the chemical <inline-formula><mml:math id="M94" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> sink by oxidation with <inline-formula><mml:math id="M95" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">OH</mml:mi></mml:mrow></mml:math></inline-formula> based on Reaction (<xref ref-type="disp-formula" rid="Ch1.R4"/>) using the reaction rate constant from JPL (Jet Propulsion Laboratory; <xref ref-type="bibr" rid="bib1.bibx44" id="altparen.49"/>). We use 3 h <inline-formula><mml:math id="M96" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">OH</mml:mi></mml:mrow></mml:math></inline-formula> fields from CAMS EAC4 (Copernicus Atmosphere Monitoring Service ECMWF Atmospheric Composition Reanalysis 4; <xref ref-type="bibr" rid="bib1.bibx28" id="altparen.50"/>) for the period 2003–2023 at <inline-formula><mml:math id="M97" display="inline"><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mi mathvariant="italic">°</mml:mi><mml:mo>×</mml:mo><mml:mn mathvariant="normal">1</mml:mn><mml:mi mathvariant="italic">°</mml:mi></mml:mrow></mml:math></inline-formula> resolution with 60 vertical levels, together with the three-dimensional <inline-formula><mml:math id="M98" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> fields from the TM5 simulation.

            <disp-formula id="Ch1.R4" content-type="numbered reaction"><label>R4</label><mml:math id="M99" display="block"><mml:mrow><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow><mml:mo>+</mml:mo><mml:mrow class="chem"><mml:mi mathvariant="normal">OH</mml:mi></mml:mrow><mml:mo>→</mml:mo><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow><mml:mo>+</mml:mo><mml:mrow class="chem"><mml:mi mathvariant="normal">H</mml:mi></mml:mrow></mml:mrow></mml:math></disp-formula>

          In the stratosphere, we prescribe <inline-formula><mml:math id="M100" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> mole fractions based on the <inline-formula><mml:math id="M101" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">4</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>–<inline-formula><mml:math id="M102" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> relationship of <xref ref-type="bibr" rid="bib1.bibx33" id="text.51"/>, following <xref ref-type="bibr" rid="bib1.bibx47" id="text.52"/>. This is needed since TM5 was mainly designed to simulate the troposphere and lacks advanced stratospheric chemistry and high vertical resolution. For example, important reactions such as methane oxidation by chlorine radicals and electronically excited oxygen atoms, as described by <xref ref-type="bibr" rid="bib1.bibx53" id="text.53"/> and <xref ref-type="bibr" rid="bib1.bibx56" id="text.54"/>, are not included in TM5. To define the troposphere–stratosphere boundary, above which the <inline-formula><mml:math id="M103" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> mole fractions are fixed, we use the latitude-dependent pressure threshold given by Eq. (1) of <xref ref-type="bibr" rid="bib1.bibx46" id="text.55"/>. To drive the parametrisation, we use monthly <inline-formula><mml:math id="M104" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">4</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> fields from CAMS EAC4 at <inline-formula><mml:math id="M105" display="inline"><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mi mathvariant="italic">°</mml:mi><mml:mo>×</mml:mo><mml:mn mathvariant="normal">1</mml:mn><mml:mi mathvariant="italic">°</mml:mi></mml:mrow></mml:math></inline-formula> resolution, interpolated to 34 vertical levels.</p>
</sec>
<sec id="Ch1.S2.SS4">
  <label>2.4</label><title>Balancing global mean sources and sinks</title>
      <p id="d2e1767">To explore the wide range of photochemical source and soil sink estimates in the literature (Table <xref ref-type="table" rid="T2"/>), we use the TM5 model to identify source–sink configurations consistent with the observed evolution of global <inline-formula><mml:math id="M106" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> mole fractions. We perform forward simulations for 2003–2013 at <inline-formula><mml:math id="M107" display="inline"><mml:mrow><mml:mn mathvariant="normal">6</mml:mn><mml:mi mathvariant="italic">°</mml:mi><mml:mo>×</mml:mo><mml:mn mathvariant="normal">4</mml:mn><mml:mi mathvariant="italic">°</mml:mi></mml:mrow></mml:math></inline-formula> resolution with 25 vertical levels, applying uniform global scaling factors to the soil sink and combined photochemical <inline-formula><mml:math id="M108" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source. We then compute four metrics using the NOAA ccgcrv curve fitting package <xref ref-type="bibr" rid="bib1.bibx68 bib1.bibx38" id="paren.56"/> based on the simulated and observed <inline-formula><mml:math id="M109" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> mole fractions at five NOAA-GML sites (ALT, CGO, MHD, MLO, SPO): (a) detrended RMSE, (b) difference between simulated and observed mean growth rate at MLO, (c) absolute difference between simulated and observed interhemispheric gradient between ALT and SPO, and (d) absolute difference between simulated and observed seasonal amplitude at MHD. The configurations that perform best across these metrics are then used as the starting point for the subsequent Bayesian optimisation of <inline-formula><mml:math id="M110" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> sources and sinks.</p>
</sec>
<sec id="Ch1.S2.SS5">
  <label>2.5</label><title>Setup of the inversion experiments</title>
      <p id="d2e1844">We use the CarbonTracker Europe v3 framework <xref ref-type="bibr" rid="bib1.bibx71" id="paren.57"/> to perform a month-by-month optimisation of the soil sink and photochemical <inline-formula><mml:math id="M111" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source. We run the inversion over the 2003–2023 period, with atmospheric transport simulated at <inline-formula><mml:math id="M112" display="inline"><mml:mrow><mml:mn mathvariant="normal">3</mml:mn><mml:mi mathvariant="italic">°</mml:mi><mml:mo>×</mml:mo><mml:mn mathvariant="normal">2</mml:mn><mml:mi mathvariant="italic">°</mml:mi></mml:mrow></mml:math></inline-formula> horizontal resolution and 34 vertical levels. We use an ensemble size of 300 members to balance performance and computational cost. We optimise the photochemical source by applying monthly regional scaling factors to the prior photochemical production fields described in Sect. <xref ref-type="sec" rid="Ch1.S2.SS3"/>. For the soil sink, we do not directly scale the fluxes as calculated within the respective schemes, as test inversions showed that this led to unstable Kalman filter performance. Instead, we optimise linearised representations of the soil sink fluxes, as explained in Sect. <xref ref-type="sec" rid="Ch1.S2.SS5.SSS1"/>.</p>
<sec id="Ch1.S2.SS5.SSS1">
  <label>2.5.1</label><title>Linearisation of soil sink and stratospheric flux</title>
      <p id="d2e1888">The soil sink and stratospheric fixing terms were linearised in the inversions to ensure stable Kalman filter performance. This addresses a well-known limitation in Kalman-based flux optimisation arising from strongly non-linear flux-state dependencies or from initial states far from the solution, as documented in carbon cycle inversions <xref ref-type="bibr" rid="bib1.bibx69" id="paren.58"/>. Such conditions are applicable here, since test inversions showed that large posterior-prior adjustments in <inline-formula><mml:math id="M113" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> mole fractions (up to 100 ppb) were required to reconcile global mass balance, large-scale gradients, and seasonal cycles. Since both processes depend directly on the <inline-formula><mml:math id="M114" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> mole fraction, this introduced a non-linear coupling between fluxes and mole fractions that led to instability of the Kalman filter. Therefore, daily linearised soil sink and stratospheric fluxes were obtained from forward simulations at <inline-formula><mml:math id="M115" display="inline"><mml:mrow><mml:mn mathvariant="normal">3</mml:mn><mml:mi mathvariant="italic">°</mml:mi><mml:mo>×</mml:mo><mml:mn mathvariant="normal">2</mml:mn><mml:mi mathvariant="italic">°</mml:mi></mml:mrow></mml:math></inline-formula> resolution with 34 vertical levels over 2003–2023, using the Sanderson and Bertagni schemes under the best-performing source-sink configurations identified in Sect. <xref ref-type="sec" rid="Ch1.S3.SS1"/> in relation to four observational metrics. Our inversion results support the linearisation, as posterior mole fractions remain close to the prior (Sect. <xref ref-type="sec" rid="Ch1.S3.SS2.SSS1"/>). Specifically, mean posterior deviations are <inline-formula><mml:math id="M116" display="inline"><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">2.0</mml:mn><mml:mspace linebreak="nobreak" width="0.125em"/><mml:mi mathvariant="italic">%</mml:mi></mml:mrow></mml:math></inline-formula> (Bertagni inversion) and <inline-formula><mml:math id="M117" display="inline"><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">11.6</mml:mn><mml:mspace linebreak="nobreak" width="0.125em"/><mml:mi mathvariant="italic">%</mml:mi></mml:mrow></mml:math></inline-formula> (Sanderson inversion), corresponding to flux differences of 1.1 and 6.6 <inline-formula><mml:math id="M118" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> under a fully first-order soil sink formulation. We deem this acceptable, as the differences are modest compared to the range of alternative prior soil sink perturbations (<inline-formula><mml:math id="M119" display="inline"><mml:mrow><mml:mo>±</mml:mo><mml:mspace width="0.25em" linebreak="nobreak"/><mml:mn mathvariant="normal">13</mml:mn></mml:mrow></mml:math></inline-formula> Tg yr<sup>−1</sup> relative to the base case) explored in the sensitivity analysis, which nevertheless showed convergence towards a stable posterior soil sink magnitude (Appendix Sect. <xref ref-type="sec" rid="App1.Ch1.S4"/>).</p>

<table-wrap id="T1" specific-use="star"><label>Table 1</label><caption><p id="d2e2009">Summary of parameters used in the inversions, as derived in Sects. <xref ref-type="sec" rid="App1.Ch1.S3.SS1"/> and <xref ref-type="sec" rid="App1.Ch1.S3.SS3"/> of the Appendix.</p></caption><oasis:table frame="topbot"><oasis:tgroup cols="4">
     <oasis:colspec colnum="1" colname="col1" align="left"/>
     <oasis:colspec colnum="2" colname="col2" align="left"/>
     <oasis:colspec colnum="3" colname="col3" align="left"/>
     <oasis:colspec colnum="4" colname="col4" align="left"/>
     <oasis:thead>
       <oasis:row rowsep="1">
         <oasis:entry colname="col1"/>
         <oasis:entry colname="col2">Soil sink</oasis:entry>
         <oasis:entry colname="col3">Isop. source</oasis:entry>
         <oasis:entry colname="col4">Non-isop. source</oasis:entry>
       </oasis:row>
     </oasis:thead>
     <oasis:tbody>
       <oasis:row>
         <oasis:entry colname="col1">Spatial discretisation</oasis:entry>
         <oasis:entry colname="col2">14 ecosystems (Sanderson)</oasis:entry>
         <oasis:entry colname="col3">14 ecosystems</oasis:entry>
         <oasis:entry colname="col4"><inline-formula><mml:math id="M121" display="inline"><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mi mathvariant="italic">°</mml:mi></mml:mrow></mml:math></inline-formula> latitude bands</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1"/>
         <oasis:entry colname="col2">22 soil textures (Bertagni)</oasis:entry>
         <oasis:entry colname="col3"/>
         <oasis:entry colname="col4"/>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">Temporal correlation timescale</oasis:entry>
         <oasis:entry colname="col2">24 months</oasis:entry>
         <oasis:entry colname="col3">12 months</oasis:entry>
         <oasis:entry colname="col4">36 months</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">Statevector element uncertainty</oasis:entry>
         <oasis:entry colname="col2">7.5 %</oasis:entry>
         <oasis:entry colname="col3">5 %</oasis:entry>
         <oasis:entry colname="col4">2.5 %</oasis:entry>
       </oasis:row>
     </oasis:tbody>
   </oasis:tgroup></oasis:table></table-wrap>

</sec>
<sec id="Ch1.S2.SS5.SSS2">
  <label>2.5.2</label><title>Spatial discretisation</title>
      <p id="d2e2124">We optimise the soil sink over the spatial regions used within each scheme: seven ecosystem regions for Sanderson and eleven soil texture classes for Bertagni (Table <xref ref-type="table" rid="T1"/>). In each region, the optimisation adjusts the magnitude of the linearised daily soil sink fluxes on <inline-formula><mml:math id="M122" display="inline"><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mi mathvariant="italic">°</mml:mi><mml:mo>×</mml:mo><mml:mn mathvariant="normal">1</mml:mn><mml:mi mathvariant="italic">°</mml:mi></mml:mrow></mml:math></inline-formula> resolution, which are effectively treated as negative surface emissions within the inversion. To introduce additional interhemispheric degrees of freedom in the statevector, the regions are subdivided at 25<inline-formula><mml:math id="M123" display="inline"><mml:mi mathvariant="italic">°</mml:mi></mml:math></inline-formula> N, resulting in 14 and 22 regions, respectively (Fig. <xref ref-type="fig" rid="FF1"/> in the Appendix). This choice accounts for the asymmetric land distribution between hemispheres and approximately balances land area between the northern and southern portions of the modelling domain. For the photochemical source, we separately optimise pre-calculated <inline-formula><mml:math id="M124" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> production fields from isoprene and non-isoprene oxidation, to account for differences in the uncertainties of their three-dimensional distributions. The isoprene <inline-formula><mml:math id="M125" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source is optimised over the same 14 Sanderson regions as the soil sink, assuming that isoprene emissions are predominantly biogenic and ecosystem-dependent. This approach reflects the short atmospheric lifetime of isoprene, which leads to relatively localised <inline-formula><mml:math id="M126" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula> production from its oxidation with respect to the underlying emissions. We optimise the non-isoprene <inline-formula><mml:math id="M127" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source using <inline-formula><mml:math id="M128" display="inline"><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mi mathvariant="italic">°</mml:mi></mml:mrow></mml:math></inline-formula> latitudinal bands, with a correlation length scale of 1000 km, yielding roughly 18 degrees of freedom from pole to pole. This reflects the relatively long atmospheric lifetime of <inline-formula><mml:math id="M129" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">4</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>, the dominant precursor of <inline-formula><mml:math id="M130" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula> in this term, yielding a relatively uniform spatial <inline-formula><mml:math id="M131" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> production pattern from its oxidation.</p>
</sec>
<sec id="Ch1.S2.SS5.SSS3">
  <label>2.5.3</label><title>Temporal correlation structure</title>
      <p id="d2e2255">We apply a temporal correlation structure to the prior flux to ensure coherent adjustments to fluxes over seasonal and sub-seasonal timescales. Annually recurring changes to the seasonal cycle are imposed through flux correlations between the same month of year (MOY) of neighbouring years. This correlation drops off exponentially with a decay constant (<inline-formula><mml:math id="M132" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="italic">σ</mml:mi><mml:mi mathvariant="normal">MOY</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>) of 36, 24, and 12 months for the non-isoprene <inline-formula><mml:math id="M133" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source, soil sink, and the isoprene <inline-formula><mml:math id="M134" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source respectively. These <inline-formula><mml:math id="M135" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="italic">σ</mml:mi><mml:mi mathvariant="normal">MOY</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> values were derived from the sensitivity tests described in Sect. <xref ref-type="sec" rid="App1.Ch1.S3.SS1"/> of the Appendix. In addition, a month-to-month correlation is imposed across all prior fluxes to reduce the degrees of freedom and to avoid unrealistically high flux variability. This correlation (<inline-formula><mml:math id="M136" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="italic">σ</mml:mi><mml:mi mathvariant="normal">MON</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>) decays exponentially with a time scale of 3 months.</p>
</sec>
<sec id="Ch1.S2.SS5.SSS4">
  <label>2.5.4</label><title>Statevector element uncertainty</title>
      <p id="d2e2325">To reflect the relative confidence in our prior fluxes, we assign statevector element uncertainties across space and time of 7.5 % for the soil sink, 5 % for the isoprene <inline-formula><mml:math id="M137" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source, and 2.5 % for the non-isoprene <inline-formula><mml:math id="M138" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source. We note that these values can not be directly interpreted in flux space (e.g. to compute prior uncertainties in <inline-formula><mml:math id="M139" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>), as their effect on fluxes is mediated by the inversion and cannot be directly inferred.</p>
</sec>
<sec id="Ch1.S2.SS5.SSS5">
  <label>2.5.5</label><title>Model-data mismatch</title>
      <p id="d2e2375">We define a model–data mismatch (MDM) of 3.0 ppb for most sites, following <xref ref-type="bibr" rid="bib1.bibx48" id="text.59"/>. The MDM represents the uncertainty assigned to each observation, accounting for measurement, transport, and representativeness errors. The MDM determines the relative weight of each observation in the inversion with smaller values imposing a stronger constraint on the posterior flux. Our value of 3.0 ppb exceeds the measurement uncertainties reported in <xref ref-type="bibr" rid="bib1.bibx52" id="text.60"/>, allowing for errors beyond instrumental uncertainty. For Deep Southern Hemisphere sites, we assign a MDM of 1.5 ppb, reflecting lower transport errors (Appendix Sect. <xref ref-type="sec" rid="App1.Ch1.S3.SS2"/>). When multiple observations are available on a given day for a site, the MDM for that day is scaled by the square root of the number of observations to avoid overweighting. We exclude the NOAA-GML surface flask sites AMY (Anmyeon-do, South Korea) and SDZ (Shangdianzi, China) from the optimisation, as they are strongly influenced by local fossil fuel emissions that cannot be resolved at the model's spatial resolution. Aircraft <inline-formula><mml:math id="M140" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> observations are also excluded to retain them for independent evaluation.</p>

      <fig id="F3" specific-use="star"><label>Figure 3</label><caption><p id="d2e2399">Sensitivity analysis for various source-sink configurations using the Sanderson (top row) and Bertagni (bottom row) schemes. Shown are: <bold>(a)</bold> mean RMSE across five stations (ALT, CGO, MHD, MLO, SPO); <bold>(b)</bold> <inline-formula><mml:math id="M141" display="inline"><mml:mi mathvariant="normal">Δ</mml:mi></mml:math></inline-formula> Growth Rate at MLO; <bold>(c)</bold> <inline-formula><mml:math id="M142" display="inline"><mml:mi mathvariant="normal">Δ</mml:mi></mml:math></inline-formula> Interpolar Difference (ALT-SPO); and <bold>(d)</bold> <inline-formula><mml:math id="M143" display="inline"><mml:mi mathvariant="normal">Δ</mml:mi></mml:math></inline-formula> Absolute Seasonal Amplitude at MHD.</p></caption>
            <graphic xlink:href="https://acp.copernicus.org/articles/26/16261/2026/acp-26-16261-2026-f03.png"/>

          </fig>

</sec>
</sec>
</sec>
<sec id="Ch1.S3">
  <label>3</label><title>Results</title>
<sec id="Ch1.S3.SS1">
  <label>3.1</label><title>Balancing global mean sources and sinks</title>
      <p id="d2e2459">Figure <xref ref-type="fig" rid="F3"/> shows the model performance evaluated against four observational metrics, across different global scaling factors applied to the soil sink (<inline-formula><mml:math id="M144" display="inline"><mml:mi>x</mml:mi></mml:math></inline-formula>-axis) and photochemical <inline-formula><mml:math id="M145" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source (<inline-formula><mml:math id="M146" display="inline"><mml:mi>y</mml:mi></mml:math></inline-formula>-axis). Here, the upper and lower panels correspond to the Sanderson and Bertagni simulations, respectively (see Sect. <xref ref-type="sec" rid="Ch1.S2.SS4"/>). Because the Bertagni scheme yields a lower prior soil sink, a broader range of scaling factors is required to span the full literature range.</p>
      <p id="d2e2491">We find that the source–sink balance that best agrees with global <inline-formula><mml:math id="M147" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> observations is close to that reported by <xref ref-type="bibr" rid="bib1.bibx41" id="text.61"/> and earlier bottom-up studies. This optimum lies in the lower-left quadrants of Fig. <xref ref-type="fig" rid="F3"/>, where low RMSE (panel a), a realistic growth rate (panel b), and a realistic interhemispheric difference (panel c) are simultaneously achieved, as indicated by the green shading. In contrast, simulating a realistic seasonal amplitude (panel d) requires a substantially larger soil sink, particularly in the Sanderson simulations. This highlights the limitations of using only two global scaling factors and motivates a more elaborate inversion approach. For the Bertagni simulations, the different metrics point to a more consistent optimum, with the seasonal amplitude in better agreement with the other diagnostics. The upper-right quadrants of Fig. <xref ref-type="fig" rid="F3"/> also show that the large soil sink and photochemical source magnitudes reported by <xref ref-type="bibr" rid="bib1.bibx54" id="text.62"/> and <xref ref-type="bibr" rid="bib1.bibx80" id="text.63"/>, derived using 2D box model inversions, perform poorly when implemented in a 3D global chemical transport model.</p>
      <p id="d2e2519">We identify the best-performing configurations as indicated by the black crosses in Fig. <xref ref-type="fig" rid="F3"/>. Here, the global photochemical source is scaled by 0.7 and the soil sink by 0.8 (Sanderson) and 1.2 (Bertagni), corresponding to a photochemical source of 38 <inline-formula><mml:math id="M148" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> and soil sinks of 59 <inline-formula><mml:math id="M149" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> (Sanderson) and 54 <inline-formula><mml:math id="M150" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> (Bertagni) in the final simulation year. These fluxes are consistent with most previous literature, including bottom-up <xref ref-type="bibr" rid="bib1.bibx15" id="paren.64"/> and 3D inverse modelling studies <xref ref-type="bibr" rid="bib1.bibx48 bib1.bibx84 bib1.bibx7" id="paren.65"/>. We therefore use them as the starting point for our detailed inversions (hereafter referred to as the “priors”). Additionally, we define two alternative prior configurations for the Sanderson and Bertagni scheme, with lower and higher soil sinks and photochemical sources. These alternative configurations (black triangles) are used to assess the sensitivity of the inversions to the prior magnitudes, as described in Appendix <xref ref-type="sec" rid="App1.Ch1.S4"/>.</p>

      <fig id="F4" specific-use="star"><label>Figure 4</label><caption><p id="d2e2587">Time series of observed (black dots) and simulated (coloured lines) <inline-formula><mml:math id="M151" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> mole fractions at ALT <bold>(a)</bold> and SPO <bold>(b)</bold>. Dotted and dashed lines show simulations before optimisation (prior), solid lines show simulations after optimisation (posterior). Red indicates Sanderson inversion, blue indicates Bertagni inversion. Panel <bold>(c)</bold> shows the prior and posterior mean absolute error (MAE), quantifying how well our simulated mole fractions match independent aircraft measurements from four campaigns. Here, the bars are ordered from left to right as Sanderson prior, Sanderson posterior, Bertagni prior, and Bertagni posterior for each campaign.</p></caption>
          <graphic xlink:href="https://acp.copernicus.org/articles/26/16261/2026/acp-26-16261-2026-f04.png"/>

        </fig>

</sec>
<sec id="Ch1.S3.SS2">
  <label>3.2</label><title>Inversions</title>
<sec id="Ch1.S3.SS2.SSS1">
  <label>3.2.1</label><title>Evaluation with surface and aircraft observations</title>
      <p id="d2e2631">Our detailed inversions substantially improve the agreement between simulated and observed <inline-formula><mml:math id="M152" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> mole fractions used in the inversions. Globally, the mean residual across all sites decreases from 69.6 (prior) to <inline-formula><mml:math id="M153" display="inline"><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">0.8</mml:mn></mml:mrow></mml:math></inline-formula> ppb (posterior) in the Sanderson inversion, and from 11.0 (prior) to <inline-formula><mml:math id="M154" display="inline"><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">0.7</mml:mn></mml:mrow></mml:math></inline-formula> ppb (posterior) in the Bertagni inversion. Figure <xref ref-type="fig" rid="F4"/> shows that our inversions accurately capture the long-term growth rate and seasonal cycle at ALT (panel a) and SPO (panel b), representing the Northern and Southern Hemisphere, respectively. In addition, panel c shows that the inversions generally improve the fit to independent aircraft observations from the ATom, HIPPO, NOAA-GML, and IAGOS-CARIBIC campaigns, giving additional confidence in the results. A more detailed evaluation against aircraft measurements is provided in Table <xref ref-type="table" rid="TF2"/> in the Appendix, showing a modest latitudinal gradient in MAE, with generally lower values in the Southern Hemisphere, while altitudinal differences are less systematic and vary between campaigns.</p>
</sec>
<sec id="Ch1.S3.SS2.SSS2">
  <label>3.2.2</label><title>The global hydrogen budget</title>
      <p id="d2e2677">We find a global mean photochemical source of 34.6 [29.2–38.2] <inline-formula><mml:math id="M155" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> and a soil sink of 52.8 [47.8–56.7] <inline-formula><mml:math id="M156" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> for the period 2003–2023 (Table <xref ref-type="table" rid="T2"/>). Here, the central values represent the mean of our best-estimate Sanderson and Bertagni inversions (hereafter referred to as the “posterior”), and the bracketed values denote the posterior error range derived in Appendix <xref ref-type="sec" rid="App1.Ch1.S4"/>. Figure <xref ref-type="fig" rid="F5"/> shows that our posterior photochemical source is lower than most previous estimates (12 of 15 studies) but remains within the one-standard-deviation range given by <xref ref-type="bibr" rid="bib1.bibx41" id="text.66"/>. It also shows that our posterior soil sink agrees well with most previous studies and is close to the <xref ref-type="bibr" rid="bib1.bibx41" id="text.67"/> ensemble mean. However, our posterior error range is much narrower than their soil sink one-standard-deviation range (32–68 <inline-formula><mml:math id="M157" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>), despite covering a broad range of inversion parameters.</p>

<table-wrap id="T2" specific-use="star"><label>Table 2</label><caption><p id="d2e2747">Optimised global <inline-formula><mml:math id="M158" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> budget for 2003–2023 based on the mean of the best-estimate inversions using the Sanderson and Bertagni soil sink schemes. Bracketed values indicate posterior error ranges for the optimised photochemical source and soil sink, and variability across inversion experiments for the non-optimised photochemical sink (see Appendix <xref ref-type="sec" rid="App1.Ch1.S4"/>). Literature values are shown for comparison, noting differences in analysis periods.</p></caption><oasis:table frame="topbot"><oasis:tgroup cols="5">
     <oasis:colspec colnum="1" colname="col1" align="left"/>
     <oasis:colspec colnum="2" colname="col2" align="right"/>
     <oasis:colspec colnum="3" colname="col3" align="right"/>
     <oasis:colspec colnum="4" colname="col4" align="right"/>
     <oasis:colspec colnum="5" colname="col5" align="right"/>
     <oasis:thead>
       <oasis:row rowsep="1">
         <oasis:entry colname="col1"/>
         <oasis:entry colname="col2">This study<sup>a</sup></oasis:entry>
         <oasis:entry colname="col3">BU<sup>b</sup></oasis:entry>
         <oasis:entry colname="col4">TD<sup>c</sup></oasis:entry>
         <oasis:entry colname="col5">
                      <xref ref-type="bibr" rid="bib1.bibx41" id="text.71"/>
                    <sup>d</sup></oasis:entry>
       </oasis:row>
     </oasis:thead>
     <oasis:tbody>
       <oasis:row rowsep="1">
         <oasis:entry namest="col1" nameend="col5">Sources </oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">Fossil fuel burning</oasis:entry>
         <oasis:entry colname="col2">14.1</oasis:entry>
         <oasis:entry colname="col3">14.4–20.0</oasis:entry>
         <oasis:entry colname="col4">15.0–22.0</oasis:entry>
         <oasis:entry colname="col5">7.5 [3.6–11.4]</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">Biomass burning</oasis:entry>
         <oasis:entry colname="col2">13.6</oasis:entry>
         <oasis:entry colname="col3">9.0–20.0</oasis:entry>
         <oasis:entry colname="col4">16.0</oasis:entry>
         <oasis:entry colname="col5">11.6 [7.9–15.3]</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">Land <inline-formula><mml:math id="M170" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">N</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>-fixation</oasis:entry>
         <oasis:entry colname="col2">3.0</oasis:entry>
         <oasis:entry colname="col3">0.0–4.0</oasis:entry>
         <oasis:entry colname="col4">6.0</oasis:entry>
         <oasis:entry colname="col5">3.1 [0.6–5.6]</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">Ocean <inline-formula><mml:math id="M171" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">N</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>-fixation</oasis:entry>
         <oasis:entry colname="col2">5.0</oasis:entry>
         <oasis:entry colname="col3">4.0–6.0</oasis:entry>
         <oasis:entry colname="col4">6.0</oasis:entry>
         <oasis:entry colname="col5">4.9 [1.8–8.0]</oasis:entry>
       </oasis:row>
       <oasis:row rowsep="1">
         <oasis:entry colname="col1">Photochemical source</oasis:entry>
         <oasis:entry colname="col2">34.6 [29.2–38.2]</oasis:entry>
         <oasis:entry colname="col3">30.2–47</oasis:entry>
         <oasis:entry colname="col4">38.6–76.0</oasis:entry>
         <oasis:entry colname="col5">38.4 [32.3–44.5]</oasis:entry>
       </oasis:row>
       <oasis:row rowsep="1">
         <oasis:entry namest="col1" nameend="col5">Sinks </oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">Soil sink</oasis:entry>
         <oasis:entry colname="col2">52.8 [47.8–56.7]</oasis:entry>
         <oasis:entry colname="col3">52.9–58.5</oasis:entry>
         <oasis:entry colname="col4">43.5–88.0</oasis:entry>
         <oasis:entry colname="col5">50 [32–68]</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">Photochemical sink</oasis:entry>
         <oasis:entry colname="col2">19.2 [19.1–19.7]</oasis:entry>
         <oasis:entry colname="col3">15.0–22.8</oasis:entry>
         <oasis:entry colname="col4">15.0–19.0</oasis:entry>
         <oasis:entry colname="col5">18.4 [16.2–20.6]</oasis:entry>
       </oasis:row>
     </oasis:tbody>
   </oasis:tgroup></oasis:table><table-wrap-foot><p id="d2e2763"><sup>a</sup> Includes stratospheric fixing term, with mean magnitude of 3.3 Tg yr<sup>−1</sup> (Sanderson) and 0.2 Tg yr<sup>−1</sup> (Bertagni) over 2003–2023. <sup>b</sup> Bottom-up studies <xref ref-type="bibr" rid="bib1.bibx39 bib1.bibx23 bib1.bibx60 bib1.bibx50 bib1.bibx83 bib1.bibx46 bib1.bibx47 bib1.bibx43 bib1.bibx57" id="paren.68"/>. <sup>c</sup> Top-down studies using 2D <xref ref-type="bibr" rid="bib1.bibx54 bib1.bibx80" id="paren.69"/> and 3D models <xref ref-type="bibr" rid="bib1.bibx48 bib1.bibx7 bib1.bibx84" id="paren.70"/>. <sup>d</sup> Recent key bottom-up study. First value denotes ensemble mean; brackets denote <inline-formula><mml:math id="M165" display="inline"><mml:mrow><mml:mo>±</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:math></inline-formula> standard deviation.</p></table-wrap-foot></table-wrap>

      <fig id="F5" specific-use="star"><label>Figure 5</label><caption><p id="d2e3082">Comparison of our fluxes with previous studies. Purple dots and triangles indicate the mean of our pre-balanced priors and alternative priors (Sanderson and Bertagni), respectively (see Sect. <xref ref-type="sec" rid="Ch1.S3.SS1"/>). Horizontal purple lines with shading show our posterior and posterior error range as derived in Appendix <xref ref-type="sec" rid="App1.Ch1.S4"/>. Previous studies are summarised in box plots grouped by methodological approach (see Table <xref ref-type="table" rid="T2"/>), with <inline-formula><mml:math id="M172" display="inline"><mml:mi>N</mml:mi></mml:math></inline-formula> denoting the number of studies in each group and whiskers extending to <inline-formula><mml:math id="M173" display="inline"><mml:mrow><mml:mn mathvariant="normal">1.5</mml:mn><mml:mspace width="0.125em" linebreak="nobreak"/><mml:mi mathvariant="normal">IQR</mml:mi></mml:mrow></mml:math></inline-formula>. Estimates from <xref ref-type="bibr" rid="bib1.bibx41" id="text.72"/> are shown as black circles with ±1 standard deviation whiskers.</p></caption>
            <graphic xlink:href="https://acp.copernicus.org/articles/26/16261/2026/acp-26-16261-2026-f05.png"/>

          </fig>

</sec>
<sec id="Ch1.S3.SS2.SSS3">
  <label>3.2.3</label><title>Spatial variability</title>
      <p id="d2e3127">We find a high degree of consistency between the posterior-prior flux corrections applied in the Sanderson and Bertagni inversions. For the soil sink, both inversions infer weaker uptake in the Northern Hemisphere mid-latitudes and stronger uptake in the tropics compared to their priors. This reflects reduced uptake in the RECCAP2 (REgional Carbon Cycle Assessment and Processes Phase 2; <xref ref-type="bibr" rid="bib1.bibx9" id="altparen.73"/>) regions of Europe, North America, East Asia, and Russia, and increased uptake in South America, Africa, and Southeast Asia (see Appendix Figs. <xref ref-type="fig" rid="FF2"/>, <xref ref-type="fig" rid="FF3"/>, <xref ref-type="fig" rid="FF4"/>, <xref ref-type="fig" rid="FF5"/>). Figure <xref ref-type="fig" rid="F6"/> shows that relative to the <xref ref-type="bibr" rid="bib1.bibx41" id="text.74"/> ensemble mean, our posterior soil sink is roughly 45 % and 70 % stronger in South America and Russia and 45 % and 35 % weaker in the Middle East and Oceania, respectively. For the non-isoprene <inline-formula><mml:math id="M174" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source, both inversions yield an increase in the Northern Hemisphere (15.7 <inline-formula><mml:math id="M175" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> Sanderson; 15.9 <inline-formula><mml:math id="M176" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> Bertagni, up from 15.4 <inline-formula><mml:math id="M177" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>) and a decrease in the Southern Hemisphere (10.9 <inline-formula><mml:math id="M178" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> Sanderson; 10.8 <inline-formula><mml:math id="M179" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> Bertagni, down from 11.8 <inline-formula><mml:math id="M180" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>) relative to the prior. This corresponds to larger sources over Europe, North America, East Asia, and the Middle East, and smaller sources over South America and Oceania (Appendix Figs. <xref ref-type="fig" rid="FF2"/>, <xref ref-type="fig" rid="FF3"/>, <xref ref-type="fig" rid="FF4"/>, <xref ref-type="fig" rid="FF5"/>). For the isoprene <inline-formula><mml:math id="M181" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source, both inversions infer a consistent decrease across latitudes, with the largest reductions in the Southern Hemisphere (4.7 <inline-formula><mml:math id="M182" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> Sanderson; 5.0 <inline-formula><mml:math id="M183" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> Bertagni, down from 7.0 <inline-formula><mml:math id="M184" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>).</p>

      <fig id="F6" specific-use="star"><label>Figure 6</label><caption><p id="d2e3334">Soil sink fluxes (2003–2023 mean) for four selected RECCAP2 regions. Light bars show the prior and dark bars the posterior, taken as the mean of the Sanderson and Bertagni inversions. Black dots with whiskers show the ensemble mean and one standard deviation range from <xref ref-type="bibr" rid="bib1.bibx41" id="text.75"/>.</p></caption>
            <graphic xlink:href="https://acp.copernicus.org/articles/26/16261/2026/acp-26-16261-2026-f06.png"/>

          </fig>


</sec>
<sec id="Ch1.S3.SS2.SSS4">
  <label>3.2.4</label><title>Long-term growth rate of atmospheric hydrogen</title>
      <p id="d2e3356">Figures <xref ref-type="fig" rid="F7"/>a and <xref ref-type="fig" rid="F7"/>b show the observed annual <inline-formula><mml:math id="M185" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> growth rate from 2003–2023 (grey bars) at SPO and MLO, alongside the simulated posterior growth rate (purple bars). The record exhibits three distinct stages, most clearly at SPO: a moderate growth rate from 2003–2008, a decline from 2008–2016, and a subsequent increase after 2016. Our prior simulations do not adequately capture this temporal evolution, most notably failing to reproduce the pronounced increase in <inline-formula><mml:math id="M186" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> after 2016 (Fig. <xref ref-type="fig" rid="F4"/>). The prior biomass burning emissions (GFED5) partly explain the decline in growth rate after 2008 (Fig. <xref ref-type="fig" rid="F7"/>c), whereas the prior fossil fuel burning source (Fig. <xref ref-type="fig" rid="F7"/>d) and prior isoprene <inline-formula><mml:math id="M187" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source (Fig. <xref ref-type="fig" rid="F7"/>e; dotted grey line) remain relatively constant over time. The prior non-isoprene <inline-formula><mml:math id="M188" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source (Fig. <xref ref-type="fig" rid="F7"/>f; dotted grey line) contributes to the increase after 2016, reflecting the increasing global <inline-formula><mml:math id="M189" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">4</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> mole fractions in this period. This contribution is, however, partly offset by a strengthening prior soil sink (Fig. <xref ref-type="fig" rid="F7"/>g; dotted grey line).</p>

      <fig id="F7" specific-use="star"><label>Figure 7</label><caption><p id="d2e3434">Observed and simulated posterior annual growth rates at SPO <bold>(a)</bold> and MLO <bold>(b)</bold>. Remaining panels show global annual fluxes for <bold>(c)</bold> biomass burning, <bold>(d)</bold> fossil fuel burning, <bold>(e)</bold> isoprene <inline-formula><mml:math id="M190" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source, <bold>(f)</bold> non-isoprene <inline-formula><mml:math id="M191" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source, <bold>(g)</bold> soil sink, and <bold>(h)</bold> net sources and sinks. The vertical lines delineate three periods (2003–2008, 2008–2016, 2016–2024) that characterise the observed <inline-formula><mml:math id="M192" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> growth rate.</p></caption>
            <graphic xlink:href="https://acp.copernicus.org/articles/26/16261/2026/acp-26-16261-2026-f07.png"/>

          </fig>

      <p id="d2e3501">In the posterior, the fit to the growth rate is primarily improved by adjusting the non-isoprene <inline-formula><mml:math id="M193" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source (Fig. <xref ref-type="fig" rid="F7"/>f), consistent with the hypothesis of <xref ref-type="bibr" rid="bib1.bibx41" id="text.76"/> that <inline-formula><mml:math id="M194" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">4</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>-driven photochemical production is a dominant control on the atmospheric <inline-formula><mml:math id="M195" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> growth rate. The posterior isoprene <inline-formula><mml:math id="M196" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source instead decreases after 2016 (Fig. <xref ref-type="fig" rid="F7"/>e), but this decrease is smaller than the increase in the non-isoprene source, resulting in a net increase in posterior photochemical <inline-formula><mml:math id="M197" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> sources. Crucially, both inversions also infer a weakening global soil sink over the last two decades, at an average rate of 0.23 Tg yr<sup>−2</sup> (Sanderson: 0.26; Bertagni: 0.20 Tg yr<sup>−2</sup>). This result is noteworthy because a larger increase in the posterior non-isoprene source than inferred here could also have explained the observed increase in <inline-formula><mml:math id="M200" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> growth after 2016. Instead, both inversions consistently infer a weakening soil sink, despite the increasing soil sink in their respective priors. This suggests that environmental limitations on soil uptake have strengthened over time, a change not captured by either prior deposition scheme. Given the theoretical temperature optimum of roughly 30 <inline-formula><mml:math id="M201" display="inline"><mml:mi mathvariant="italic">°</mml:mi></mml:math></inline-formula>C for microbial activity <xref ref-type="bibr" rid="bib1.bibx61 bib1.bibx16" id="paren.77"/>, we consider a temperature-driven explanation less likely in light of the observed climate warming. Changes in soil moisture or other controls on microbial activity may instead contribute to the inferred trend. This finding warrants further investigation given its potential implications for future atmospheric <inline-formula><mml:math id="M202" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> levels under the expected increase in anthropogenic emissions during the energy transition.</p>
</sec>
<sec id="Ch1.S3.SS2.SSS5">
  <label>3.2.5</label><title>Seasonality and Interannual variability of the soil sink</title>
      <p id="d2e3633">Both inversions infer an increase in the seasonal amplitude of the <inline-formula><mml:math id="M203" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> soil sink, together with a shift towards a later seasonal maximum in uptake relative to the priors. The mean seasonal range of the monthly soil sink over 2003–2023 increases in all RECCAP2 regions except Europe in the Sanderson inversion and the Middle East in the Bertagni inversion (Appendix Table <xref ref-type="table" rid="TF3"/> and Fig. <xref ref-type="fig" rid="FF5"/>), with the largest increases in South America, Africa, Russia, and North America. In South America, the seasonal range increases from 3.3 to 8.4 <inline-formula><mml:math id="M204" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> (Sanderson) and from 2.1 to 5.5 <inline-formula><mml:math id="M205" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> (Bertagni), in Africa from 1.2 to 6.0 and from 2.5 to 4.7 <inline-formula><mml:math id="M206" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>, respectively, in Russia from 6.5 to 9.0 <inline-formula><mml:math id="M207" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> and from 7.4 to 10.3 <inline-formula><mml:math id="M208" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>, and in North America from 5.4 to 8.1 and from 6.2 to 7.4 <inline-formula><mml:math id="M209" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>. The strongest shifts in seasonal timing occur in Asia, particularly South Asia, East Asia, and Southeast Asia. Across all RECCAP2 regions, maximum uptake occurs 25.0 d later on average in the Sanderson posterior and 28.3 d later in the Bertagni posterior. We compare our inferred soil sink seasonality with independent estimates by <xref ref-type="bibr" rid="bib1.bibx63" id="text.78"/> and <xref ref-type="bibr" rid="bib1.bibx66" id="text.79"/> in Appendix <xref ref-type="sec" rid="App1.Ch1.S2"/>. We find excellent agreement with <xref ref-type="bibr" rid="bib1.bibx63" id="text.80"/> in both seasonal range and timing, but poor agreement with <xref ref-type="bibr" rid="bib1.bibx66" id="text.81"/>, likely due to substantial methodological differences between our studies.</p>

      <fig id="F8" specific-use="star"><label>Figure 8</label><caption><p id="d2e3771">Monthly soil sink flux anomalies for tropical South America (RECCAP2 region), defined relative to a month-specific climatology over the final 10 years of the inversions (2014–2024), plotted against the Oceanic Niño Index <xref ref-type="bibr" rid="bib1.bibx3" id="paren.82"/>. Left panels correspond to the Sanderson inversion, right panels to the Bertagni inversion.</p></caption>
            <graphic xlink:href="https://acp.copernicus.org/articles/26/16261/2026/acp-26-16261-2026-f08.png"/>

          </fig>

      <p id="d2e3783">Our inversions also show a clear sensitivity of the soil sink to the El Niño–Southern Oscillation (ENSO). This response is strongest in tropical South America (Fig. <xref ref-type="fig" rid="F8"/>) but remains statistically significant globally, including weaker but significant correlations for several other regions (Table <xref ref-type="table" rid="TF4"/> in the Appendix). We hypothesise that ENSO primarily influences the soil sink through soil moisture. According to <xref ref-type="bibr" rid="bib1.bibx6" id="text.83"/>, in most regions, including tropical South America, the soil sink is dominantly diffusion-limited (i.e., too wet). Therefore, the drier conditions in the Amazon associated with El Niño may temporarily reduce the diffusive barrier and increase the soil sink, whereas wetter conditions during La Niña increase the diffusive barrier and dampen the soil sink. This could explain the upper right and lower left quadrants of Fig. <xref ref-type="fig" rid="F8"/>, respectively. Notably, this ENSO–soil sink relationship was only very weakly present in the priors, but emerges much more clearly in the posteriors, highlighting the additional information provided by the atmospheric constraints.</p>
</sec>
</sec>
</sec>
<sec id="Ch1.S4">
  <label>4</label><title>Discussion and outlook</title>
<sec id="Ch1.S4.SS1">
  <label>4.1</label><title>Uncertainty and potential aliasing in the inferred H<sub>2</sub> budget</title>
      <p id="d2e3822">We have relatively high confidence in our inferred soil sink, supported by the use of observationally constrained <inline-formula><mml:math id="M211" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">OH</mml:mi></mml:mrow></mml:math></inline-formula> fields and the agreement of our estimated photochemical sink with other bottom-up estimates. To estimate the photochemical sink, we used <inline-formula><mml:math id="M212" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">OH</mml:mi></mml:mrow></mml:math></inline-formula> fields from the CAMS EAC4 reanalysis, which provides globally consistent atmospheric composition fields constrained by satellite observations of <inline-formula><mml:math id="M213" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">O</mml:mi><mml:mn mathvariant="normal">3</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>, <inline-formula><mml:math id="M214" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">CO</mml:mi></mml:mrow></mml:math></inline-formula>, <inline-formula><mml:math id="M215" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">NO</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>, and aerosol optical depth, alongside meteorological observations <xref ref-type="bibr" rid="bib1.bibx28" id="paren.84"/>. Our resulting photochemical sink of 19.2 [19.1–19.7] <inline-formula><mml:math id="M216" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> falls within the range of other bottom-up estimates (15.0–22.8 <inline-formula><mml:math id="M217" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>; see Table <xref ref-type="table" rid="T2"/>) and is very close to the mean estimate of 18.4 [16.2–20.6] <inline-formula><mml:math id="M218" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> reported by <xref ref-type="bibr" rid="bib1.bibx41" id="text.85"/>. We note that <xref ref-type="bibr" rid="bib1.bibx57" id="text.86"/> reported a higher photochemical sink of 25 [22–30] <inline-formula><mml:math id="M219" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> from an ensemble of global three-dimensional atmospheric chemistry models. However, <xref ref-type="bibr" rid="bib1.bibx81" id="text.87"/> showed that these models tend to overestimate OH, which explains the higher estimates of the photochemical <inline-formula><mml:math id="M220" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> sink. Supporting this, <xref ref-type="bibr" rid="bib1.bibx57" id="text.88"/> report <inline-formula><mml:math id="M221" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">4</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> lifetimes due to <inline-formula><mml:math id="M222" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">OH</mml:mi></mml:mrow></mml:math></inline-formula> oxidation of 8.0 [6.9–9.2] years across their ensemble of models, substantially shorter than the lifetime of 11.2 <inline-formula><mml:math id="M223" display="inline"><mml:mo>±</mml:mo></mml:math></inline-formula> 1.3 years inferred from observations <xref ref-type="bibr" rid="bib1.bibx49" id="paren.89"/>.</p>
      <p id="d2e3999">We note that our reported range of 19.2 [19.1–19.7] <inline-formula><mml:math id="M224" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> represents variability in the mean photochemical sink over 2003–2023 across our inversion experiments, rather than an uncertainty estimate that includes <inline-formula><mml:math id="M225" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">OH</mml:mi></mml:mrow></mml:math></inline-formula> uncertainty. This variability arises from differences in the posterior photochemical source, posterior soil sink, and spatiotemporal distribution of posterior mole fractions, as the photochemical sink is not directly optimised in our inversions. We are aware that the range in our photochemical <inline-formula><mml:math id="M226" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> sink is narrow relative to the broader uncertainties in global <inline-formula><mml:math id="M227" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">OH</mml:mi></mml:mrow></mml:math></inline-formula> distributions, which partly reflects our use of the same CAMS EAC4 <inline-formula><mml:math id="M228" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">OH</mml:mi></mml:mrow></mml:math></inline-formula> fields in all experiments. Therefore, to illustrate the potential implications of <inline-formula><mml:math id="M229" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">OH</mml:mi></mml:mrow></mml:math></inline-formula>-field uncertainties, we consider a hypothetical 20 % bias in <inline-formula><mml:math id="M230" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">OH</mml:mi></mml:mrow></mml:math></inline-formula> fields. Under a first-order approximation, assuming full compensation in the inversion by an opposite change in the soil sink, this would result in a change in our posterior soil sink of approximately 3.9 <inline-formula><mml:math id="M231" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>. This change is relatively small compared to the posterior error range on the soil sink (47.8–56.7 <inline-formula><mml:math id="M232" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>), supporting our conclusion that the inferred soil sink is relatively robust against 20 % variations in <inline-formula><mml:math id="M233" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">OH</mml:mi></mml:mrow></mml:math></inline-formula>. Changes in the north–south gradient of <inline-formula><mml:math id="M234" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">OH</mml:mi></mml:mrow></mml:math></inline-formula> <xref ref-type="bibr" rid="bib1.bibx36 bib1.bibx37" id="paren.90"/> or in its interannual variations <xref ref-type="bibr" rid="bib1.bibx42" id="paren.91"/> would likely affect the photochemical sink and hence soil sink patterns within this global constraint, but these effects are not quantified in our study.</p>
      <p id="d2e4128">We are also confident in our inferred total sources, although constraining the photochemical source is more challenging. This arises from the fact that several other sources in the hydrogen budget are prescribed in our inversions rather than optimised, introducing a potential for aliasing, whereby biases in these prescribed terms are compensated by adjustments in the posterior photochemical source or soil sink. For example, the three-stage pattern in our posterior photochemical source could also partly reflect a systematic underestimation of biomass burning <inline-formula><mml:math id="M235" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> emissions in the GFED5 prior, which showed a similar temporal structure. However, this is unlikely to fully explain our inferred trend, as GFED5 <inline-formula><mml:math id="M236" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> emissions already lie at the upper end of current literature estimates <xref ref-type="bibr" rid="bib1.bibx41" id="paren.92"/>. Similarly, sources not included in our model provide an additional pathway for aliasing. For example, geological sources and leakage-related industrial emissions were omitted due to the lack of robust emission inventories and their expected small magnitude <xref ref-type="bibr" rid="bib1.bibx41" id="paren.93"/>. If sufficiently large and increasing over time, these fluxes could in principle contribute to the inferred weakening of the global soil sink. This is unlikely for geological sources, for which no clear mechanisms support a sustained long-term increase. Industrial emissions provide a more plausible pathway, as global hydrogen demand increased from 52.5 Tg in 2000 <xref ref-type="bibr" rid="bib1.bibx26" id="paren.94"/> to 95 Tg in 2023 <xref ref-type="bibr" rid="bib1.bibx27" id="paren.95"/>, implying an approximate doubling of emissions under constant loss rates. However, fully explaining our inferred soil sink trend (4.8 <inline-formula><mml:math id="M237" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> reduction over 2003–2023) would require losses of approximately 11.3 % across the hydrogen supply chain. This lies at the upper end of literature estimates (<inline-formula><mml:math id="M238" display="inline"><mml:mrow><mml:mo>∼</mml:mo><mml:mn mathvariant="normal">10</mml:mn><mml:mspace linebreak="nobreak" width="0.125em"/><mml:mi mathvariant="italic">%</mml:mi></mml:mrow></mml:math></inline-formula>) and well above current best estimates (1.3 %; <xref ref-type="bibr" rid="bib1.bibx70" id="altparen.96"/>), suggesting that industrial emissions may contribute but are unlikely to fully explain our inferred trend.</p>
      <p id="d2e4200">A possible misattribution of sources and sinks of <inline-formula><mml:math id="M239" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> could be evaluated by propagating our <inline-formula><mml:math id="M240" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> fluxes into simulated hydrogen deuteride (<inline-formula><mml:math id="M241" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">HD</mml:mi></mml:mrow></mml:math></inline-formula>) abundances, following an approach similar to <xref ref-type="bibr" rid="bib1.bibx45" id="text.97"/>. While the photochemical source has limited isotopic leverage, the soil sink enriches the atmosphere in HD <xref ref-type="bibr" rid="bib1.bibx8" id="paren.98"/>. We therefore expect that a decreasing soil sink would contribute to declining global HD levels over time. In the case of a strong increase in industrial or geological emissions, we also expect declining HD levels, since these sources are strongly isotopically depleted <xref ref-type="bibr" rid="bib1.bibx18" id="paren.99"/>. However, no <inline-formula><mml:math id="M242" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">HD</mml:mi></mml:mrow></mml:math></inline-formula> observations have been published over the past 15 years, meaning this cannot be tested. In addition, other sources and sinks with greater isotopic fractionation have likely evolved in recent decades, as well as their corresponding source signatures. For example, advances in engine technology have likely altered both the <inline-formula><mml:math id="M243" display="inline"><mml:mrow><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow><mml:mspace width="0.125em" linebreak="nobreak"/><mml:mo>:</mml:mo><mml:mspace width="0.125em" linebreak="nobreak"/><mml:mrow class="chem"><mml:mi mathvariant="normal">CO</mml:mi></mml:mrow></mml:mrow></mml:math></inline-formula> emission ratio used to infer fossil fuel burning <inline-formula><mml:math id="M244" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> emissions from <inline-formula><mml:math id="M245" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">CO</mml:mi></mml:mrow></mml:math></inline-formula> emissions and the isotopic signature of the resulting <inline-formula><mml:math id="M246" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> <xref ref-type="bibr" rid="bib1.bibx73" id="paren.100"/>; however, such changes remain poorly characterised.</p>
      <p id="d2e4304">Our inverse problem is further complicated by the opposing signs and substantial spatial overlap of the soil sink and photochemical source (particularly the isoprene-related partition), both of which peak in tropical regions, leading to strong posterior correlations between these flux components (see Appendix <xref ref-type="sec" rid="App1.Ch1.S5"/>). Flux separability is further complicated by the low observational density in these regions. As recently noted by <xref ref-type="bibr" rid="bib1.bibx66" id="text.101"/>, local process-based measurements are crucial for improving understanding of tropical soil <inline-formula><mml:math id="M247" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> fluxes. We further emphasise the need for expanded atmospheric mole fraction observations to constrain future inverse modelling studies, as these regions remain among the least observed globally. A challenge specific to the photochemical source is its three-dimensional nature, making it more difficult to constrain using surface observations alone <xref ref-type="bibr" rid="bib1.bibx7" id="paren.102"/>. Despite this, we consider multi-species inversions a promising approach to better quantify this term, given the strong chemical coupling between <inline-formula><mml:math id="M248" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>, <inline-formula><mml:math id="M249" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula>, <inline-formula><mml:math id="M250" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">CO</mml:mi></mml:mrow></mml:math></inline-formula>, and <inline-formula><mml:math id="M251" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">OH</mml:mi></mml:mrow></mml:math></inline-formula>. <xref ref-type="bibr" rid="bib1.bibx48" id="text.103"/> applied such a framework, however we identify a clear opportunity to revisit it in light of nearly two decades of satellite observations of <inline-formula><mml:math id="M252" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">CO</mml:mi></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M253" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula> since then.</p>
</sec>
<sec id="Ch1.S4.SS2">
  <label>4.2</label><title>Causes of an overestimated prior photochemical <inline-formula><mml:math id="M254" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source</title>
      <p id="d2e4411">We found a significantly lower posterior photochemical <inline-formula><mml:math id="M255" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source compared to the prior derived from the full-chemistry simulation. This reduction occurred mainly in the pre-optimisation with global scaling factors (Sect. <xref ref-type="sec" rid="Ch1.S3.SS1"/>), but also in the subsequent detailed inversions (Sect. <xref ref-type="sec" rid="Ch1.S3.SS2"/>). Comparison with ATom aircraft observations (Appendix A) shows that the inversions have likely compensated for an overestimation of the <inline-formula><mml:math id="M256" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula> photolysis rate constant through the molecular channel (i.e., the <inline-formula><mml:math id="M257" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>-producing pathway), with the magnitude of this overestimation increasing towards the surface. In addition, the inversions likely corrected for an overestimation of <inline-formula><mml:math id="M258" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula> mole fractions over land, particularly in the lower atmosphere. The overestimation of the molecular-channel photolysis rate constant is possibly related to its 30 % increase in the updated CBM4 (Carbon Bond v4) and subsequent CB05 mechanism relative to the original CBM4, introduced to improve simulated <inline-formula><mml:math id="M259" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">CO</mml:mi></mml:mrow></mml:math></inline-formula>, <inline-formula><mml:math id="M260" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">4</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>, and <inline-formula><mml:math id="M261" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">O</mml:mi><mml:mn mathvariant="normal">3</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> <xref ref-type="bibr" rid="bib1.bibx75" id="paren.104"/>. However, the impact of this enhancement of the <inline-formula><mml:math id="M262" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>-producing photolysis channel on simulated <inline-formula><mml:math id="M263" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> was not evaluated in this study.</p>
</sec>
<sec id="Ch1.S4.SS3">
  <label>4.3</label><title>Implications of ENSO-driven soil uptake variability for other trace gases</title>
      <p id="d2e4531">Our inferred <inline-formula><mml:math id="M264" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> soil sink sensitivity to ENSO likely extends to other trace gases such as <inline-formula><mml:math id="M265" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">4</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>, <inline-formula><mml:math id="M266" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">CO</mml:mi></mml:mrow></mml:math></inline-formula>, <inline-formula><mml:math id="M267" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">N</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M268" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">COS</mml:mi></mml:mrow></mml:math></inline-formula>, since their uptake efficiency also depends on gas diffusion through the soil pore network. When soils become water-saturated, diffusion is reduced as pores fill with water, limiting access to microbial or chemical sinks. Field measurements support this theory, as <xref ref-type="bibr" rid="bib1.bibx2" id="text.105"/> conducted soil chamber measurements in a tropical forest in Costa Rica and showed that ENSO cycles influence biogenic <inline-formula><mml:math id="M269" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">4</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> fluxes through soil moisture changes, with <inline-formula><mml:math id="M270" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">4</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> consumption highest during the driest phase of El Niño and reduced uptake during La Niña and wetter conditions. These results align with our findings and support an ENSO-driven modulation of the <inline-formula><mml:math id="M271" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> soil sink in tropical regions, with similar moisture-driven mechanisms proposed.</p>
</sec>
</sec>
<sec id="Ch1.S5" sec-type="conclusions">
  <label>5</label><title>Conclusions</title>
      <p id="d2e4632">In this study, we performed atmospheric inversions of <inline-formula><mml:math id="M272" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> using the TM5 chemical transport model to derive a spatiotemporally resolved budget over the period 2003–2023. Using this approach, we infer a global mean <inline-formula><mml:math id="M273" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> soil sink of 52.8 [47.8–56.7] <inline-formula><mml:math id="M274" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> and a global photochemical <inline-formula><mml:math id="M275" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source of 34.6 [29.2–38.2] <inline-formula><mml:math id="M276" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> over 2003–2023. Notably, our soil sink posterior error range is substantially narrower than the one standard deviation range reported by <xref ref-type="bibr" rid="bib1.bibx41" id="text.106"/>. We also infer a soil sink that is 45 % and 35 % weaker in the Middle East and Oceania and 45 % and 70 % stronger in South America and Russia, respectively, relative to <xref ref-type="bibr" rid="bib1.bibx41" id="text.107"/>. Our results further suggest that variability in atmospheric <inline-formula><mml:math id="M277" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> growth rates between 2003 and 2023 was primarily driven by changes in the non-isoprene photochemical source, consistent with the recent hypothesis of <xref ref-type="bibr" rid="bib1.bibx41" id="text.108"/> that <inline-formula><mml:math id="M278" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">4</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>-driven photochemical production is a dominant control on the atmospheric <inline-formula><mml:math id="M279" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> growth rate. Additionally, our results show a weakening of the global soil sink at a mean rate of 0.23 <inline-formula><mml:math id="M280" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> over this period. We recommend further investigation into this, given the potential implications for future atmospheric <inline-formula><mml:math id="M281" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> levels in light of the projected increase in anthropogenic emissions during the ongoing energy transition. Finally, our results show a clear sensitivity of the <inline-formula><mml:math id="M282" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> soil sink to the El Niño–Southern Oscillation, strongest over diffusion-limited soils in tropical South America, with increased uptake during drier El Niño conditions and reduced uptake during wetter La Niña conditions. This sensitivity may extend to other tracers, such as <inline-formula><mml:math id="M283" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">4</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>, <inline-formula><mml:math id="M284" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">CO</mml:mi></mml:mrow></mml:math></inline-formula>, <inline-formula><mml:math id="M285" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">N</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula>, and <inline-formula><mml:math id="M286" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">COS</mml:mi></mml:mrow></mml:math></inline-formula>, given that their uptake is also dependent on diffusion through the soil pore network.</p>
      <p id="d2e4826">To further improve quantification of the global hydrogen budget, progress is needed along two complementary directions: improved process understanding to refine prior flux estimates and expanded observational constraints for future inverse modelling studies. For process understanding, additional field measurements of the soil sink across a broad range of environments are key, with particular emphasis on tropical regions, as noted by <xref ref-type="bibr" rid="bib1.bibx66" id="text.109"/>. For observational constraints, expanding <inline-formula><mml:math id="M287" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> mole fraction measurements in the tropics should be a priority, given the large fluxes and sparse observational coverage in these regions. Measurements and simulations of <inline-formula><mml:math id="M288" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">HD</mml:mi></mml:mrow></mml:math></inline-formula> may provide additional constraints on the global <inline-formula><mml:math id="M289" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> budget, as demonstrated by <xref ref-type="bibr" rid="bib1.bibx47" id="text.110"/>. However, the potential of <inline-formula><mml:math id="M290" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">HD</mml:mi></mml:mrow></mml:math></inline-formula> is limited by the absence of observations over the past 15 years, as well as limited progress in constraining source signatures and fractionation factors. Finally, for the photochemical <inline-formula><mml:math id="M291" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source, we consider multi-species inversions a promising approach given the strong chemical coupling between <inline-formula><mml:math id="M292" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>, <inline-formula><mml:math id="M293" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula>, <inline-formula><mml:math id="M294" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">CO</mml:mi></mml:mrow></mml:math></inline-formula>, and <inline-formula><mml:math id="M295" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">OH</mml:mi></mml:mrow></mml:math></inline-formula>. <xref ref-type="bibr" rid="bib1.bibx48" id="text.111"/> applied such a framework, and we identify a clear opportunity to revisit it in light of nearly two decades of satellite observations of <inline-formula><mml:math id="M296" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">CO</mml:mi></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M297" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula> since then.</p>
</sec>

      
      </body>
    <back><app-group>

<app id="App1.Ch1.S1">
  <label>Appendix A</label><title>Evaluation of simulated formaldehyde levels and the photolysis rate constant with aircraft observations</title>
      <p id="d2e4961">To assess whether our prior photochemical <inline-formula><mml:math id="M298" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source fields provide a reasonable starting point for the inversions, we evaluate simulated <inline-formula><mml:math id="M299" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula> mole fractions and the photolysis rate constant (<inline-formula><mml:math id="M300" display="inline"><mml:mrow><mml:msub><mml:mi>J</mml:mi><mml:mrow><mml:msub><mml:mi mathvariant="normal">a</mml:mi><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:msub></mml:mrow></mml:msub></mml:mrow></mml:math></inline-formula>) against aircraft observations from the ATom campaign <xref ref-type="bibr" rid="bib1.bibx21 bib1.bibx22 bib1.bibx78" id="paren.112"/>. The latter refers to the molecular photolysis channel of <inline-formula><mml:math id="M301" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula>, which is the <inline-formula><mml:math id="M302" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>-producing pathway.</p>
      <p id="d2e5037">Figure <xref ref-type="fig" rid="FA1"/> shows that our simulations adequately capture most of the observed large-scale variability in <inline-formula><mml:math id="M303" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula>, including enhanced mole fractions over polluted continental regions. The simulations reproduce the near-surface land-ocean gradient of approximately 1 ppb over North America and the increase toward the surface due to enhanced precursor emissions. Across all flights (<inline-formula><mml:math id="M304" display="inline"><mml:mrow><mml:mi>N</mml:mi><mml:mo>=</mml:mo><mml:mn mathvariant="normal">41</mml:mn></mml:mrow></mml:math></inline-formula>), the simulated and observed latitudinal and vertical gradients agree relatively well (Fig. <xref ref-type="fig" rid="FA2"/>), although our simulations show a general negative bias. The mean <inline-formula><mml:math id="M305" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula> mole fraction across all ATom observations was 0.138 ppb, compared to 0.090 ppb in our simulations, yielding a mean bias of <inline-formula><mml:math id="M306" display="inline"><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">0.048</mml:mn></mml:mrow></mml:math></inline-formula> ppb (<inline-formula><mml:math id="M307" display="inline"><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">34.5</mml:mn></mml:mrow></mml:math></inline-formula> %; Table <xref ref-type="table" rid="TA1"/>). For the six flights with mean observed <inline-formula><mml:math id="M308" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula> mole fractions above 0.2 ppb, however, the bias is reduced to <inline-formula><mml:math id="M309" display="inline"><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">0.018</mml:mn></mml:mrow></mml:math></inline-formula> ppb (<inline-formula><mml:math id="M310" display="inline"><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">7.6</mml:mn><mml:mspace linebreak="nobreak" width="0.125em"/><mml:mi mathvariant="italic">%</mml:mi></mml:mrow></mml:math></inline-formula>). Five of these flights sampled tropical ocean regions, where enhanced <inline-formula><mml:math id="M311" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">4</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> oxidation likely contributed to elevated <inline-formula><mml:math id="M312" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula> levels. The sixth flight sampled continental air over the United States, where greater <inline-formula><mml:math id="M313" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">OH</mml:mi></mml:mrow></mml:math></inline-formula> availability, together with anthropogenic and biogenic VOC emissions, likely contributed to the elevated <inline-formula><mml:math id="M314" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula> levels.</p>
      <p id="d2e5187">Figure <xref ref-type="fig" rid="FA3"/> shows that our hourly simulated <inline-formula><mml:math id="M315" display="inline"><mml:mrow><mml:msub><mml:mi>J</mml:mi><mml:mrow><mml:msub><mml:mi mathvariant="normal">a</mml:mi><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:msub></mml:mrow></mml:msub></mml:mrow></mml:math></inline-formula> values display a mean bias of <inline-formula><mml:math id="M316" display="inline"><mml:mrow><mml:mo>+</mml:mo><mml:mn mathvariant="normal">7.4</mml:mn><mml:mspace linebreak="nobreak" width="0.125em"/><mml:mi mathvariant="italic">%</mml:mi></mml:mrow></mml:math></inline-formula> across all altitudes relative to the ATom observations. However, this bias increases towards the surface, with values of <inline-formula><mml:math id="M317" display="inline"><mml:mrow><mml:mo>+</mml:mo><mml:mn mathvariant="normal">23.5</mml:mn><mml:mspace width="0.125em" linebreak="nobreak"/><mml:mi mathvariant="italic">%</mml:mi></mml:mrow></mml:math></inline-formula> between 0–2 km and <inline-formula><mml:math id="M318" display="inline"><mml:mrow><mml:mo>+</mml:mo><mml:mn mathvariant="normal">17.1</mml:mn><mml:mspace linebreak="nobreak" width="0.125em"/><mml:mi mathvariant="italic">%</mml:mi></mml:mrow></mml:math></inline-formula> between 2–4 km altitude. Besides this vertical structure, the bias is relatively uniform across latitudes, except between 60 and 90<inline-formula><mml:math id="M319" display="inline"><mml:mi mathvariant="italic">°</mml:mi></mml:math></inline-formula> S, although this region contains few ATom observations and is relatively unimportant for <inline-formula><mml:math id="M320" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> production given that both <inline-formula><mml:math id="M321" display="inline"><mml:mrow><mml:msub><mml:mi>J</mml:mi><mml:mrow><mml:msub><mml:mi mathvariant="normal">a</mml:mi><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:msub></mml:mrow></mml:msub></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M322" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula> mole fractions are low here. Based on the identified vertical bias structure, we calculated correction factors for <inline-formula><mml:math id="M323" display="inline"><mml:mrow><mml:msub><mml:mi>J</mml:mi><mml:mrow><mml:msub><mml:mi mathvariant="normal">a</mml:mi><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:msub></mml:mrow></mml:msub></mml:mrow></mml:math></inline-formula> for each model vertical layer and reran the full-chemistry simulation for the 2016–2018 period. As expected, the lower <inline-formula><mml:math id="M324" display="inline"><mml:mrow><mml:msub><mml:mi>J</mml:mi><mml:mrow><mml:msub><mml:mi mathvariant="normal">a</mml:mi><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:msub></mml:mrow></mml:msub></mml:mrow></mml:math></inline-formula> values resulted in slightly larger simulated <inline-formula><mml:math id="M325" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula> mole fractions (Figs. <xref ref-type="fig" rid="FA1"/> and <xref ref-type="fig" rid="FA2"/>). Applying this correction reduced the annual mean photochemical <inline-formula><mml:math id="M326" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source from 55.0 to 50.3 <inline-formula><mml:math id="M327" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> in 2016–2017, still substantially above our posterior estimate of 36.0 <inline-formula><mml:math id="M328" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> obtained after the global scaling in Sect. <xref ref-type="sec" rid="Ch1.S3.SS1"/> and subsequent detailed inversions in Sect. <xref ref-type="sec" rid="Ch1.S3.SS2"/>.</p>
      <p id="d2e5415">We hypothesize that the remaining difference is related to an overestimation of <inline-formula><mml:math id="M329" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula> in regions important for photochemical <inline-formula><mml:math id="M330" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> production. Evidence for such a bias is provided by the flight over the continental United States (Fig. <xref ref-type="fig" rid="FA1"/>), where simulated <inline-formula><mml:math id="M331" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula> exceeds observations in the lower troposphere, where both <inline-formula><mml:math id="M332" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula> concentrations and photochemical <inline-formula><mml:math id="M333" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> production are high. This regional bias is not apparent in the overall ATom mean because ATom predominantly sampled remote oceanic and background conditions, where <inline-formula><mml:math id="M334" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula> concentrations and photochemical <inline-formula><mml:math id="M335" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> production are lower. The ATom mean therefore does not necessarily represent <inline-formula><mml:math id="M336" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula> biases in regions that contribute most strongly to the global photochemical <inline-formula><mml:math id="M337" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source. While the <inline-formula><mml:math id="M338" display="inline"><mml:mrow><mml:msub><mml:mi>J</mml:mi><mml:mrow><mml:msub><mml:mi mathvariant="normal">a</mml:mi><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:msub></mml:mrow></mml:msub></mml:mrow></mml:math></inline-formula> bias can be directly corrected using the observational constraint presented here, addressing the <inline-formula><mml:math id="M339" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula> bias would require changes to precursor emissions or the chemical mechanism and is beyond the scope of this study.</p>
      <p id="d2e5566">Since the <inline-formula><mml:math id="M340" display="inline"><mml:mrow><mml:msub><mml:mi>J</mml:mi><mml:mrow><mml:msub><mml:mi mathvariant="normal">a</mml:mi><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:msub></mml:mrow></mml:msub></mml:mrow></mml:math></inline-formula> correction reduced the prior photochemical <inline-formula><mml:math id="M341" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source by approximately 3 <inline-formula><mml:math id="M342" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> after applying the global scaling factor of 0.7 from Sect. <xref ref-type="sec" rid="Ch1.S3.SS1"/>, we assessed its impact on posterior fluxes. We performed two inversions for 2016–2018 using photochemical <inline-formula><mml:math id="M343" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source fields with and without the correction. The uncorrected case used the original source fields from the full inversions but was rerun over 2016–2018 to ensure a consistent comparison, excluding observations outside this period that influenced the original inversion through the temporal correlations described in Appendix <xref ref-type="sec" rid="App1.Ch1.S3.SS1"/>. For 2016–2017, applying the correction changed the posterior fluxes from 50.0 to 48.3 <inline-formula><mml:math id="M344" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> for the soil sink, from 27.8 to 26.7 <inline-formula><mml:math id="M345" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> for the non-isoprene <inline-formula><mml:math id="M346" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source, and from 8.2 to 7.5 <inline-formula><mml:math id="M347" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> for the isoprene <inline-formula><mml:math id="M348" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source. Although these changes are not negligible, they remain well within the uncertainty ranges established in Appendix <xref ref-type="sec" rid="App1.Ch1.S4"/>. Therefore, we do not deem it necessary to rerun all inversions with the bias-corrected photochemical sources. </p>

      <fig id="FA1"><label>Figure A1</label><caption><p id="d2e5713">Observed and simulated <inline-formula><mml:math id="M349" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula> mole fractions for three ATom flights: <bold>(a)</bold> 160 822 (mean bias: base run <inline-formula><mml:math id="M350" display="inline"><mml:mrow><mml:mo>=</mml:mo><mml:mn mathvariant="normal">36.2</mml:mn><mml:mspace width="0.125em" linebreak="nobreak"/><mml:mi mathvariant="italic">%</mml:mi></mml:mrow></mml:math></inline-formula>, bias-corrected run <inline-formula><mml:math id="M351" display="inline"><mml:mrow><mml:mo>=</mml:mo><mml:mn mathvariant="normal">47.3</mml:mn><mml:mspace width="0.125em" linebreak="nobreak"/><mml:mi mathvariant="italic">%</mml:mi></mml:mrow></mml:math></inline-formula>), <bold>(b)</bold> 160 817 (mean bias: base run <inline-formula><mml:math id="M352" display="inline"><mml:mrow><mml:mo>=</mml:mo><mml:mo>-</mml:mo><mml:mn mathvariant="normal">5.3</mml:mn><mml:mspace linebreak="nobreak" width="0.125em"/><mml:mi mathvariant="italic">%</mml:mi></mml:mrow></mml:math></inline-formula>, bias-corrected run <inline-formula><mml:math id="M353" display="inline"><mml:mrow><mml:mo>=</mml:mo><mml:mn mathvariant="normal">0.3</mml:mn><mml:mspace linebreak="nobreak" width="0.125em"/><mml:mi mathvariant="italic">%</mml:mi></mml:mrow></mml:math></inline-formula>), and <bold>(c)</bold> 160 823 (mean bias: base run <inline-formula><mml:math id="M354" display="inline"><mml:mrow><mml:mo>=</mml:mo><mml:mn mathvariant="normal">44.5</mml:mn><mml:mspace linebreak="nobreak" width="0.125em"/><mml:mi mathvariant="italic">%</mml:mi></mml:mrow></mml:math></inline-formula>, bias-corrected run <inline-formula><mml:math id="M355" display="inline"><mml:mrow><mml:mo>=</mml:mo><mml:mn mathvariant="normal">55.2</mml:mn><mml:mspace linebreak="nobreak" width="0.125em"/><mml:mi mathvariant="italic">%</mml:mi></mml:mrow></mml:math></inline-formula>). Left axis: <inline-formula><mml:math id="M356" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula> mole fraction; right axis: altitude along flight track.</p></caption>
        
        <graphic xlink:href="https://acp.copernicus.org/articles/26/16261/2026/acp-26-16261-2026-f09.png"/>

      </fig>

<fig id="FA2"><label>Figure A2</label><caption><p id="d2e5843">Latitudinal and vertical gradients in observed and simulated <inline-formula><mml:math id="M357" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula> mole fractions for the ATom flights. Results are shown for simulations with and without <inline-formula><mml:math id="M358" display="inline"><mml:mrow><mml:msub><mml:mi>J</mml:mi><mml:mrow><mml:msub><mml:mi mathvariant="normal">a</mml:mi><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:msub></mml:mrow></mml:msub></mml:mrow></mml:math></inline-formula> bias correction.</p></caption>
        
        <graphic xlink:href="https://acp.copernicus.org/articles/26/16261/2026/acp-26-16261-2026-f10.png"/>

      </fig>

      <fig id="FA3"><label>Figure A3</label><caption><p id="d2e5890">Vertical and latitudinal gradient in simulated bias in the photolysis rate constant for the photolysis of <inline-formula><mml:math id="M359" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula> through the molecular channel (<inline-formula><mml:math id="M360" display="inline"><mml:mrow><mml:msub><mml:mi>J</mml:mi><mml:mrow><mml:msub><mml:mi mathvariant="normal">a</mml:mi><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:msub></mml:mrow></mml:msub></mml:mrow></mml:math></inline-formula>).</p></caption>
        
        <graphic xlink:href="https://acp.copernicus.org/articles/26/16261/2026/acp-26-16261-2026-f11.png"/>

      </fig>

<table-wrap id="TA1"><label>Table A1</label><caption><p id="d2e5940">Summary statistics of observed and simulated <inline-formula><mml:math id="M361" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula> mole fractions from ATom flights for all flights and subsets with increasing minimum observed <inline-formula><mml:math id="M362" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub><mml:mi mathvariant="normal">O</mml:mi></mml:mrow></mml:math></inline-formula>. The table compares observations with the baserun and bias-corrected simulations, reporting the mean simulated mole fractions and the corresponding mean biases relative to observations.</p></caption><oasis:table frame="topbot"><oasis:tgroup cols="7">
     <oasis:colspec colnum="1" colname="col1" align="left"/>
     <oasis:colspec colnum="2" colname="col2" align="right"/>
     <oasis:colspec colnum="3" colname="col3" align="center"/>
     <oasis:colspec colnum="4" colname="col4" align="center"/>
     <oasis:colspec colnum="5" colname="col5" align="center"/>
     <oasis:colspec colnum="6" colname="col6" align="right"/>
     <oasis:colspec colnum="7" colname="col7" align="right"/>
     <oasis:thead>
       <oasis:row>
         <oasis:entry colname="col1">Flights</oasis:entry>
         <oasis:entry colname="col2"><inline-formula><mml:math id="M363" display="inline"><mml:mi>N</mml:mi></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col3">Mean obs.</oasis:entry>
         <oasis:entry colname="col4">Mean sim.</oasis:entry>
         <oasis:entry colname="col5">Mean sim.</oasis:entry>
         <oasis:entry colname="col6">Mean bias</oasis:entry>
         <oasis:entry colname="col7">Mean bias</oasis:entry>
       </oasis:row>
       <oasis:row rowsep="1">
         <oasis:entry colname="col1"/>
         <oasis:entry colname="col2"/>
         <oasis:entry colname="col3">[ppb]</oasis:entry>
         <oasis:entry colname="col4">(base) [ppb]</oasis:entry>
         <oasis:entry colname="col5">(bias-corr.) [ppb]</oasis:entry>
         <oasis:entry colname="col6">(base) [%]</oasis:entry>
         <oasis:entry colname="col7">(bias-corr.) [%]</oasis:entry>
       </oasis:row>
     </oasis:thead>
     <oasis:tbody>
       <oasis:row>
         <oasis:entry colname="col1">All</oasis:entry>
         <oasis:entry colname="col2">41</oasis:entry>
         <oasis:entry colname="col3">0.138</oasis:entry>
         <oasis:entry colname="col4">0.090</oasis:entry>
         <oasis:entry colname="col5">0.096</oasis:entry>
         <oasis:entry colname="col6"><inline-formula><mml:math id="M364" display="inline"><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">34.47</mml:mn></mml:mrow></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col7"><inline-formula><mml:math id="M365" display="inline"><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">30.29</mml:mn></mml:mrow></mml:math></inline-formula></oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1"><inline-formula><mml:math id="M366" display="inline"><mml:mrow><mml:mo>&gt;</mml:mo><mml:mn mathvariant="normal">0.05</mml:mn></mml:mrow></mml:math></inline-formula> ppb</oasis:entry>
         <oasis:entry colname="col2">40</oasis:entry>
         <oasis:entry colname="col3">0.141</oasis:entry>
         <oasis:entry colname="col4">0.093</oasis:entry>
         <oasis:entry colname="col5">0.099</oasis:entry>
         <oasis:entry colname="col6"><inline-formula><mml:math id="M367" display="inline"><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">34.06</mml:mn></mml:mrow></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col7"><inline-formula><mml:math id="M368" display="inline"><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">29.86</mml:mn></mml:mrow></mml:math></inline-formula></oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1"><inline-formula><mml:math id="M369" display="inline"><mml:mrow><mml:mo>&gt;</mml:mo><mml:mn mathvariant="normal">0.1</mml:mn></mml:mrow></mml:math></inline-formula> ppb</oasis:entry>
         <oasis:entry colname="col2">30</oasis:entry>
         <oasis:entry colname="col3">0.160</oasis:entry>
         <oasis:entry colname="col4">0.114</oasis:entry>
         <oasis:entry colname="col5">0.121</oasis:entry>
         <oasis:entry colname="col6"><inline-formula><mml:math id="M370" display="inline"><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">29.20</mml:mn></mml:mrow></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col7"><inline-formula><mml:math id="M371" display="inline"><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">24.65</mml:mn></mml:mrow></mml:math></inline-formula></oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1"><inline-formula><mml:math id="M372" display="inline"><mml:mrow><mml:mo>&gt;</mml:mo><mml:mn mathvariant="normal">0.15</mml:mn></mml:mrow></mml:math></inline-formula> ppb</oasis:entry>
         <oasis:entry colname="col2">15</oasis:entry>
         <oasis:entry colname="col3">0.196</oasis:entry>
         <oasis:entry colname="col4">0.154</oasis:entry>
         <oasis:entry colname="col5">0.164</oasis:entry>
         <oasis:entry colname="col6"><inline-formula><mml:math id="M373" display="inline"><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">21.47</mml:mn></mml:mrow></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col7"><inline-formula><mml:math id="M374" display="inline"><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">16.15</mml:mn></mml:mrow></mml:math></inline-formula></oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1"><inline-formula><mml:math id="M375" display="inline"><mml:mrow><mml:mo>&gt;</mml:mo><mml:mn mathvariant="normal">0.2</mml:mn></mml:mrow></mml:math></inline-formula> ppb</oasis:entry>
         <oasis:entry colname="col2">6</oasis:entry>
         <oasis:entry colname="col3">0.247</oasis:entry>
         <oasis:entry colname="col4">0.229</oasis:entry>
         <oasis:entry colname="col5">0.244</oasis:entry>
         <oasis:entry colname="col6"><inline-formula><mml:math id="M376" display="inline"><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">7.63</mml:mn></mml:mrow></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col7"><inline-formula><mml:math id="M377" display="inline"><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1.39</mml:mn></mml:mrow></mml:math></inline-formula></oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1"><inline-formula><mml:math id="M378" display="inline"><mml:mrow><mml:mo>&gt;</mml:mo><mml:mn mathvariant="normal">0.25</mml:mn></mml:mrow></mml:math></inline-formula> ppb</oasis:entry>
         <oasis:entry colname="col2">2</oasis:entry>
         <oasis:entry colname="col3">0.347</oasis:entry>
         <oasis:entry colname="col4">0.418</oasis:entry>
         <oasis:entry colname="col5">0.449</oasis:entry>
         <oasis:entry colname="col6">20.33</oasis:entry>
         <oasis:entry colname="col7">29.23</oasis:entry>
       </oasis:row>
     </oasis:tbody>
   </oasis:tgroup></oasis:table></table-wrap>


</app>

<app id="App1.Ch1.S2">
  <label>Appendix B</label><title>Comparison of <inline-formula><mml:math id="M379" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> sink seasonality with previous studies</title>
      <p id="d2e6346">We compare our inferred <inline-formula><mml:math id="M380" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> sink seasonality with the independent estimates of <xref ref-type="bibr" rid="bib1.bibx63" id="text.113"/> and <xref ref-type="bibr" rid="bib1.bibx66" id="text.114"/>. For this, we calculated the seasonal range and the timing of maximum uptake of our simulated soil sink using the NOAA CCGCRV curve-fitting package <xref ref-type="bibr" rid="bib1.bibx68 bib1.bibx38" id="paren.115"/>.</p>
      <p id="d2e6369"><xref ref-type="bibr" rid="bib1.bibx63" id="text.116"/> used HFC-152a observations in a Bayesian three-box inversion to jointly constrain <inline-formula><mml:math id="M381" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">OH</mml:mi></mml:mrow></mml:math></inline-formula> seasonality and infer the <inline-formula><mml:math id="M382" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> soil-sink cycle for the Northern Hemisphere (90–20<inline-formula><mml:math id="M383" display="inline"><mml:mi mathvariant="italic">°</mml:mi></mml:math></inline-formula> N), tropics (20<inline-formula><mml:math id="M384" display="inline"><mml:mi mathvariant="italic">°</mml:mi></mml:math></inline-formula> N–20<inline-formula><mml:math id="M385" display="inline"><mml:mi mathvariant="italic">°</mml:mi></mml:math></inline-formula> S), and Southern Hemisphere (20–90<inline-formula><mml:math id="M386" display="inline"><mml:mi mathvariant="italic">°</mml:mi></mml:math></inline-formula> S). We find excellent agreement in seasonal range and timing, particularly in the Northern Hemisphere (Table <xref ref-type="table" rid="TB1"/>). Here, our posterior seasonal ranges of 21.0 and 20.0 <inline-formula><mml:math id="M387" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> for Sanderson and Bertagni closely match their estimate of 18–21 <inline-formula><mml:math id="M388" display="inline"><mml:mo>±</mml:mo></mml:math></inline-formula> 8 Tg yr<sup>−1</sup>, and our posterior soil sink peaks in July and August, respectively, coinciding with the July–August peak reported by <xref ref-type="bibr" rid="bib1.bibx63" id="text.117"/>. In the tropics, our Sanderson posterior seasonal range of 9.0 <inline-formula><mml:math id="M390" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> agrees very well with their estimate of 8 <inline-formula><mml:math id="M391" display="inline"><mml:mo>±</mml:mo></mml:math></inline-formula> 8 Tg yr<sup>−1</sup>, while the Bertagni posterior of 4.5 <inline-formula><mml:math id="M393" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> still falls well within their reported uncertainty range. <xref ref-type="bibr" rid="bib1.bibx63" id="text.118"/> find substantial variation in the timing of maximum uptake across their four experiments, with peaks in January, April, and August. Our inferred maxima in August and October are in good agreement with two of their four experiments, which showed an August maximum. In the Southern Hemisphere, our posterior seasonal ranges of 1.4 and 1.8 <inline-formula><mml:math id="M394" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> both fall within their reported range of 2–3 <inline-formula><mml:math id="M395" display="inline"><mml:mo>±</mml:mo></mml:math></inline-formula> 2.5 Tg yr<sup>−1</sup>. Interestingly, our Southern Hemisphere soil sink peaks in November, consistent with <xref ref-type="bibr" rid="bib1.bibx7" id="text.119"/> and <xref ref-type="bibr" rid="bib1.bibx80" id="text.120"/>, whereas <xref ref-type="bibr" rid="bib1.bibx63" id="text.121"/> find a January–March maximum, closer to the February peak reported by <xref ref-type="bibr" rid="bib1.bibx54" id="text.122"/>.</p>
      <p id="d2e6569"><xref ref-type="bibr" rid="bib1.bibx63" id="text.123"/> also constrained the seasonality of chemical loss by <inline-formula><mml:math id="M397" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">OH</mml:mi></mml:mrow></mml:math></inline-formula>. In the Northern Hemisphere, our posterior seasonal ranges of 8.8 and 8.9 Tg yr<sup>−1</sup> for Sanderson and Bertagni closely match their estimate of 8 <inline-formula><mml:math id="M399" display="inline"><mml:mo>±</mml:mo></mml:math></inline-formula> 1 Tg yr<sup>−1</sup>, and both inversions peak in July, consistent with their June–July maximum. In the tropics, our chemical loss term is relatively constant throughout the year, yielding a seasonal range of only 1.1 Tg yr<sup>−1</sup>, which falls outside their reported range of 4 <inline-formula><mml:math id="M402" display="inline"><mml:mo>±</mml:mo></mml:math></inline-formula> 2.5 Tg yr<sup>−1</sup>, with maximum loss occurring in September compared with their February–March maximum. This larger discrepancy in the tropics may reflect differences in the seasonality of photochemical production and other <inline-formula><mml:math id="M404" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> sources between the two approaches. In the Southern Hemisphere, both inversions yield a seasonal range of 4.9 Tg yr<sup>−1</sup>, lower than their estimate of 7 <inline-formula><mml:math id="M406" display="inline"><mml:mo>±</mml:mo></mml:math></inline-formula> 1 <inline-formula><mml:math id="M407" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>, while our January maximum is consistent with their December–January maximum.</p>
      <p id="d2e6693">Interestingly, our inferred soil-sink seasonality differs substantially from <xref ref-type="bibr" rid="bib1.bibx66" id="text.124"/>, who used a zonal-mean two-dimensional model with four vertical levels to optimise <inline-formula><mml:math id="M408" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> soil-sink seasonality in three latitude bands. In the 30<inline-formula><mml:math id="M409" display="inline"><mml:mi mathvariant="italic">°</mml:mi></mml:math></inline-formula> S–0<inline-formula><mml:math id="M410" display="inline"><mml:mi mathvariant="italic">°</mml:mi></mml:math></inline-formula> N band, our posterior seasonal ranges are 6.4 and 2.2 Tg yr<sup>−1</sup> for Sanderson and Bertagni, respectively, compared with their substantially larger estimate of 21.9 Tg yr<sup>−1</sup>. The timing of maximum uptake also differs, occurring in September and June in our inversions, respectively, compared with May in their study. Similarly, in the 0–30<inline-formula><mml:math id="M413" display="inline"><mml:mi mathvariant="italic">°</mml:mi></mml:math></inline-formula> N band, our posterior seasonal ranges of 5.9 and 1.7 Tg yr<sup>−1</sup> are lower than their estimate of 18.1 Tg yr<sup>−1</sup>, with maximum uptake occurring in March and December, respectively, compared with October. In the 30–60<inline-formula><mml:math id="M416" display="inline"><mml:mi mathvariant="italic">°</mml:mi></mml:math></inline-formula> N band, our seasonal ranges of 12.7 and 14.0 Tg yr<sup>−1</sup> are again lower than their estimate of 23.1 Tg yr<sup>−1</sup>, while our maxima occur in June and May, respectively, compared with September. Thus, compared with <xref ref-type="bibr" rid="bib1.bibx66" id="text.125"/>, our inversions consistently infer weaker seasonal variations and generally different timing of maximum uptake. We attribute this to the different methodological approaches used in the two studies: <xref ref-type="bibr" rid="bib1.bibx66" id="text.126"/> optimise only the soil sink using a zonal-mean transport framework, whereas we jointly optimise the soil sink and photochemical <inline-formula><mml:math id="M419" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source using a three-dimensional framework. Furthermore, <xref ref-type="bibr" rid="bib1.bibx66" id="text.127"/> used a smaller set of observations (i.e., only NOAA-GML data). </p>

<table-wrap id="TB1"><label>Table B1</label><caption><p id="d2e6837">Seasonality of the soil sink in our inversions compared with <xref ref-type="bibr" rid="bib1.bibx63" id="text.128"/>. Our fluxes are evaluated over 2010–2022 to match <xref ref-type="bibr" rid="bib1.bibx63" id="text.129"/>. We calculate the seasonal range and timing of maximum uptake using the NOAA CCGCRV curve-fitting package <xref ref-type="bibr" rid="bib1.bibx68 bib1.bibx38" id="paren.130"/>. Values show the seasonal range in <inline-formula><mml:math id="M420" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> followed by the month of maximum sink strength. <xref ref-type="bibr" rid="bib1.bibx63" id="text.131"/> provide only one seasonal range value for the tropics box (from their “<inline-formula><mml:math id="M421" display="inline"><mml:mrow class="chem"><mml:mi mathvariant="normal">OH</mml:mi></mml:mrow></mml:math></inline-formula> case”) and no month of maximum sink strength, since the latter varies between cases.</p></caption><oasis:table frame="topbot"><oasis:tgroup cols="4">
     <oasis:colspec colnum="1" colname="col1" align="left"/>
     <oasis:colspec colnum="2" colname="col2" align="left"/>
     <oasis:colspec colnum="3" colname="col3" align="left"/>
     <oasis:colspec colnum="4" colname="col4" align="left"/>
     <oasis:thead>
       <oasis:row rowsep="1">
         <oasis:entry colname="col1"/>
         <oasis:entry colname="col2">90–20<inline-formula><mml:math id="M422" display="inline"><mml:mi mathvariant="italic">°</mml:mi></mml:math></inline-formula> N</oasis:entry>
         <oasis:entry colname="col3">20<inline-formula><mml:math id="M423" display="inline"><mml:mi mathvariant="italic">°</mml:mi></mml:math></inline-formula> N–20<inline-formula><mml:math id="M424" display="inline"><mml:mi mathvariant="italic">°</mml:mi></mml:math></inline-formula> S</oasis:entry>
         <oasis:entry colname="col4">20–90<inline-formula><mml:math id="M425" display="inline"><mml:mi mathvariant="italic">°</mml:mi></mml:math></inline-formula> S</oasis:entry>
       </oasis:row>
     </oasis:thead>
     <oasis:tbody>
       <oasis:row>
         <oasis:entry colname="col1">Sanderson prior</oasis:entry>
         <oasis:entry colname="col2">17.0 (Jun)</oasis:entry>
         <oasis:entry colname="col3">2.1 (Aug)</oasis:entry>
         <oasis:entry colname="col4">1.3 (Dec)</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">Sanderson posterior</oasis:entry>
         <oasis:entry colname="col2">21.0 (Jul)</oasis:entry>
         <oasis:entry colname="col3">9.0 (Aug)</oasis:entry>
         <oasis:entry colname="col4">1.4 (Sep)</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">Bertagni prior</oasis:entry>
         <oasis:entry colname="col2">16.7 (Jun)</oasis:entry>
         <oasis:entry colname="col3">3.4 (Jun)</oasis:entry>
         <oasis:entry colname="col4">1.3 (Nov)</oasis:entry>
       </oasis:row>
       <oasis:row rowsep="1">
         <oasis:entry colname="col1">Bertagni posterior</oasis:entry>
         <oasis:entry colname="col2">20.0 (Aug)</oasis:entry>
         <oasis:entry colname="col3">4.5 (Oct)</oasis:entry>
         <oasis:entry colname="col4">1.8 (Nov)</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">
                  <xref ref-type="bibr" rid="bib1.bibx63" id="text.132"/>
                </oasis:entry>
         <oasis:entry colname="col2">18–21 <inline-formula><mml:math id="M426" display="inline"><mml:mo>±</mml:mo></mml:math></inline-formula> 8 (Jul–Aug)</oasis:entry>
         <oasis:entry colname="col3">8 <inline-formula><mml:math id="M427" display="inline"><mml:mo>±</mml:mo></mml:math></inline-formula> 8</oasis:entry>
         <oasis:entry colname="col4">2–3 <inline-formula><mml:math id="M428" display="inline"><mml:mo>±</mml:mo></mml:math></inline-formula> 2.5 (Jan–Mar)</oasis:entry>
       </oasis:row>
     </oasis:tbody>
   </oasis:tgroup></oasis:table></table-wrap>

</app>

<app id="App1.Ch1.S3">
  <label>Appendix C</label><title>Evaluation of inversion parameters</title>
      <p id="d2e7047">To determine the best-performing inversion settings, we conducted various experiments at a <inline-formula><mml:math id="M429" display="inline"><mml:mrow><mml:mn mathvariant="normal">6</mml:mn><mml:mi mathvariant="italic">°</mml:mi><mml:mo>×</mml:mo><mml:mn mathvariant="normal">4</mml:mn><mml:mi mathvariant="italic">°</mml:mi></mml:mrow></mml:math></inline-formula> horizontal resolution with 25 vertical levels, covering 2003–2013. We used the optimal prior Bertagni soil sink and photochemical source fluxes derived in Sect. <xref ref-type="sec" rid="Ch1.S2.SS4"/> and the settings described in Sect. <xref ref-type="sec" rid="Ch1.S2.SS5"/> unless stated otherwise. We use an ensemble size of 150 members, which is sufficient given our chosen uncertainties for each statevector element in space and time. However, our main results in Sect. <xref ref-type="sec" rid="Ch1.S3.SS2"/> are still obtained with 300 members to be conservative and account for potential sensitivity to random sampling, particularly since tests showed that larger uncertainties on the statevector elements increase this sensitivity.</p>
<sec id="App1.Ch1.S3.SS1">
  <label>C1</label><title>Temporal correlations</title>
      <p id="d2e7079">To establish robust temporal correlations, we first examined the effect of the <inline-formula><mml:math id="M430" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="italic">σ</mml:mi><mml:mi mathvariant="normal">MOY</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> parameter in our statevector (Table <xref ref-type="table" rid="TC1"/>). This parameter describes the extent to which observations from neighbouring years influence monthly scaling, and its use is based on the assumption that a large part of the error shows seasonal behaviour. In our first experiment (A1), we use a 60-month value for the non-isoprene <inline-formula><mml:math id="M431" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source, assuming that <inline-formula><mml:math id="M432" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">CH</mml:mi><mml:mn mathvariant="normal">4</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> mole fractions show low interannual variability. For the soil sink, we use a 36-month value, assuming that this term is more variable, but that this variability is already partly captured in our input datasets. Finally, we use a 24-month value for the isoprene <inline-formula><mml:math id="M433" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source, since its interannual variability is poorly understood and poorly constrained. Shorter timescales or uniform long timescales were also tested (Table <xref ref-type="table" rid="TC1"/>).</p>

<table-wrap id="TC1" specific-use="star"><label>Table C1</label><caption><p id="d2e7134">Impact of the <inline-formula><mml:math id="M434" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="italic">σ</mml:mi><mml:mi mathvariant="normal">MOY</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> parameter, which prescribes the decay time of correlations between the same months of adjacent years (J–J, F–F, M–M, ...), on the fraction of grid cells with optimised scaling factors that reverse the sign of the prior flux.</p></caption><oasis:table frame="topbot"><oasis:tgroup cols="7">
     <oasis:colspec colnum="1" colname="col1" align="left"/>
     <oasis:colspec colnum="2" colname="col2" align="center"/>
     <oasis:colspec colnum="3" colname="col3" align="center"/>
     <oasis:colspec colnum="4" colname="col4" align="center"/>
     <oasis:colspec colnum="5" colname="col5" align="center"/>
     <oasis:colspec colnum="6" colname="col6" align="center"/>
     <oasis:colspec colnum="7" colname="col7" align="center"/>
     <oasis:thead>
       <oasis:row rowsep="1">
         <oasis:entry colname="col1">Experiment</oasis:entry>
         <oasis:entry colname="col2"><inline-formula><mml:math id="M435" display="inline"><mml:mrow><mml:msub><mml:msub><mml:mi mathvariant="italic">σ</mml:mi><mml:mi mathvariant="normal">MOY</mml:mi></mml:msub><mml:mtext>soil sink</mml:mtext></mml:msub></mml:mrow></mml:math></inline-formula> [m]</oasis:entry>
         <oasis:entry colname="col3"><inline-formula><mml:math id="M436" display="inline"><mml:mrow><mml:msub><mml:msub><mml:mi mathvariant="italic">σ</mml:mi><mml:mi mathvariant="normal">MOY</mml:mi></mml:msub><mml:mtext>non-isoprene</mml:mtext></mml:msub></mml:mrow></mml:math></inline-formula> [m]</oasis:entry>
         <oasis:entry colname="col4"><inline-formula><mml:math id="M437" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="italic">σ</mml:mi><mml:mi mathvariant="normal">MOY</mml:mi></mml:msub><mml:mtext>isoprene</mml:mtext></mml:mrow></mml:math></inline-formula> [m]</oasis:entry>
         <oasis:entry colname="col5">Pos<sub>soil</sub> [%]</oasis:entry>
         <oasis:entry colname="col6">Neg<sub>isoprene</sub> [%]</oasis:entry>
         <oasis:entry colname="col7">Neg<sub>non-isoprene</sub> [%]</oasis:entry>
       </oasis:row>
     </oasis:thead>
     <oasis:tbody>
       <oasis:row>
         <oasis:entry colname="col1">A1</oasis:entry>
         <oasis:entry colname="col2">36</oasis:entry>
         <oasis:entry colname="col3">60</oasis:entry>
         <oasis:entry colname="col4">24</oasis:entry>
         <oasis:entry colname="col5">3.48</oasis:entry>
         <oasis:entry colname="col6">2.26</oasis:entry>
         <oasis:entry colname="col7">6.34</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">A2</oasis:entry>
         <oasis:entry colname="col2">24</oasis:entry>
         <oasis:entry colname="col3">36</oasis:entry>
         <oasis:entry colname="col4">12</oasis:entry>
         <oasis:entry colname="col5">2.85</oasis:entry>
         <oasis:entry colname="col6">1.00</oasis:entry>
         <oasis:entry colname="col7">2.41</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">A3</oasis:entry>
         <oasis:entry colname="col2">60</oasis:entry>
         <oasis:entry colname="col3">60</oasis:entry>
         <oasis:entry colname="col4">60</oasis:entry>
         <oasis:entry colname="col5">3.56</oasis:entry>
         <oasis:entry colname="col6">3.05</oasis:entry>
         <oasis:entry colname="col7">10.98</oasis:entry>
       </oasis:row>
     </oasis:tbody>
   </oasis:tgroup></oasis:table></table-wrap>

<table-wrap id="TC2"><label>Table C2</label><caption><p id="d2e7342">Impacts of progressively refined MDM classes for different site categories on posterior chi-squared (<inline-formula><mml:math id="M441" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="italic">χ</mml:mi><mml:mi mathvariant="normal">post</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msubsup></mml:mrow></mml:math></inline-formula>) at MAA (Mawson, Antarctica), MLO (Mauna Loa, Hawaii), MHD (Mace Head, Ireland), and NWR (Niwot Ridge, USA). These sites represent Deep Southern Hemisphere, Marine Boundary Layer, Mixed, and Land environments, respectively. Additionally, the fraction of grid cells with optimised scaling factors that reverse the prior flux sign is shown.</p></caption><oasis:table frame="topbot"><oasis:tgroup cols="5">
     <oasis:colspec colnum="1" colname="col1" align="left"/>
     <oasis:colspec colnum="2" colname="col2" align="center"/>
     <oasis:colspec colnum="3" colname="col3" align="center"/>
     <oasis:colspec colnum="4" colname="col4" align="center"/>
     <oasis:colspec colnum="5" colname="col5" align="center"/>
     <oasis:thead>
       <oasis:row rowsep="1">
         <oasis:entry colname="col1"/>
         <oasis:entry colname="col2">A4</oasis:entry>
         <oasis:entry colname="col3">A5<sup>a</sup></oasis:entry>
         <oasis:entry colname="col4">A6<sup>b</sup></oasis:entry>
         <oasis:entry colname="col5">A7<sup>c</sup></oasis:entry>
       </oasis:row>
     </oasis:thead>
     <oasis:tbody>
       <oasis:row rowsep="1">
         <oasis:entry namest="col1" nameend="col5">Median <inline-formula><mml:math id="M451" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="italic">χ</mml:mi><mml:mi mathvariant="normal">post</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msubsup></mml:mrow></mml:math></inline-formula></oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">MAA</oasis:entry>
         <oasis:entry colname="col2">0.56</oasis:entry>
         <oasis:entry colname="col3">1.33</oasis:entry>
         <oasis:entry colname="col4">1.45</oasis:entry>
         <oasis:entry colname="col5">1.49</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">MLO</oasis:entry>
         <oasis:entry colname="col2">6.65</oasis:entry>
         <oasis:entry colname="col3">6.87</oasis:entry>
         <oasis:entry colname="col4">14.61</oasis:entry>
         <oasis:entry colname="col5">14.82</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">MHD</oasis:entry>
         <oasis:entry colname="col2">1.98</oasis:entry>
         <oasis:entry colname="col3">1.87</oasis:entry>
         <oasis:entry colname="col4">1.95</oasis:entry>
         <oasis:entry colname="col5">2.78</oasis:entry>
       </oasis:row>
       <oasis:row rowsep="1">
         <oasis:entry colname="col1">NWR</oasis:entry>
         <oasis:entry colname="col2">12.54</oasis:entry>
         <oasis:entry colname="col3">12.52</oasis:entry>
         <oasis:entry colname="col4">12.94</oasis:entry>
         <oasis:entry colname="col5">12.94</oasis:entry>
       </oasis:row>
       <oasis:row rowsep="1">
         <oasis:entry namest="col1" nameend="col5">Flux sign reversal fractions [%] </oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">Pos<sub>soil sink</sub></oasis:entry>
         <oasis:entry colname="col2">2.83</oasis:entry>
         <oasis:entry colname="col3">2.77</oasis:entry>
         <oasis:entry colname="col4">2.94</oasis:entry>
         <oasis:entry colname="col5">2.79</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">Neg<sub>isoprene</sub></oasis:entry>
         <oasis:entry colname="col2">1.07</oasis:entry>
         <oasis:entry colname="col3">1.19</oasis:entry>
         <oasis:entry colname="col4">1.29</oasis:entry>
         <oasis:entry colname="col5">1.29</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">Neg<sub>non-isoprene</sub></oasis:entry>
         <oasis:entry colname="col2">2.41</oasis:entry>
         <oasis:entry colname="col3">1.54</oasis:entry>
         <oasis:entry colname="col4">2.26</oasis:entry>
         <oasis:entry colname="col5">2.01</oasis:entry>
       </oasis:row>
     </oasis:tbody>
   </oasis:tgroup></oasis:table><table-wrap-foot><p id="d2e7358"><sup>a</sup> Deep Southern Hemisphere (1.5 ppb MDM). <sup>b</sup> Deep Southern Hemisphere (1.5 ppb MDM) <inline-formula><mml:math id="M444" display="inline"><mml:mo>+</mml:mo></mml:math></inline-formula> Marine Boundary Layer (2.0 ppb MDM). <sup>c</sup> Deep Southern Hemisphere (1.5 ppb MDM) <inline-formula><mml:math id="M446" display="inline"><mml:mo>+</mml:mo></mml:math></inline-formula> Marine Boundary Layer (2.0 ppb MDM) <inline-formula><mml:math id="M447" display="inline"><mml:mo>+</mml:mo></mml:math></inline-formula> Mixed (2.5 ppb MDM).</p></table-wrap-foot></table-wrap>

      <p id="d2e7647">We identify experiment A2 as the best-performing configuration, as it yields the most physically consistent and interpretable flux estimates, with the lowest fraction of grid cells in which the optimised scaling factors reverse the sign of the prior flux, resulting in positive soil sinks or negative photochemical sources (Table <xref ref-type="table" rid="TC1"/>). This is illustrated for the South America RECCAP2 region in Fig. <xref ref-type="fig" rid="FC1"/>, where experiments with longer temporal correlation scales (A1 and A3) show a reduced soil sink, compensated by a smaller photochemical source (panel a). This is unlikely, as it largely arises from a more negative photochemical source (panel b) and a more positive soil sink (panel c). We also find this in other regions, including Africa and Oceania (not shown here). Experiment A2 also performs better because it has a more spatially coherent optimised soil sink and a non-isoprene <inline-formula><mml:math id="M455" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source, in contrast to the scattered and dipole-like patterns obtained in experiments A1 and A3. Finally, the regional fluxes of A2 show realistic interannual variability, whereas variability is strongly suppressed in A1 and A3. Based on these considerations, we adopt the <inline-formula><mml:math id="M456" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="italic">σ</mml:mi><mml:mi mathvariant="normal">MOY</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> values from experiment A2 for all remaining inversions.</p>
</sec>
<sec id="App1.Ch1.S3.SS2">
  <label>C2</label><title>Model-data mismatch</title>
      <p id="d2e7689">Next, we investigated the sensitivity of the inversions to the value of the model–data mismatch (MDM). This was achieved through offline inversions, which avoided rerunning the full transport model and reduced computational costs. In the base configuration (experiment A4), a uniform MDM of 3.0 ppb was assigned to all sites. To introduce spatially varying but physically motivated uncertainties, we performed a sequence of experiments with progressively refined site classifications. In experiment A5, Deep Southern Hemisphere (Deep SH) sites were assigned a reduced MDM of 1.5 ppb to reflect their relatively low transport and representation errors (Fig. <xref ref-type="fig" rid="FC2"/>), while all remaining sites retained an MDM of 3.0 ppb. Experiment A6 further distinguished marine boundary layer (MBL) sites, which were assigned an intermediate MDM of 2.0 ppb in addition to the reduced MDM for Deep SH sites. Finally, in experiment A7, mixed sites were assigned an MDM of 2.5 ppb, while land sites, characterised by stronger flux heterogeneity and transport uncertainty, retained the highest MDM of 3.0 ppb. We find experiment A5 performs best, because it yields a posterior chi-squared (<inline-formula><mml:math id="M457" display="inline"><mml:mrow><mml:msup><mml:mi mathvariant="italic">χ</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msup></mml:mrow></mml:math></inline-formula>) closest to unity at three out of four sites that we selected to represent the MDM classes (Deep SH, MBL, Mixed, and Land). Furthermore, experiment A5 produces the lowest fraction of grid cells with sign-reversing optimised scaling factors, leading to a negative non-isoprene <inline-formula><mml:math id="M458" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source or a positive soil sink (i.e. physically implausible fluxes), and hence we adopt its MDM classification in the remaining experiments (Table <xref ref-type="table" rid="TC2"/>).</p>
</sec>
<sec id="App1.Ch1.S3.SS3">
  <label>C3</label><title>Statevector element uncertainty</title>
      <p id="d2e7726">Lastly, we conducted offline inversion experiments by perturbing the uncertainties in our statevector elements across space and time. This was motivated by the observed relationship between the choice of these uncertainties and the occurrence of unphysical scaling factors, which could produce either a negative photochemical source or a positive soil sink (Fig. <xref ref-type="fig" rid="FC3"/>), even with a 300-member inversion. We found that reducing these uncertainties improved the physical consistency of the optimised fluxes while still allowing meaningful adjustments. The initial uncertainties were 15 %, 10 %, and 5 % for the soil sink, isoprene <inline-formula><mml:math id="M459" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source, and non-isoprene <inline-formula><mml:math id="M460" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source, respectively. Multiplying these values by a factor of 0.5 provided an optimum balance between minimising unphysical fluxes and maintaining meaningful spatial and temporal adjustments, as shown in Fig. <xref ref-type="fig" rid="FC3"/>. Another reason for choosing the factor 0.5 was that lowering it further led to increasingly large posterior residuals, as denoted by panel d. Consequently, in the main inversions, uncertainties of 7.5 %, 5 %, and 2.5 % are used for the soil sink, isoprene <inline-formula><mml:math id="M461" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source, and non-isoprene <inline-formula><mml:math id="M462" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source, respectively. </p>

      <fig id="FC1"><label>Figure C1</label><caption><p id="d2e7781">Annual mean H<sub>2</sub> fluxes over the South America RECCAP2 region (2003–2013) for the three  <inline-formula><mml:math id="M464" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="italic">σ</mml:mi><mml:mi mathvariant="normal">MOY</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> experiments. Panel <bold>(a)</bold> shows the soil sink versus the photochemical source, panel <bold>(b)</bold> shows the soil sink versus the negative photochemical source, panel <bold>(c)</bold> shows the soil sink versus the positive soil sink over this region.</p></caption>
          
          <graphic xlink:href="https://acp.copernicus.org/articles/26/16261/2026/acp-26-16261-2026-f12.png"/>

        </fig>

      <fig id="FC2"><label>Figure C2</label><caption><p id="d2e7823">Global distribution of <inline-formula><mml:math id="M465" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> surface flask and in situ sampling sites used in this study. Sites are coloured according to their site classification: marine boundary layer (MBL), land, mixed, polluted, and deep Southern Hemisphere (Deep SH).</p></caption>
          
          <graphic xlink:href="https://acp.copernicus.org/articles/26/16261/2026/acp-26-16261-2026-f13.png"/>

        </fig>

<fig id="FC3"><label>Figure C3</label><caption><p id="d2e7849">Sensitivity of unphysical <inline-formula><mml:math id="M466" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> fluxes and inversion performance metrics to the statevector element uncertainty scaling factor (PU). <bold>(a)</bold> Monthly mean positive soil dry deposition (sink) and negative photochemical production from isoprene and non-isoprene precursors (sources). <bold>(b)</bold> Maximum monthly values of these unphysical fluxes during the runtime (2003–2023). <bold>(c)</bold> Mean of the median innovation chi-squared (<inline-formula><mml:math id="M467" display="inline"><mml:mrow><mml:msubsup><mml:mi mathvariant="italic">χ</mml:mi><mml:mi mathvariant="normal">ino</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msubsup></mml:mrow></mml:math></inline-formula>) across all sites. <bold>(d)</bold> Mean of the median posterior residual across all sites.</p></caption>
          
          <graphic xlink:href="https://acp.copernicus.org/articles/26/16261/2026/acp-26-16261-2026-f14.png"/>

        </fig>

</sec>
</app>

<app id="App1.Ch1.S4">
  <label>Appendix D</label><title>Approximation of flux uncertainty</title>
      <p id="d2e7906">As shown in the previous sections, the choice of inversion parameters strongly influences the posterior fluxes inferred by the inversion. To quantify this impact, we systematically vary the number of ensemble members, observation subsets, scaling factors for the observation error variance (<inline-formula><mml:math id="M468" display="inline"><mml:mi>R</mml:mi></mml:math></inline-formula>), and scaling factors for the uncertainty assigned to each statevector element in space and time (<inline-formula><mml:math id="M469" display="inline"><mml:mi>P</mml:mi></mml:math></inline-formula>). This approach also provides a practical estimate of uncertainty within our ensemble Kalman filter-based inversion framework, for which formal posterior error estimates are otherwise not available. We perform the analysis using the best-performing configurations for the Sanderson and Bertagni schemes identified in Fig. <xref ref-type="sec" rid="Ch1.S3.SS1"/>, as well as alternative configurations shown in the same figure to assess sensitivity to the absolute magnitude of the priors. These alternative configurations correspond to prior source–sink pairs (photochemical source, soil sink) of (33.1, 40.3) and (44.1, 51.8) <inline-formula><mml:math id="M470" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> for the Sanderson experiments, and (33.1, 40.4) and (44.1, 67.3) <inline-formula><mml:math id="M471" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> for the Bertagni experiments. We note that scaling the prior flux magnitude also proportionally scales the assumed absolute uncertainty in flux space, thereby modifying both the prior statevector and its associated covariance matrix.</p>
      <p id="d2e7959">Table <xref ref-type="table" rid="TD1"/> shows that the posterior fluxes are most dependent on the absolute magnitude of the prior fluxes. In comparison, the posterior fluxes are relatively insensitive to the remaining inversion parameters. We find that the soil sink and the non-isoprene photochemical source are the most sta-ble flux components, whereas the isoprene <inline-formula><mml:math id="M472" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source shows larger variability, likely reflecting its spatial correlation with the soil sink. Apart from the absolute prior magnitudes, the largest deviations from the Base run occur in experiments with suboptimal configurations, i.e. when using a reduced number of ensemble members or a strongly limited observational network. This highlights the importance of sufficient ensemble size and observational coverage for robust flux optimisation. Even under these less constrained configurations, the inferred flux changes remain moderate, giving additional confidence to the robustness of our derived global <inline-formula><mml:math id="M473" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> budget. Based on all experiments in Table <xref ref-type="table" rid="TD1"/>, we derive a posterior error range of 47.8–56.7 <inline-formula><mml:math id="M474" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> for the soil sink and 29.2–38.2 <inline-formula><mml:math id="M475" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> for the photochemical source. For the individual photochemical source components, we derive an isoprene <inline-formula><mml:math id="M476" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source of 6.4–9.0 <inline-formula><mml:math id="M477" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>, and a non-isoprene <inline-formula><mml:math id="M478" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source of 22.8–30.2 <inline-formula><mml:math id="M479" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>. </p>

<table-wrap id="TD1"><label>Table D1</label><caption><p id="d2e8089">Posterior H<sub>2</sub> fluxes from 11 offline inversions (2003–2023) in which key inversion parameters were systematically perturbed. The final two rows show posterior H<sub>2</sub> fluxes from two online inversions (2014–2023) using alternative source–sink configurations for the photochemical source and soil sink. All fluxes are in <inline-formula><mml:math id="M482" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>. Experiment definitions: Base denotes the “best-performing” inversions used to derive our main results; P01/P10 denote scaling factors for the prior statevector uncertainty <inline-formula><mml:math id="M483" display="inline"><mml:mi>P</mml:mi></mml:math></inline-formula> (0.1<inline-formula><mml:math id="M484" display="inline"><mml:mo>×</mml:mo></mml:math></inline-formula>, 1.0<inline-formula><mml:math id="M485" display="inline"><mml:mo>×</mml:mo></mml:math></inline-formula>); R05/R20 denote scaling factors for the observational error variance <inline-formula><mml:math id="M486" display="inline"><mml:mi>R</mml:mi></mml:math></inline-formula> (0.5<inline-formula><mml:math id="M487" display="inline"><mml:mo>×</mml:mo></mml:math></inline-formula>, 2.0<inline-formula><mml:math id="M488" display="inline"><mml:mo>×</mml:mo></mml:math></inline-formula>); NMEM75/150/225 denote the number of ensemble members; NOBS10/25/50 denote reduced observation subsets (10 %, 25 %, 50 %); PRLOW/PRHIGH denote alternative (lower or higher) prior magnitudes.</p></caption><oasis:table frame="topbot"><oasis:tgroup cols="7">
     <oasis:colspec colnum="1" colname="col1" align="left"/>
     <oasis:colspec colnum="2" colname="col2" align="center"/>
     <oasis:colspec colnum="3" colname="col3" align="center" colsep="1"/>
     <oasis:colspec colnum="4" colname="col4" align="center"/>
     <oasis:colspec colnum="5" colname="col5" align="center" colsep="1"/>
     <oasis:colspec colnum="6" colname="col6" align="center"/>
     <oasis:colspec colnum="7" colname="col7" align="center"/>
     <oasis:thead>
       <oasis:row>
         <oasis:entry colname="col1"/>
         <oasis:entry rowsep="1" namest="col2" nameend="col3" colsep="1">Soil sink </oasis:entry>
         <oasis:entry rowsep="1" namest="col4" nameend="col5" colsep="1">isoprene <inline-formula><mml:math id="M489" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source </oasis:entry>
         <oasis:entry rowsep="1" namest="col6" nameend="col7">Non-isoprene <inline-formula><mml:math id="M490" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source </oasis:entry>
       </oasis:row>
       <oasis:row rowsep="1">
         <oasis:entry colname="col1"/>
         <oasis:entry colname="col2">Bertagni</oasis:entry>
         <oasis:entry colname="col3">Sanderson</oasis:entry>
         <oasis:entry colname="col4">Bertagni</oasis:entry>
         <oasis:entry colname="col5">Sanderson</oasis:entry>
         <oasis:entry colname="col6">Bertagni</oasis:entry>
         <oasis:entry colname="col7">Sanderson</oasis:entry>
       </oasis:row>
     </oasis:thead>
     <oasis:tbody>
       <oasis:row>
         <oasis:entry colname="col1">Base</oasis:entry>
         <oasis:entry colname="col2">51.8</oasis:entry>
         <oasis:entry colname="col3">53.9</oasis:entry>
         <oasis:entry colname="col4">8.2</oasis:entry>
         <oasis:entry colname="col5">7.7</oasis:entry>
         <oasis:entry colname="col6">26.7</oasis:entry>
         <oasis:entry colname="col7">26.6</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">P01</oasis:entry>
         <oasis:entry colname="col2">51.7</oasis:entry>
         <oasis:entry colname="col3">54.5</oasis:entry>
         <oasis:entry colname="col4">8.3</oasis:entry>
         <oasis:entry colname="col5">8.3</oasis:entry>
         <oasis:entry colname="col6">26.8</oasis:entry>
         <oasis:entry colname="col7">26.7</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">P10</oasis:entry>
         <oasis:entry colname="col2">51.6</oasis:entry>
         <oasis:entry colname="col3">53.5</oasis:entry>
         <oasis:entry colname="col4">8.1</oasis:entry>
         <oasis:entry colname="col5">7.5</oasis:entry>
         <oasis:entry colname="col6">26.7</oasis:entry>
         <oasis:entry colname="col7">26.6</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">R05</oasis:entry>
         <oasis:entry colname="col2">51.7</oasis:entry>
         <oasis:entry colname="col3">53.7</oasis:entry>
         <oasis:entry colname="col4">8.1</oasis:entry>
         <oasis:entry colname="col5">7.6</oasis:entry>
         <oasis:entry colname="col6">26.7</oasis:entry>
         <oasis:entry colname="col7">26.6</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">R20</oasis:entry>
         <oasis:entry colname="col2">51.8</oasis:entry>
         <oasis:entry colname="col3">54.1</oasis:entry>
         <oasis:entry colname="col4">8.2</oasis:entry>
         <oasis:entry colname="col5">7.9</oasis:entry>
         <oasis:entry colname="col6">26.7</oasis:entry>
         <oasis:entry colname="col7">26.6</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">NMEM75</oasis:entry>
         <oasis:entry colname="col2">51.2</oasis:entry>
         <oasis:entry colname="col3">55.3</oasis:entry>
         <oasis:entry colname="col4">8.4</oasis:entry>
         <oasis:entry colname="col5">8.5</oasis:entry>
         <oasis:entry colname="col6">26.3</oasis:entry>
         <oasis:entry colname="col7">27.4</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">NMEM150</oasis:entry>
         <oasis:entry colname="col2">52.0</oasis:entry>
         <oasis:entry colname="col3">55.9</oasis:entry>
         <oasis:entry colname="col4">8.7</oasis:entry>
         <oasis:entry colname="col5">8.4</oasis:entry>
         <oasis:entry colname="col6">26.4</oasis:entry>
         <oasis:entry colname="col7">27.7</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">NMEM225</oasis:entry>
         <oasis:entry colname="col2">52.1</oasis:entry>
         <oasis:entry colname="col3">54.1</oasis:entry>
         <oasis:entry colname="col4">8.2</oasis:entry>
         <oasis:entry colname="col5">7.5</oasis:entry>
         <oasis:entry colname="col6">26.7</oasis:entry>
         <oasis:entry colname="col7">27.0</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">NOBS10</oasis:entry>
         <oasis:entry colname="col2">51.5</oasis:entry>
         <oasis:entry colname="col3">54.5</oasis:entry>
         <oasis:entry colname="col4">8.1</oasis:entry>
         <oasis:entry colname="col5">8.2</oasis:entry>
         <oasis:entry colname="col6">26.7</oasis:entry>
         <oasis:entry colname="col7">26.6</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">NOBS25</oasis:entry>
         <oasis:entry colname="col2">51.5</oasis:entry>
         <oasis:entry colname="col3">54.4</oasis:entry>
         <oasis:entry colname="col4">8.1</oasis:entry>
         <oasis:entry colname="col5">8.2</oasis:entry>
         <oasis:entry colname="col6">26.6</oasis:entry>
         <oasis:entry colname="col7">26.5</oasis:entry>
       </oasis:row>
       <oasis:row rowsep="1">
         <oasis:entry colname="col1">NOBS50</oasis:entry>
         <oasis:entry colname="col2">51.5</oasis:entry>
         <oasis:entry colname="col3">54.4</oasis:entry>
         <oasis:entry colname="col4">8.2</oasis:entry>
         <oasis:entry colname="col5">8.1</oasis:entry>
         <oasis:entry colname="col6">26.3</oasis:entry>
         <oasis:entry colname="col7">26.7</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">PRLOW</oasis:entry>
         <oasis:entry colname="col2">48.8</oasis:entry>
         <oasis:entry colname="col3">47.8</oasis:entry>
         <oasis:entry colname="col4">6.4</oasis:entry>
         <oasis:entry colname="col5">6.6</oasis:entry>
         <oasis:entry colname="col6">22.8</oasis:entry>
         <oasis:entry colname="col7">22.7</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">PRHIGH</oasis:entry>
         <oasis:entry colname="col2">56.7</oasis:entry>
         <oasis:entry colname="col3">55.5</oasis:entry>
         <oasis:entry colname="col4">8.0</oasis:entry>
         <oasis:entry colname="col5">9.0</oasis:entry>
         <oasis:entry colname="col6">30.2</oasis:entry>
         <oasis:entry colname="col7">28.0</oasis:entry>
       </oasis:row>
     </oasis:tbody>
   </oasis:tgroup></oasis:table></table-wrap>

</app>

<app id="App1.Ch1.S5">
  <label>Appendix E</label><title>Spatiotemporal correlations</title>
      <p id="d2e8591">We find that both inversions show a strongly negative global correlation between the posterior soil sink and isoprene <inline-formula><mml:math id="M491" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source fluxes (Fig. <xref ref-type="fig" rid="FE1"/>). This shows that regions with stronger soil uptake (more negative flux values) tend to coincide with regions of higher isoprene <inline-formula><mml:math id="M492" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> emissions (more positive flux values), reflecting spatial co-location of strong flux magnitudes with opposite sign conventions, particularly in the tropics. Both inversions also show a negative correlation between the soil sink and the non-isoprene <inline-formula><mml:math id="M493" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source. This similarly reflects that regions of enhanced soil uptake tend to coincide with regions of enhanced emissions from this source category, despite limited spatial overlap in their dominant source regions. In contrast, the two <inline-formula><mml:math id="M494" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> source terms (isoprene and non-isoprene) show a positive correlation, indicating that regions with stronger emissions in one source category tend to also exhibit stronger emissions in the other. This suggests a shared spatial structure in the distribution of <inline-formula><mml:math id="M495" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> sources across regions. </p>

      <fig id="FE1"><label>Figure E1</label><caption><p id="d2e8655">Temporal evolution of area-weighted spatial correlations between posterior ensemble-mean <inline-formula><mml:math id="M496" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> flux field. Each panel shows the Pearson correlation computed across grid cells for a given month, using cosine-latitude weighting to account for grid-cell area. Correlations are calculated from posterior ensemble-mean flux fields rather than individual ensemble members, because the ensemble Kalman filter inversion substantially reduces posterior spread, leaving residual ensemble-member variability at the grid scale comparatively small.</p></caption>
        
        <graphic xlink:href="https://acp.copernicus.org/articles/26/16261/2026/acp-26-16261-2026-f15.png"/>

      </fig>

</app>

<app id="App1.Ch1.S6">
  <label>Appendix F</label><title>Additional figures and tables</title>

      <fig id="FF1"><label>Figure F1</label><caption><p id="d2e8688">Regions used in the soil sink schemes and inversions (statevector regions). The left panel shows the 22 GLDAS soil texture classes, while the right panel shows the 14 Sanderson ecosystem regions. Both are split at 25<inline-formula><mml:math id="M497" display="inline"><mml:mi mathvariant="italic">°</mml:mi></mml:math></inline-formula> N to approximately balance the land area between the Northern and Southern portion of the domain (dotted black line).</p></caption>
        
        <graphic xlink:href="https://acp.copernicus.org/articles/26/16261/2026/acp-26-16261-2026-f16.png"/>

      </fig>

<fig id="FF2"><label>Figure F2</label><caption><p id="d2e8709">Annual mean fluxes (2003–2023) for the Sanderson inversion, with corresponding latitudinal distributions.</p></caption>
        
        <graphic xlink:href="https://acp.copernicus.org/articles/26/16261/2026/acp-26-16261-2026-f17.png"/>

      </fig>

<fig id="FF3"><label>Figure F3</label><caption><p id="d2e8724">Annual mean fluxes (2003–2023) for the Bertagni inversion, with corresponding latitudinal distributions.</p></caption>
        
        <graphic xlink:href="https://acp.copernicus.org/articles/26/16261/2026/acp-26-16261-2026-f18.png"/>

      </fig>

<fig id="FF4"><label>Figure F4</label><caption><p id="d2e8738">Fluxes per RECCAP2 region (2003–2023 mean), ordered from left to right as: Sanderson prior, Sanderson posterior, Bertagni prior, and Bertagni posterior. Black dots with whiskers show the ensemble mean and one standard deviation range from <xref ref-type="bibr" rid="bib1.bibx41" id="text.133"/>. A small soil sink occurs over global ocean regions in our results (Sanderson: 2.7 <inline-formula><mml:math id="M498" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>; Bertagni: 0.6 <inline-formula><mml:math id="M499" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>), as some islands are included in the RECCAP2 ocean regions.</p></caption>
        
        <graphic xlink:href="https://acp.copernicus.org/articles/26/16261/2026/acp-26-16261-2026-f19.png"/>

      </fig>

<fig id="FF5"><label>Figure F5</label><caption><p id="d2e8789">Monthly simulated fluxes for each RECCAP2 region in the Sanderson and Bertagni inversions. Prior fluxes are shown as dotted lines and posterior fluxes as solid lines. A small soil sink occurs over global ocean regions in our results (Sanderson: 2.7 <inline-formula><mml:math id="M500" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>; Bertagni: 0.6 <inline-formula><mml:math id="M501" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>), as some islands are included in the RECCAP2 ocean regions.</p></caption>
        
        <graphic xlink:href="https://acp.copernicus.org/articles/26/16261/2026/acp-26-16261-2026-f20.jpg"/>

      </fig>

<table-wrap id="TF1"><label>Table F1</label><caption><p id="d2e8840">Overview of measurement networks used in this study, filtered for our study period (2003–2023). Sampling interval denotes the modal (i.e. most frequently occurring) time between consecutive measurements, computed across all stations or flight tracks within each network.</p></caption><oasis:table frame="topbot"><oasis:tgroup cols="5">
     <oasis:colspec colnum="1" colname="col1" align="left"/>
     <oasis:colspec colnum="2" colname="col2" align="right"/>
     <oasis:colspec colnum="3" colname="col3" align="left"/>
     <oasis:colspec colnum="4" colname="col4" align="left"/>
     <oasis:colspec colnum="5" colname="col5" align="left"/>
     <oasis:thead>
       <oasis:row rowsep="1">
         <oasis:entry colname="col1">Network</oasis:entry>
         <oasis:entry colname="col2">Stations/Flights</oasis:entry>
         <oasis:entry colname="col3">Start date</oasis:entry>
         <oasis:entry colname="col4">End date</oasis:entry>
         <oasis:entry colname="col5">Sampling interval</oasis:entry>
       </oasis:row>
     </oasis:thead>
     <oasis:tbody>
       <oasis:row rowsep="1">
         <oasis:entry namest="col1" nameend="col5">Surface observations </oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">NOAA-GML</oasis:entry>
         <oasis:entry colname="col2">66</oasis:entry>
         <oasis:entry colname="col3">2007–12–13</oasis:entry>
         <oasis:entry colname="col4">2023–12–31</oasis:entry>
         <oasis:entry colname="col5">7 d</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">ICOS</oasis:entry>
         <oasis:entry colname="col2">18</oasis:entry>
         <oasis:entry colname="col3">2019–06–26</oasis:entry>
         <oasis:entry colname="col4">2023–12–31</oasis:entry>
         <oasis:entry colname="col5">3 d</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">MPI-BGC</oasis:entry>
         <oasis:entry colname="col2">13</oasis:entry>
         <oasis:entry colname="col3">2003–03–13</oasis:entry>
         <oasis:entry colname="col4">2023–12–20</oasis:entry>
         <oasis:entry colname="col5">7 d</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">CSIRO</oasis:entry>
         <oasis:entry colname="col2">11</oasis:entry>
         <oasis:entry colname="col3">2003–01–02</oasis:entry>
         <oasis:entry colname="col4">2023–12–20</oasis:entry>
         <oasis:entry colname="col5">7 d</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">EUROHYDROS</oasis:entry>
         <oasis:entry colname="col2">6</oasis:entry>
         <oasis:entry colname="col3">2005–02–09</oasis:entry>
         <oasis:entry colname="col4">2009–10–15</oasis:entry>
         <oasis:entry colname="col5">7 d</oasis:entry>
       </oasis:row>
       <oasis:row rowsep="1">
         <oasis:entry colname="col1">AGAGE</oasis:entry>
         <oasis:entry colname="col2">2</oasis:entry>
         <oasis:entry colname="col3">2003–01–04</oasis:entry>
         <oasis:entry colname="col4">2023–06–30</oasis:entry>
         <oasis:entry colname="col5">40 min</oasis:entry>
       </oasis:row>
       <oasis:row rowsep="1">
         <oasis:entry namest="col1" nameend="col5">Aircraft observations </oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">ATom</oasis:entry>
         <oasis:entry colname="col2">41</oasis:entry>
         <oasis:entry colname="col3">2016–07–29</oasis:entry>
         <oasis:entry colname="col4">2018–05–21</oasis:entry>
         <oasis:entry colname="col5">2 min</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">IAGOS-CARIBIC</oasis:entry>
         <oasis:entry colname="col2">36</oasis:entry>
         <oasis:entry colname="col3">2007–05–22</oasis:entry>
         <oasis:entry colname="col4">2010–09–23</oasis:entry>
         <oasis:entry colname="col5">40 min</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">HIPPO</oasis:entry>
         <oasis:entry colname="col2">16</oasis:entry>
         <oasis:entry colname="col3">2009–01–09</oasis:entry>
         <oasis:entry colname="col4">2011–09–09</oasis:entry>
         <oasis:entry colname="col5">2 min</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">NOAA-GML</oasis:entry>
         <oasis:entry colname="col2">25</oasis:entry>
         <oasis:entry colname="col3">2009–09–28</oasis:entry>
         <oasis:entry colname="col4">2021–12–30</oasis:entry>
         <oasis:entry colname="col5">3 min</oasis:entry>
       </oasis:row>
     </oasis:tbody>
   </oasis:tgroup></oasis:table></table-wrap>

<table-wrap id="TF2"><label>Table F2</label><caption><p id="d2e9081">Evaluation of posterior <inline-formula><mml:math id="M502" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> mole fractions against independent aircraft data. Shown are the mean posterior MAE [ppb] binned by altitude and latitude for the Sanderson (left) and Bertagni (right) inversions for each measurement campaign.</p></caption><oasis:table frame="topbot"><oasis:tgroup cols="7">
     <oasis:colspec colnum="1" colname="col1" align="left"/>
     <oasis:colspec colnum="2" colname="col2" align="right"/>
     <oasis:colspec colnum="3" colname="col3" align="right"/>
     <oasis:colspec colnum="4" colname="col4" align="right" colsep="1"/>
     <oasis:colspec colnum="5" colname="col5" align="right"/>
     <oasis:colspec colnum="6" colname="col6" align="right"/>
     <oasis:colspec colnum="7" colname="col7" align="right"/>
     <oasis:thead>
       <oasis:row>
         <oasis:entry colname="col1"/>
         <oasis:entry rowsep="1" namest="col2" nameend="col4" align="center" colsep="1">Sanderson </oasis:entry>
         <oasis:entry rowsep="1" namest="col5" nameend="col7" align="center">Bertagni </oasis:entry>
       </oasis:row>
       <oasis:row rowsep="1">
         <oasis:entry colname="col1"/>
         <oasis:entry colname="col2">90–30° S</oasis:entry>
         <oasis:entry colname="col3">30° S–30° N</oasis:entry>
         <oasis:entry colname="col4">30–90° N</oasis:entry>
         <oasis:entry colname="col5">90–30° S</oasis:entry>
         <oasis:entry colname="col6">30° S–30° N</oasis:entry>
         <oasis:entry colname="col7">30–90° N</oasis:entry>
       </oasis:row>
     </oasis:thead>
     <oasis:tbody>
       <oasis:row rowsep="1">
         <oasis:entry namest="col1" nameend="col7">Below 5000 m </oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">ATom</oasis:entry>
         <oasis:entry colname="col2">30.15</oasis:entry>
         <oasis:entry colname="col3">24.43</oasis:entry>
         <oasis:entry colname="col4">26.68</oasis:entry>
         <oasis:entry colname="col5">28.04</oasis:entry>
         <oasis:entry colname="col6">25.10</oasis:entry>
         <oasis:entry colname="col7">25.82</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">IAGOS–CARIBIC</oasis:entry>
         <oasis:entry colname="col2">–</oasis:entry>
         <oasis:entry colname="col3">–</oasis:entry>
         <oasis:entry colname="col4">–</oasis:entry>
         <oasis:entry colname="col5">–</oasis:entry>
         <oasis:entry colname="col6">–</oasis:entry>
         <oasis:entry colname="col7">–</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">HIPPO</oasis:entry>
         <oasis:entry colname="col2">25.23</oasis:entry>
         <oasis:entry colname="col3">20.11</oasis:entry>
         <oasis:entry colname="col4">19.67</oasis:entry>
         <oasis:entry colname="col5">23.46</oasis:entry>
         <oasis:entry colname="col6">19.22</oasis:entry>
         <oasis:entry colname="col7">20.13</oasis:entry>
       </oasis:row>
       <oasis:row rowsep="1">
         <oasis:entry colname="col1">NOAA–GML</oasis:entry>
         <oasis:entry colname="col2">3.00</oasis:entry>
         <oasis:entry colname="col3">5.77</oasis:entry>
         <oasis:entry colname="col4">8.98</oasis:entry>
         <oasis:entry colname="col5">2.36</oasis:entry>
         <oasis:entry colname="col6">5.64</oasis:entry>
         <oasis:entry colname="col7">10.25</oasis:entry>
       </oasis:row>
       <oasis:row rowsep="1">
         <oasis:entry namest="col1" nameend="col7">Above 5000 m </oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">ATom</oasis:entry>
         <oasis:entry colname="col2">23.73</oasis:entry>
         <oasis:entry colname="col3">22.24</oasis:entry>
         <oasis:entry colname="col4">22.06</oasis:entry>
         <oasis:entry colname="col5">17.79</oasis:entry>
         <oasis:entry colname="col6">22.21</oasis:entry>
         <oasis:entry colname="col7">16.97</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">IAGOS–CARIBIC</oasis:entry>
         <oasis:entry colname="col2">3.04</oasis:entry>
         <oasis:entry colname="col3">15.70</oasis:entry>
         <oasis:entry colname="col4">14.44</oasis:entry>
         <oasis:entry colname="col5">7.93</oasis:entry>
         <oasis:entry colname="col6">15.99</oasis:entry>
         <oasis:entry colname="col7">17.45</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">HIPPO</oasis:entry>
         <oasis:entry colname="col2">26.96</oasis:entry>
         <oasis:entry colname="col3">23.21</oasis:entry>
         <oasis:entry colname="col4">22.27</oasis:entry>
         <oasis:entry colname="col5">20.44</oasis:entry>
         <oasis:entry colname="col6">21.90</oasis:entry>
         <oasis:entry colname="col7">14.81</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">NOAA–GML</oasis:entry>
         <oasis:entry colname="col2">3.88</oasis:entry>
         <oasis:entry colname="col3">6.08</oasis:entry>
         <oasis:entry colname="col4">9.35</oasis:entry>
         <oasis:entry colname="col5">8.20</oasis:entry>
         <oasis:entry colname="col6">5.88</oasis:entry>
         <oasis:entry colname="col7">10.65</oasis:entry>
       </oasis:row>
     </oasis:tbody>
   </oasis:tgroup></oasis:table></table-wrap>

<table-wrap id="TF3"><label>Table F3</label><caption><p id="d2e9374">Seasonality of the monthly <inline-formula><mml:math id="M503" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> soil sink across RECCAP2 regions. Values show the mean annual seasonal range in <inline-formula><mml:math id="M504" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">Tg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">yr</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> followed by the mean month of maximum uptake calculated over the 2003–2023 region.</p></caption><oasis:table frame="topbot"><oasis:tgroup cols="5">
     <oasis:colspec colnum="1" colname="col1" align="left"/>
     <oasis:colspec colnum="2" colname="col2" align="center"/>
     <oasis:colspec colnum="3" colname="col3" align="center" colsep="1"/>
     <oasis:colspec colnum="4" colname="col4" align="center"/>
     <oasis:colspec colnum="5" colname="col5" align="center"/>
     <oasis:thead>
       <oasis:row>
         <oasis:entry colname="col1"/>
         <oasis:entry rowsep="1" namest="col2" nameend="col3" colsep="1">Sanderson </oasis:entry>
         <oasis:entry rowsep="1" namest="col4" nameend="col5">Bertagni </oasis:entry>
       </oasis:row>
       <oasis:row rowsep="1">
         <oasis:entry colname="col1">Region</oasis:entry>
         <oasis:entry colname="col2">Prior</oasis:entry>
         <oasis:entry colname="col3">Posterior</oasis:entry>
         <oasis:entry colname="col4">Prior</oasis:entry>
         <oasis:entry colname="col5">Posterior</oasis:entry>
       </oasis:row>
     </oasis:thead>
     <oasis:tbody>
       <oasis:row>
         <oasis:entry colname="col1">North America</oasis:entry>
         <oasis:entry colname="col2">5.4 (Jun)</oasis:entry>
         <oasis:entry colname="col3">8.1 (Jun)</oasis:entry>
         <oasis:entry colname="col4">6.2 (May)</oasis:entry>
         <oasis:entry colname="col5">7.4 (Jul)</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">South America</oasis:entry>
         <oasis:entry colname="col2">3.3 (Aug)</oasis:entry>
         <oasis:entry colname="col3">8.4 (Aug)</oasis:entry>
         <oasis:entry colname="col4">2.1 (Aug)</oasis:entry>
         <oasis:entry colname="col5">5.5 (Aug)</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">Europe</oasis:entry>
         <oasis:entry colname="col2">1.7 (Jun)</oasis:entry>
         <oasis:entry colname="col3">1.1 (Jul)</oasis:entry>
         <oasis:entry colname="col4">1.9 (May)</oasis:entry>
         <oasis:entry colname="col5">2.1 (Jul)</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">Africa</oasis:entry>
         <oasis:entry colname="col2">1.2 (Jul)</oasis:entry>
         <oasis:entry colname="col3">6.0 (Jul)</oasis:entry>
         <oasis:entry colname="col4">2.5 (Sep)</oasis:entry>
         <oasis:entry colname="col5">4.7 (Sep)</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">Russia</oasis:entry>
         <oasis:entry colname="col2">6.5 (Jun)</oasis:entry>
         <oasis:entry colname="col3">9.0 (Jun)</oasis:entry>
         <oasis:entry colname="col4">7.4 (May)</oasis:entry>
         <oasis:entry colname="col5">10.3 (Jun)</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">Middle East</oasis:entry>
         <oasis:entry colname="col2">2.2 (Jun)</oasis:entry>
         <oasis:entry colname="col3">2.4 (Jun)</oasis:entry>
         <oasis:entry colname="col4">1.9 (Apr)</oasis:entry>
         <oasis:entry colname="col5">1.8 (Apr)</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">South Asia</oasis:entry>
         <oasis:entry colname="col2">1.1 (Mar)</oasis:entry>
         <oasis:entry colname="col3">1.6 (Jun)</oasis:entry>
         <oasis:entry colname="col4">0.7 (Aug)</oasis:entry>
         <oasis:entry colname="col5">1.3 (Aug)</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">East Asia</oasis:entry>
         <oasis:entry colname="col2">1.3 (Mar)</oasis:entry>
         <oasis:entry colname="col3">1.4 (Jun)</oasis:entry>
         <oasis:entry colname="col4">1.9 (May)</oasis:entry>
         <oasis:entry colname="col5">3.1 (Jun)</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">South East Asia</oasis:entry>
         <oasis:entry colname="col2">1.1 (Feb)</oasis:entry>
         <oasis:entry colname="col3">1.5 (May)</oasis:entry>
         <oasis:entry colname="col4">0.5 (Mar)</oasis:entry>
         <oasis:entry colname="col5">1.2 (Jun)</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">Oceania</oasis:entry>
         <oasis:entry colname="col2">0.4 (Oct)</oasis:entry>
         <oasis:entry colname="col3">1.0 (Jul)</oasis:entry>
         <oasis:entry colname="col4">1.0 (Jul)</oasis:entry>
         <oasis:entry colname="col5">1.8 (Jul)</oasis:entry>
       </oasis:row>
     </oasis:tbody>
   </oasis:tgroup></oasis:table></table-wrap>

<table-wrap id="TF4"><label>Table F4</label><caption><p id="d2e9641">Correlation statistics between monthly posterior soil sink anomalies and the Oceanic Niño Index (ONI; units <inline-formula><mml:math id="M505" display="inline"><mml:mi mathvariant="italic">°</mml:mi></mml:math></inline-formula>C). Monthly soil sink anomalies are computed as deviations from a month-specific climatology over the final 10 years of the inversions (2014–2024). Metric <inline-formula><mml:math id="M506" display="inline"><mml:mi>r</mml:mi></mml:math></inline-formula> is the Pearson correlation coefficient, and <inline-formula><mml:math id="M507" display="inline"><mml:mi>p</mml:mi></mml:math></inline-formula> the <inline-formula><mml:math id="M508" display="inline"><mml:mi>p</mml:mi></mml:math></inline-formula>-value. Only regions where <inline-formula><mml:math id="M509" display="inline"><mml:mrow><mml:mi>p</mml:mi><mml:mo>&lt;</mml:mo><mml:mn mathvariant="normal">0.05</mml:mn></mml:mrow></mml:math></inline-formula> in either the Sanderson or Bertagni inversion are shown.</p></caption><oasis:table frame="topbot"><oasis:tgroup cols="7">
     <oasis:colspec colnum="1" colname="col1" align="left"/>
     <oasis:colspec colnum="2" colname="col2" align="right"/>
     <oasis:colspec colnum="3" colname="col3" align="right"/>
     <oasis:colspec colnum="4" colname="col4" align="center" colsep="1"/>
     <oasis:colspec colnum="5" colname="col5" align="center"/>
     <oasis:colspec colnum="6" colname="col6" align="center"/>
     <oasis:colspec colnum="7" colname="col7" align="center"/>
     <oasis:thead>
       <oasis:row>
         <oasis:entry colname="col1"/>
         <oasis:entry rowsep="1" namest="col2" nameend="col4" align="center" colsep="1">Sanderson </oasis:entry>
         <oasis:entry rowsep="1" namest="col5" nameend="col7">Bertagni </oasis:entry>
       </oasis:row>
       <oasis:row rowsep="1">
         <oasis:entry colname="col1">Region</oasis:entry>
         <oasis:entry colname="col2">Slope [Tg yr<sup>−1</sup> ONI<sup>−1</sup>]</oasis:entry>
         <oasis:entry colname="col3"><inline-formula><mml:math id="M512" display="inline"><mml:mi>r</mml:mi></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col4"><inline-formula><mml:math id="M513" display="inline"><mml:mi>p</mml:mi></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col5">Slope [Tg yr<sup>−1</sup> ONI<sup>−1</sup>]</oasis:entry>
         <oasis:entry colname="col6"><inline-formula><mml:math id="M516" display="inline"><mml:mi>r</mml:mi></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col7"><inline-formula><mml:math id="M517" display="inline"><mml:mi>p</mml:mi></mml:math></inline-formula></oasis:entry>
       </oasis:row>
     </oasis:thead>
     <oasis:tbody>
       <oasis:row>
         <oasis:entry colname="col1">Global</oasis:entry>
         <oasis:entry colname="col2">0.98</oasis:entry>
         <oasis:entry colname="col3">0.21</oasis:entry>
         <oasis:entry colname="col4">0.02</oasis:entry>
         <oasis:entry colname="col5">2.11</oasis:entry>
         <oasis:entry colname="col6">0.50</oasis:entry>
         <oasis:entry colname="col7">0.00</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">South American Tropical</oasis:entry>
         <oasis:entry colname="col2">0.63</oasis:entry>
         <oasis:entry colname="col3">0.48</oasis:entry>
         <oasis:entry colname="col4">0.00</oasis:entry>
         <oasis:entry colname="col5">0.97</oasis:entry>
         <oasis:entry colname="col6">0.63</oasis:entry>
         <oasis:entry colname="col7">0.00</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">Southern Africa</oasis:entry>
         <oasis:entry colname="col2"><inline-formula><mml:math id="M518" display="inline"><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">0.46</mml:mn></mml:mrow></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col3"><inline-formula><mml:math id="M519" display="inline"><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">0.33</mml:mn></mml:mrow></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col4">0.00</oasis:entry>
         <oasis:entry colname="col5">0.13</oasis:entry>
         <oasis:entry colname="col6">0.16</oasis:entry>
         <oasis:entry colname="col7">0.09</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">South American Temperate</oasis:entry>
         <oasis:entry colname="col2">0.21</oasis:entry>
         <oasis:entry colname="col3">0.26</oasis:entry>
         <oasis:entry colname="col4">0.00</oasis:entry>
         <oasis:entry colname="col5">0.34</oasis:entry>
         <oasis:entry colname="col6">0.54</oasis:entry>
         <oasis:entry colname="col7">0.00</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">Tropical Asia</oasis:entry>
         <oasis:entry colname="col2">0.19</oasis:entry>
         <oasis:entry colname="col3">0.29</oasis:entry>
         <oasis:entry colname="col4">0.00</oasis:entry>
         <oasis:entry colname="col5">0.25</oasis:entry>
         <oasis:entry colname="col6">0.32</oasis:entry>
         <oasis:entry colname="col7">0.00</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">Europe</oasis:entry>
         <oasis:entry colname="col2">0.16</oasis:entry>
         <oasis:entry colname="col3">0.40</oasis:entry>
         <oasis:entry colname="col4">0.00</oasis:entry>
         <oasis:entry colname="col5">0.16</oasis:entry>
         <oasis:entry colname="col6">0.31</oasis:entry>
         <oasis:entry colname="col7">0.00</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">North American Temperate</oasis:entry>
         <oasis:entry colname="col2">0.10</oasis:entry>
         <oasis:entry colname="col3">0.24</oasis:entry>
         <oasis:entry colname="col4">0.01</oasis:entry>
         <oasis:entry colname="col5">0.19</oasis:entry>
         <oasis:entry colname="col6">0.33</oasis:entry>
         <oasis:entry colname="col7">0.00</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">North American Boreal</oasis:entry>
         <oasis:entry colname="col2">0.09</oasis:entry>
         <oasis:entry colname="col3">0.18</oasis:entry>
         <oasis:entry colname="col4">0.05</oasis:entry>
         <oasis:entry colname="col5">0.14</oasis:entry>
         <oasis:entry colname="col6">0.31</oasis:entry>
         <oasis:entry colname="col7">0.00</oasis:entry>
       </oasis:row>
     </oasis:tbody>
   </oasis:tgroup></oasis:table></table-wrap>

</app>
  </app-group><notes notes-type="codedataavailability"><title>Code and data availability</title>

      <p id="d2e10035">The CarbonTracker and TM5-MP living codebases are openly available via <uri>https://git.wur.nl/ctdas</uri> (last access: 28 May 2026) and <uri>https://ci.tno.nl/gitlab/tm5/tm5-mp</uri> (last access: 28 May 2026), respectively. The two-dimensional prior and posterior monthly <inline-formula><mml:math id="M520" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">H</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> fluxes produced in this study are archived on <ext-link xlink:href="https://doi.org/10.5281/zenodo.20427212" ext-link-type="DOI">10.5281/zenodo.20427212</ext-link> <xref ref-type="bibr" rid="bib1.bibx64" id="paren.134"/> at 1 <inline-formula><mml:math id="M521" display="inline"><mml:mo>×</mml:mo></mml:math></inline-formula> 1 resolution. Daily three-dimensional flux fields are available upon request.</p>
  </notes><notes notes-type="authorcontribution"><title>Author contributions</title>

      <p id="d2e10072">FS performed the simulations, analysed and synthesised the data, produced the figures, and led the writing of the manuscript. WP supervised the study and contributed to the conceptualisation and design of the project, manuscript structure, and writing. JH contributed to the development of the model code and the design of the simulations. MK contributed to the conceptualisation and design of the project. WP, JH, and MK reviewed the manuscript.</p>
  </notes><notes notes-type="competinginterests"><title>Competing interests</title>

      <p id="d2e10079">The contact author has declared that none of the authors has any competing interests.</p>
  </notes><notes notes-type="disclaimer"><title>Disclaimer</title>

      <p id="d2e10085">Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. The authors bear the ultimate responsibility for providing appropriate place names. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.</p>
  </notes><ack><title>Acknowledgements</title><p id="d2e10091">We contacted the PIs of all observational datasets used in this study and gratefully acknowledge the EUROHYDROS <xref ref-type="bibr" rid="bib1.bibx5" id="paren.135"/>, NOAA-GML <xref ref-type="bibr" rid="bib1.bibx52" id="paren.136"/>, CSIRO-GASLAB <xref ref-type="bibr" rid="bib1.bibx32" id="paren.137"/>, MPI-BGC <xref ref-type="bibr" rid="bib1.bibx30" id="paren.138"/>, and ICOS <xref ref-type="bibr" rid="bib1.bibx1" id="paren.139"/> networks for surface flask H<sub>2</sub> observations, as well as the AGAGE network <xref ref-type="bibr" rid="bib1.bibx51" id="paren.140"/> for in situ H<sub>2</sub> observations. We also acknowledge the HIPPO <xref ref-type="bibr" rid="bib1.bibx79 bib1.bibx77" id="paren.141"/>, ATom <xref ref-type="bibr" rid="bib1.bibx35 bib1.bibx67 bib1.bibx78" id="paren.142"/>, NOAA-GML <xref ref-type="bibr" rid="bib1.bibx34 bib1.bibx65" id="paren.143"/>, and IAGOS-CARIBIC <xref ref-type="bibr" rid="bib1.bibx4" id="paren.144"/> campaigns for aircraft H<sub>2</sub> observations. ATom additionally provided CH<sub>2</sub>O and photolysis rate constant (JaCH<sub>2</sub>O) aircraft observations <xref ref-type="bibr" rid="bib1.bibx21 bib1.bibx22 bib1.bibx78" id="paren.145"/>. AI tools (ChatGPT and GitHub Copilot) were used to assist with language editing, manuscript readability, brainstorming, and coding support for analysis scripts. No generative AI tools were used to generate scientific interpretations or conclusions. All outputs were reviewed and validated by the contact author.</p></ack><notes notes-type="financialsupport"><title>Financial support</title>

      <p id="d2e10176">This research has been supported by the Nationaal Programma Groningen and the European Union via the subsidy “Waterstof Werkt: Train and Learn Hub” and by the Gas &amp; Hydrogen Partnerships Shell Nederland. All simulations were performed on the HPC cluster Aether at the University of Bremen, financed by DFG within the scope of the Excellence Initiative.</p>
  </notes><notes notes-type="reviewstatement"><title>Review statement</title>

      <p id="d2e10182">This paper was edited by Gunnar Myhre and reviewed by Alexander Tardito Chaudhri and one anonymous referee.</p>
  </notes><ref-list>
    <title>References</title>

      <ref id="bib1.bibx1"><label>Apadula et al.(2025)Apadula, Arnold, Bergamaschi, Biermann, Chen, Colomb, Conil, Couret, Cristofanelli, De Mazière, Delmotte, Di Iorio, Emmenegger, Forster, Frumau, Harris, Haszpra, Hatakka, Heliasz, Heltai, Hensen, Hermansen, Hoheisel, Kneuer, Komínková, Kubistin, Larmanou, Laurent, Laurila, Lehner, Lehtinen, Leskinen, Leuenberger, Levula, Lindauer, Lopez, Lund Myhre, Lunder, Mammarella, Manca, Manning, Marek, Marklund, Meinhardt, Miettinen, Molnár, Montaguti, Mölder, Müller-Williams, O'Doherty, Ottosson-Löfvenius, Piacentino, Pichon, Pitt, Platt, Plaß-Dülmer, Ramonet, Rivas-Soriano, Roulet, Scheeren, Schmidt, Schumacher, Sferlazzo, Sha, Smith, Stanley, Steinbacher, Sørensen, Trisolino, Vítková, Ylisirniö, Yver-Kwok, Zazzeri, Zwerschke, di Sarra, ICOS ATC-Laboratoires des Sciences du Climat et de L'Environnement (LSCE), France, ICOS Central Radiocarbon Laboratory (CRL), Germany, ICOS Flask And Calibration Laboratory (FCL), Germany, ICOS-CAL-CRL, and ICOS-CAL-FCL</label><mixed-citation>Apadula, F., Arnold, S., Bergamaschi, P., Biermann, T., Chen, H., Colomb, A., Conil, S., Couret, C., Cristofanelli, P., De Mazière, M., Delmotte, M., Di Iorio, T., Emmenegger, L., Forster, G., Frumau, A., Harris, E., Haszpra, L., Hatakka, J., Heliasz, M., Heltai, D., Hensen, A., Hermansen, O., Hoheisel, A., Kneuer, T., Komínková, K., Kubistin, D., Larmanou, E., Laurent, O., Laurila, T., Lehner, I., Lehtinen, K., Leskinen, A., Leuenberger, M., Levula, J., Lindauer, M., Lopez, M., Lund Myhre, C., Lunder, C., Mammarella, I., Manca, G., Manning, A., Marek, M. V., Marklund, P., Meinhardt, F., Miettinen, P., Molnár, M., Montaguti, S., Mölder, M., Müller-Williams, J., O'Doherty, S., Ottosson-Löfvenius, M., Piacentino, S., Pichon, J.-M., Pitt, J., Platt, S. M., Plaß-Dülmer, C., Ramonet, M., Rivas-Soriano, P., Roulet, Y.-A., Scheeren, B., Schmidt, M., Schumacher, M., Sferlazzo, D., Sha, M. K., Smith, P., Stanley, K., Steinbacher, M., Sørensen, L. L., Trisolino, P., Vítková, G., Ylisirniö, A., Yver-Kwok, C., Zazzeri, G., Zwerschke, E., di Sarra, A., ICOS ATC-Laboratoires des Sciences du Climat et de L'Environnement (LSCE), France, ICOS Central Radiocarbon Laboratory (CRL), Germany, ICOS Flask And Calibration Laboratory (FCL), Germany, ICOS-CAL-CRL, and ICOS-CAL-FCL: ICOS Atmosphere Release 2025-1 of Level 2 Greenhouse Gas Mole Fractions of CO2, CH4, N2O, CO, meteorology and 14CO2, and flask samples analysed for CO2, CH4, N2O, CO, H2, SF6, delta 13C CO2, delta18OCO2, delta O2N2 and 14C,   ICOS ATC time series,ICOS ATC Flask time series Version Number: 1.0, <ext-link xlink:href="https://doi.org/10.18160/PP29-9CNZ" ext-link-type="DOI">10.18160/PP29-9CNZ</ext-link>, 2025.</mixed-citation></ref>
      <ref id="bib1.bibx2"><label>Aronson et al.(2019)</label><mixed-citation>Aronson, E. L., Dierick, D., Botthoff, J. K., Oberbauer, S., Zelikova, T. J., Harmon, T. C., Rundel, P., Johnson, R. F., Swanson, A. C., Pinto‐Tomás, A. A., Artavia-León, A., Matarrita-Carranza, B., and Allen, M. F.: ENSO-Influenced Drought Drives Methane Flux Dynamics in a Tropical Wet Forest Soil, J. Geophys. Res.-Biogeo., 124, 2267–2276, <ext-link xlink:href="https://doi.org/10.1029/2018JG004832" ext-link-type="DOI">10.1029/2018JG004832</ext-link>, 2019.</mixed-citation></ref>
      <ref id="bib1.bibx3"><label>Barnston et al.(1997)Barnston, Chelliah, and Goldenberg</label><mixed-citation>Barnston, A. G., Chelliah, M., and Goldenberg, S. B.: Documentation of a highly ENSO-related sst region in the equatorial pacific: Research note, Atmosphere-Ocean, 35, 367–383, <ext-link xlink:href="https://doi.org/10.1080/07055900.1997.9649597" ext-link-type="DOI">10.1080/07055900.1997.9649597</ext-link>, 1997.</mixed-citation></ref>
      <ref id="bib1.bibx4"><label>Batenburg et al.(2016)Batenburg, Schuck, Baker, Zahn, Brenninkmeijer, and Roeckmann</label><mixed-citation>Batenburg, A., Schuck, T., Baker, A., Zahn, A., Brenninkmeijer, C., and Roeckmann, T.: H2 mixing ratio and isotope data from CARIBIC project, figshare, <ext-link xlink:href="https://doi.org/10.6084/M9.FIGSHARE.2068992.V1" ext-link-type="DOI">10.6084/M9.FIGSHARE.2068992.V1</ext-link>,  2016.</mixed-citation></ref>
      <ref id="bib1.bibx5"><label>Batenburg et al.(2011)Batenburg, Walter, Pieterse, Levin, Schmidt, Jordan, Hammer, Yver, and Röckmann</label><mixed-citation>Batenburg, A. M., Walter, S., Pieterse, G., Levin, I., Schmidt, M., Jordan, A., Hammer, S., Yver, C., and Röckmann, T.: Temporal and spatial variability of the stable isotopic composition of atmospheric molecular hydrogen: observations at six EUROHYDROS stations, Atmos. Chem. Phys., 11, 6985–6999, <ext-link xlink:href="https://doi.org/10.5194/acp-11-6985-2011" ext-link-type="DOI">10.5194/acp-11-6985-2011</ext-link>, 2011.</mixed-citation></ref>
      <ref id="bib1.bibx6"><label>Bertagni et al.(2021)Bertagni, Paulot, and Porporato</label><mixed-citation>Bertagni, M. B., Paulot, F., and Porporato, A.: Moisture Fluctuations Modulate Abiotic and Biotic Limitations of H2 Soil Uptake, Global Biogeochem. Cy., 35, <ext-link xlink:href="https://doi.org/10.1029/2021GB006987" ext-link-type="DOI">10.1029/2021GB006987</ext-link>, 2021.</mixed-citation></ref>
      <ref id="bib1.bibx7"><label>Bousquet et al.(2011)Bousquet, Yver, Pison, Li, Fortems, Hauglustaine, Szopa, Rayner, Novelli, Langenfelds, Steele, Ramonet, Schmidt, Foster, Morfopoulos, and Ciais</label><mixed-citation>Bousquet, P., Yver, C., Pison, I., Li, Y. S., Fortems, A., Hauglustaine, D., Szopa, S., Rayner, P. J., Novelli, P., Langenfelds, R., Steele, P., Ramonet, M., Schmidt, M., Foster, P., Morfopoulos, C., and Ciais, P.: A three-dimensional synthesis inversion of the molecular hydrogen cycle: Sources and sinks budget and implications for the soil uptake, J. Geophys. Res., 116, D01302, <ext-link xlink:href="https://doi.org/10.1029/2010JD014599" ext-link-type="DOI">10.1029/2010JD014599</ext-link>, 2011.</mixed-citation></ref>
      <ref id="bib1.bibx8"><label>Chen et al.(2015)Chen, Popa, Batenburg, and Röckmann</label><mixed-citation>Chen, Q., Popa, M. E., Batenburg, A. M., and Röckmann, T.: Isotopic signatures of production and uptake of H2 by soil, Atmos. Chem. Phys., 15, 13003–13021, <ext-link xlink:href="https://doi.org/10.5194/acp-15-13003-2015" ext-link-type="DOI">10.5194/acp-15-13003-2015</ext-link>, 2015.</mixed-citation></ref>
      <ref id="bib1.bibx9"><label>Ciais et al.(2022)Ciais, Bastos, Chevallier, Lauerwald, Poulter, Canadell, Hugelius, Jackson, Jain, Jones, Kondo, Luijkx, Patra, Peters, Pongratz, Petrescu, Piao, Qiu, Von Randow, Regnier, Saunois, Scholes, Shvidenko, Tian, Yang, Wang, and Zheng</label><mixed-citation>Ciais, P., Bastos, A., Chevallier, F., Lauerwald, R., Poulter, B., Canadell, J. G., Hugelius, G., Jackson, R. B., Jain, A., Jones, M., Kondo, M., Luijkx, I. T., Patra, P. K., Peters, W., Pongratz, J., Petrescu, A. M. R., Piao, S., Qiu, C., Von Randow, C., Regnier, P., Saunois, M., Scholes, R., Shvidenko, A., Tian, H., Yang, H., Wang, X., and Zheng, B.: Definitions and methods to estimate regional land carbon fluxes for the second phase of the REgional Carbon Cycle Assessment and Processes Project (RECCAP-2), Geosci. Model Dev., 15, 1289–1316, <ext-link xlink:href="https://doi.org/10.5194/gmd-15-1289-2022" ext-link-type="DOI">10.5194/gmd-15-1289-2022</ext-link>, 2022.</mixed-citation></ref>
      <ref id="bib1.bibx10"><label>Conrad(1996)</label><mixed-citation>Conrad, R.: Soil microorganisms as controllers of atmospheric trace gases (H2, CO, CH4, OCS, N2O, and NO), Microbiol. Rev., 60, 609–640, <ext-link xlink:href="https://doi.org/10.1128/mr.60.4.609-640.1996" ext-link-type="DOI">10.1128/mr.60.4.609-640.1996</ext-link>, 1996.</mixed-citation></ref>
      <ref id="bib1.bibx11"><label>Conrad et al.(1989)Conrad, Mayer, and Wüst</label><mixed-citation>Conrad, R., Mayer, H.-P., and Wüst, M.: Temporal change of gas metabolism by hydrogen-syntrophic methanogenic bacterial associations in anoxic paddy soil, FEMS Microbiol. Lett., 62, 265–273, <ext-link xlink:href="https://doi.org/10.1111/j.1574-6968.1989.tb03701.x" ext-link-type="DOI">10.1111/j.1574-6968.1989.tb03701.x</ext-link>, 1989.</mixed-citation></ref>
      <ref id="bib1.bibx12"><label>Conte et al.(2019)Conte, Szopa, Séférian, and Bopp</label><mixed-citation>Conte, L., Szopa, S., Séférian, R., and Bopp, L.: The oceanic cycle of carbon monoxide and its emissions to the atmosphere, Biogeosciences, 16, 881–902, <ext-link xlink:href="https://doi.org/10.5194/bg-16-881-2019" ext-link-type="DOI">10.5194/bg-16-881-2019</ext-link>, 2019.</mixed-citation></ref>
      <ref id="bib1.bibx13"><label>Cowan et al.(2025)Cowan, Roberts, Hanlon, Bezanger, Toteva, Tweedie, Yeung, Deshpande, Levy, Skiba, Nemitz, and Drewer</label><mixed-citation>Cowan, N., Roberts, T., Hanlon, M., Bezanger, A., Toteva, G., Tweedie, A., Yeung, K., Deshpande, A., Levy, P., Skiba, U., Nemitz, E., and Drewer, J.: Quantifying the soil sink of atmospheric hydrogen: a full year of field measurements from grassland and forest soils in the UK, Biogeosciences, 22, 3449–3461, <ext-link xlink:href="https://doi.org/10.5194/bg-22-3449-2025" ext-link-type="DOI">10.5194/bg-22-3449-2025</ext-link>, 2025.</mixed-citation></ref>
      <ref id="bib1.bibx14"><label>De Kleijne et al.(2024)De Kleijne, Huijbregts, Knobloch, Van Zelm, Hilbers, De Coninck, and Hanssen</label><mixed-citation>De Kleijne, K., Huijbregts, M. A. J., Knobloch, F., Van Zelm, R., Hilbers, J. P., De Coninck, H., and Hanssen, S. V.: Worldwide greenhouse gas emissions of green hydrogen production and transport, Nature Energy, <ext-link xlink:href="https://doi.org/10.1038/s41560-024-01563-1" ext-link-type="DOI">10.1038/s41560-024-01563-1</ext-link>, 2024.</mixed-citation></ref>
      <ref id="bib1.bibx15"><label>Ehhalt and Rohrer(2009)</label><mixed-citation>Ehhalt, D. H. and Rohrer, F.: The tropospheric cycle of H2: a critical review, Tellus B, 61, 500–535, <ext-link xlink:href="https://doi.org/10.1111/j.1600-0889.2009.00416.x" ext-link-type="DOI">10.1111/j.1600-0889.2009.00416.x</ext-link>, 2009.</mixed-citation></ref>
      <ref id="bib1.bibx16"><label>Ehhalt and Rohrer(2011)</label><mixed-citation>Ehhalt, D. H. and Rohrer, F.: The dependence of soil H<sub>2</sub> uptake on temperature and moisture: a reanalysis of laboratory data, Tellus B, 63, 1040, <ext-link xlink:href="https://doi.org/10.1111/j.1600-0889.2011.00581.x" ext-link-type="DOI">10.1111/j.1600-0889.2011.00581.x</ext-link>, 2011.</mixed-citation></ref>
      <ref id="bib1.bibx17"><label>Feng et al.(2020)Feng, Smith, Braun, Crippa, Gidden, Hoesly, Klimont, Van Marle, Van Den Berg, and Van Der Werf</label><mixed-citation>Feng, L., Smith, S. J., Braun, C., Crippa, M., Gidden, M. J., Hoesly, R., Klimont, Z., van Marle, M., van den Berg, M., and van der Werf, G. R.: The generation of gridded emissions data for CMIP6, Geosci. Model Dev., 13, 461–482, <ext-link xlink:href="https://doi.org/10.5194/gmd-13-461-2020" ext-link-type="DOI">10.5194/gmd-13-461-2020</ext-link>, 2020.</mixed-citation></ref>
      <ref id="bib1.bibx18"><label>Gibson et al.(2024)Gibson, Eby, and Jaggi</label><mixed-citation>Gibson, J., Eby, P., and Jaggi, A.: Natural isotope fingerprinting of produced hydrogen and its potential applications to the hydrogen economy, Int. J. Hydrogen Energ., 66, 468–478, <ext-link xlink:href="https://doi.org/10.1016/j.ijhydene.2024.04.077" ext-link-type="DOI">10.1016/j.ijhydene.2024.04.077</ext-link>, 2024.</mixed-citation></ref>
      <ref id="bib1.bibx19"><label>Gros et al.(2009)Gros, Peeken, Bluhm, Zöllner, Sarda-Esteve, and Bonsang</label><mixed-citation>Gros, V., Peeken, I., Bluhm, K., Zöllner, E., Sarda-Esteve, R., and Bonsang, B.: Carbon monoxide emissions by phytoplankton: evidence from laboratory experiments, Environ. Chem., 6, 369, <ext-link xlink:href="https://doi.org/10.1071/EN09020" ext-link-type="DOI">10.1071/EN09020</ext-link>, 2009.</mixed-citation></ref>
      <ref id="bib1.bibx20"><label>Guenther et al.(2012)Guenther, Jiang, Heald, Sakulyanontvittaya, Duhl, Emmons, and Wang</label><mixed-citation>Guenther, A. B., Jiang, X., Heald, C. L., Sakulyanontvittaya, T., Duhl, T., Emmons, L. K., and Wang, X.: The Model of Emissions of Gases and Aerosols from Nature version 2.1 (MEGAN2.1): an extended and updated framework for modeling biogenic emissions, Geosci. Model Dev., 5, 1471–1492, <ext-link xlink:href="https://doi.org/10.5194/gmd-5-1471-2012" ext-link-type="DOI">10.5194/gmd-5-1471-2012</ext-link>, 2012.</mixed-citation></ref>
      <ref id="bib1.bibx21"><label>Hall et al.(2018)Hall, Ullmann, Prather, Flynn, Murray, Fiore, Correa, Strode, Steenrod, Lamarque, Guth, Josse, Flemming, Huijnen, Abraham, and Archibald</label><mixed-citation>Hall, S. R., Ullmann, K., Prather, M. J., Flynn, C. M., Murray, L. T., Fiore, A. M., Correa, G., Strode, S. A., Steenrod, S. D., Lamarque, J.-F., Guth, J., Josse, B., Flemming, J., Huijnen, V., Abraham, N. L., and Archibald, A. T.: Cloud impacts on photochemistry: building a climatology of photolysis rates from the Atmospheric Tomography mission, Atmos. Chem. Phys., 18, 16809–16828, <ext-link xlink:href="https://doi.org/10.5194/acp-18-16809-2018" ext-link-type="DOI">10.5194/acp-18-16809-2018</ext-link>, 2018.</mixed-citation></ref>
      <ref id="bib1.bibx22"><label>Hanisco et al.(2019)Hanisco, Hannun, St.Clair, and Wolfe</label><mixed-citation>Hanisco, T., Hannun, R., St.Clair, J., and Wolfe, G.: Atmospheric Tomography Mission (ATom)ATom: L2 Measurements of In Situ Airborne Formaldehyde (ISAF), ORNL Distributed Active Archive Center, <ext-link xlink:href="https://doi.org/10.3334/ORNLDAAC/1730" ext-link-type="DOI">10.3334/ORNLDAAC/1730</ext-link>, 2019.</mixed-citation></ref>
      <ref id="bib1.bibx23"><label>Hauglustaine and Ehhalt(2002)</label><mixed-citation>Hauglustaine, D. A. and Ehhalt, D. H.: A three-dimensional model of molecular hydrogen in the troposphere: TROPOSPHERIC MOLECULAR HYDROGEN, J. Geophys. Res.-Atmos., 107, ACH 4-1–ACH 4-16, <ext-link xlink:href="https://doi.org/10.1029/2001JD001156" ext-link-type="DOI">10.1029/2001JD001156</ext-link>, 2002.</mixed-citation></ref>
      <ref id="bib1.bibx24"><label>Hersbach et al.(2020)</label><mixed-citation>Hersbach, H., Bell, B., Berrisford, P., Hirahara, S., Horányi, A., Muñoz-Sabater, J., Nicolas, J., Peubey, C., Radu, R., Schepers, D., Simmons, A., Soci, C., Abdalla, S., Abellan, X., Balsamo, G., Bechtold, P., Biavati, G., Bidlot, J., Bonavita, M., De Chiara, G., Dahlgren, P., Dee, D., Diamantakis, M., Dragani, R., Flemming, J., Forbes, R., Fuentes, M., Geer, A., Haimberger, L., Healy, S., Hogan, R. J., Hólm, E., Janisková, M., Keeley, S., Laloyaux, P., Lopez, P., Lupu, C., Radnoti, G., De Rosnay, P., Rozum, I., Vamborg, F., Villaume, S., and Thépaut, J.: The ERA5 global reanalysis, Q. J. Roy. Meteor. Soc., 146, 1999–2049, <ext-link xlink:href="https://doi.org/10.1002/qj.3803" ext-link-type="DOI">10.1002/qj.3803</ext-link>, 2020.</mixed-citation></ref>
      <ref id="bib1.bibx25"><label>Hoesly et al.(2018)Hoesly, Smith, Feng, Klimont, Janssens-Maenhout, Pitkanen, Seibert, Vu, Andres, Bolt, Bond, Dawidowski, Kholod, Kurokawa, Li, Liu, Lu, Moura, O'Rourke, and Zhang</label><mixed-citation>Hoesly, R. M., Smith, S. J., Feng, L., Klimont, Z., Janssens-Maenhout, G., Pitkanen, T., Seibert, J. J., Vu, L., Andres, R. J., Bolt, R. M., Bond, T. C., Dawidowski, L., Kholod, N., Kurokawa, J.-I., Li, M., Liu, L., Lu, Z., Moura, M. C. P., O'Rourke, P. R., and Zhang, Q.: Historical (1750–2014) anthropogenic emissions of reactive gases and aerosols from the Community Emissions Data System (CEDS), Geosci. Model Dev., 11, 369–408, <ext-link xlink:href="https://doi.org/10.5194/gmd-11-369-2018" ext-link-type="DOI">10.5194/gmd-11-369-2018</ext-link>, 2018.</mixed-citation></ref>
      <ref id="bib1.bibx26"><label>IEA(2019)</label><mixed-citation>IEA: The Future of Hydrogen, Tech. rep., International Energy Agency, <uri>https://www.iea.org/reports/the-future-of-hydrogen</uri> (last access: 28 May 2026), 2019.</mixed-citation></ref>
      <ref id="bib1.bibx27"><label>IEA(2023)</label><mixed-citation>IEA: Global Hydrogen Review 2023, Tech. rep., International Energy Agency, <uri>https://www.iea.org/reports/global-hydrogen-review-2023</uri> (last access: 28 May 2026), 2023.</mixed-citation></ref>
      <ref id="bib1.bibx28"><label>Inness et al.(2019)Inness, Ades, Agustí-Panareda, Barré, Benedictow, Blechschmidt, Dominguez, Engelen, Eskes, Flemming, Huijnen, Jones, Kipling, Massart, Parrington, Peuch, Razinger, Remy, Schulz, and Suttie</label><mixed-citation>Inness, A., Ades, M., Agustí-Panareda, A., Barré, J., Benedictow, A., Blechschmidt, A.-M., Dominguez, J. J., Engelen, R., Eskes, H., Flemming, J., Huijnen, V., Jones, L., Kipling, Z., Massart, S., Parrington, M., Peuch, V.-H., Razinger, M., Remy, S., Schulz, M., and Suttie, M.: The CAMS reanalysis of atmospheric composition, Atmos. Chem. Phys., 19, 3515–3556, <ext-link xlink:href="https://doi.org/10.5194/acp-19-3515-2019" ext-link-type="DOI">10.5194/acp-19-3515-2019</ext-link>, 2019.</mixed-citation></ref>
      <ref id="bib1.bibx29"><label>Jordan and Steinberg(2011)</label><mixed-citation>Jordan, A. and Steinberg, B.: Calibration of atmospheric hydrogen measurements, Atmos. Meas. Tech., 4, 509–521, <ext-link xlink:href="https://doi.org/10.5194/amt-4-509-2011" ext-link-type="DOI">10.5194/amt-4-509-2011</ext-link>, 2011.</mixed-citation></ref>
      <ref id="bib1.bibx30"><label>Jordan et al.(2024)Jordan, Moossen, Rothe, Brand, Heimann, Zaehle, and Rödenbeck</label><mixed-citation>Jordan, A., Moossen, H., Rothe, M., Brand, W., Heimann, M., Zaehle, S., and Rödenbeck, C.: Atmospheric flask sampling program of MPI-BGC, version 2024.1,  Edmond, <ext-link xlink:href="https://doi.org/10.17617/3.S3KHMS" ext-link-type="DOI">10.17617/3.S3KHMS</ext-link>, 2024.</mixed-citation></ref>
      <ref id="bib1.bibx31"><label>Krol et al.(2005)Krol, Houweling, Bregman, Van Den Broek, Segers, Van Velthoven, Peters, Dentener, and Bergamaschi</label><mixed-citation>Krol, M., Houweling, S., Bregman, B., van den Broek, M., Segers, A., van Velthoven, P., Peters, W., Dentener, F., and Bergamaschi, P.: The two-way nested global chemistry-transport zoom model TM5: algorithm and applications, Atmos. Chem. Phys., 5, 417–432, <ext-link xlink:href="https://doi.org/10.5194/acp-5-417-2005" ext-link-type="DOI">10.5194/acp-5-417-2005</ext-link>, 2005.</mixed-citation></ref>
      <ref id="bib1.bibx32"><label>Krummel et al.(2025)Krummel, Langenfelds, Guerette, and Loh</label><mixed-citation>Krummel, P., Langenfelds, R., Guerette, E., and Loh, Z.: Atmospheric H2 at sites from the Commonwealth Scientific and Industrial Research Organisation (CSIRO) global monitoring network, World Data Centre for Greenhouse Gases, <ext-link xlink:href="https://doi.org/10.50849/WDCGG_H2_ALL_2025" ext-link-type="DOI">10.50849/WDCGG_H2_ALL_2025</ext-link>, 2025.</mixed-citation></ref>
      <ref id="bib1.bibx33"><label>McCarthy(2004)</label><mixed-citation>McCarthy, M. C.: The hydrogen isotopic composition of water vapor entering the stratosphere inferred from high-precision measurements of CHD3 and HD, J. Geophys. Res., 109, D07304, <ext-link xlink:href="https://doi.org/10.1029/2003JD004003" ext-link-type="DOI">10.1029/2003JD004003</ext-link>, 2004.</mixed-citation></ref>
      <ref id="bib1.bibx34"><label>McKain et al.(2022)McKain, Sweeney, Crotwell, Crotwell, Higgs, Legard, Madronich, Miller, Moglia, Mund, Neff, Wolter, and NOAA Global Monitoring Laboratory</label><mixed-citation>McKain, K., Sweeney, C., Crotwell, A., Crotwell, M., Higgs, J., Legard, T., Madronich, M., Miller, J., Moglia, E., Mund, J., Neff, D., Wolter, S., and NOAA Global Monitoring Laboratory: NOAA Global Greenhouse Gas Reference Network Flask-Air PFP Sample Measurements of CO<sub>2</sub>, CH<sub>4</sub>, CO, N<sub>2</sub>O, H<sub>2</sub>, SF<sub>6</sub> and isotopic ratios collected from aircraft vertical profiles, NOAA Global Monitoring Laboratory, <ext-link xlink:href="https://doi.org/10.15138/39HR-9N34" ext-link-type="DOI">10.15138/39HR-9N34</ext-link>, 2022.</mixed-citation></ref>
      <ref id="bib1.bibx35"><label>Montzka et al.(2019)Montzka, Moore, and Sweeney</label><mixed-citation>Montzka, S., Moore, F., and Sweeney, C.: Atmospheric Tomography Mission (ATom)ATom: L2 Measurements from the Programmable Flask Package (PFP) Whole Air Sampler, ORNL Distributed Active Archive Center, <ext-link xlink:href="https://doi.org/10.3334/ORNLDAAC/1746" ext-link-type="DOI">10.3334/ORNLDAAC/1746</ext-link>,  2019.</mixed-citation></ref>
      <ref id="bib1.bibx36"><label>Naik et al.(2013)Naik, Voulgarakis, Fiore, Horowitz, Lamarque, Lin, Prather, Young, Bergmann, Cameron-Smith, Cionni, Collins, Dalsøren, Doherty, Eyring, Faluvegi, Folberth, Josse, Lee, MacKenzie, Nagashima, Van Noije, Plummer, Righi, Rumbold, Skeie, Shindell, Stevenson, Strode, Sudo, Szopa, and Zeng</label><mixed-citation>Naik, V., Voulgarakis, A., Fiore, A. M., Horowitz, L. W., Lamarque, J.-F., Lin, M., Prather, M. J., Young, P. J., Bergmann, D., Cameron-Smith, P. J., Cionni, I., Collins, W. J., Dalsøren, S. B., Doherty, R., Eyring, V., Faluvegi, G., Folberth, G. A., Josse, B., Lee, Y. H., MacKenzie, I. A., Nagashima, T., van Noije, T. P. C., Plummer, D. A., Righi, M., Rumbold, S. T., Skeie, R., Shindell, D. T., Stevenson, D. S., Strode, S., Sudo, K., Szopa, S., and Zeng, G.: Preindustrial to present-day changes in tropospheric hydroxyl radical and methane lifetime from the Atmospheric Chemistry and Climate Model Intercomparison Project (ACCMIP), Atmos. Chem. Phys., 13, 5277–5298, <ext-link xlink:href="https://doi.org/10.5194/acp-13-5277-2013" ext-link-type="DOI">10.5194/acp-13-5277-2013</ext-link>, 2013.</mixed-citation></ref>
      <ref id="bib1.bibx37"><label>Naus et al.(2021)Naus, Montzka, Patra, and Krol</label><mixed-citation>Naus, S., Montzka, S. A., Patra, P. K., and Krol, M. C.: A three-dimensional-model inversion of methyl chloroform to constrain the atmospheric oxidative capacity, Atmos. Chem. Phys., 21, 4809–4824, <ext-link xlink:href="https://doi.org/10.5194/acp-21-4809-2021" ext-link-type="DOI">10.5194/acp-21-4809-2021</ext-link>, 2021.</mixed-citation></ref>
      <ref id="bib1.bibx38"><label>NOAA Global Monitoring Laboratory(2026)</label><mixed-citation>NOAA Global Monitoring Laboratory: Curve Fitting Methods Applied to Time Series (CCGCRV), <uri>https://gml.noaa.gov/ccgg/mbl/crvfit/</uri> (last access: 28 May 2026), 2026.</mixed-citation></ref>
      <ref id="bib1.bibx39"><label>Novelli et al.(1999)Novelli, Lang, Masarie, Hurst, Myers, and Elkins</label><mixed-citation>Novelli, P., Lang, P., Masarie, K., Hurst, D., Myers, R., and Elkins, J.: Molecular hydrogen in the troposphere: Global distribution and budget, J. Geophys. Res., 104, 30427–30444, <ext-link xlink:href="https://doi.org/10.1029/1999JD900788" ext-link-type="DOI">10.1029/1999JD900788</ext-link>, 1999.</mixed-citation></ref>
      <ref id="bib1.bibx40"><label>Ocko and Hamburg(2022)</label><mixed-citation>Ocko, I. B. and Hamburg, S. P.: Climate consequences of hydrogen emissions, Atmos. Chem. Phys., 22, 9349–9368, <ext-link xlink:href="https://doi.org/10.5194/acp-22-9349-2022" ext-link-type="DOI">10.5194/acp-22-9349-2022</ext-link>, 2022.</mixed-citation></ref>
      <ref id="bib1.bibx41"><label>Ouyang et al.(2025)Ouyang, Jackson, Saunois, Canadell, Zhao, Morfopoulos, Krummel, Patra, Peters, Dennison, Gasser, Archibald, Arora, Baudoin, Chandra, Ciais, Davis, Feron, Guo, Hauglustaine, Jones, Jones, Kato, Kennedy, Knauer, Lienert, Lombardozzi, Melton, Nabel, O’Sullivan, Pétron, Poulter, Rogelj, Sandoval Calle, Smith, Suntharalingam, Tian, Wang, and Wiltshire</label><mixed-citation>Ouyang, Z., Jackson, R. B., Saunois, M., Canadell, J. G., Zhao, Y., Morfopoulos, C., Krummel, P. B., Patra, P. K., Peters, G. P., Dennison, F., Gasser, T., Archibald, A. T., Arora, V., Baudoin, G., Chandra, N., Ciais, P., Davis, S. J., Feron, S., Guo, F., Hauglustaine, D., Jones, C. D., Jones, M. W., Kato, E., Kennedy, D., Knauer, J., Lienert, S., Lombardozzi, D., Melton, J. R., Nabel, J. E., O'Sullivan, M., Pétron, G., Poulter, B., Rogelj, J., Sandoval Calle, D., Smith, P., Suntharalingam, P., Tian, H., Wang, C., and Wiltshire, A.: The global hydrogen budget, Nature, 648, 616–624, <ext-link xlink:href="https://doi.org/10.1038/s41586-025-09806-1" ext-link-type="DOI">10.1038/s41586-025-09806-1</ext-link>, 2025.</mixed-citation></ref>
      <ref id="bib1.bibx42"><label>Patra et al.(2021)Patra, Krol, Prinn, Takigawa, Mühle, Montzka, Lal, Yamashita, Naus, Chandra, Weiss, Krummel, Fraser, O'Doherty, and Elkins</label><mixed-citation>Patra, P. K., Krol, M. C., Prinn, R. G., Takigawa, M., Mühle, J., Montzka, S. A., Lal, S., Yamashita, Y., Naus, S., Chandra, N., Weiss, R. F., Krummel, P. B., Fraser, P. J., O'Doherty, S., and Elkins, J. W.: Methyl Chloroform Continues to Constrain the Hydroxyl (OH) Variability in the Troposphere, J. Geophys. Res.-Atmos., 126, e2020JD033862, <ext-link xlink:href="https://doi.org/10.1029/2020JD033862" ext-link-type="DOI">10.1029/2020JD033862</ext-link>, 2021.</mixed-citation></ref>
      <ref id="bib1.bibx43"><label>Paulot et al.(2021)Paulot, Paynter, Naik, Malyshev, Menzel, and Horowitz</label><mixed-citation>Paulot, F., Paynter, D., Naik, V., Malyshev, S., Menzel, R., and Horowitz, L. W.: Global modeling of hydrogen using GFDL-AM4.1: Sensitivity of soil removal and radiative forcing, Int. J. Hydrogen Energ., 46, 13446–13460, <ext-link xlink:href="https://doi.org/10.1016/j.ijhydene.2021.01.088" ext-link-type="DOI">10.1016/j.ijhydene.2021.01.088</ext-link>, 2021.</mixed-citation></ref>
      <ref id="bib1.bibx44"><label>Percival and Sander(2026)</label><mixed-citation>Percival, C. J. and Sander, S. P.: Chemical Kinetics and Photochemical Data for Use in Atmospheric Studies Evaluation Number 20, Tech. rep., Jet Propulsion Laboratory, California Institute of Technology, <ext-link xlink:href="https://doi.org/10.48577/JPL.KFQ0TZ" ext-link-type="DOI">10.48577/JPL.KFQ0TZ</ext-link>, 2026.</mixed-citation></ref>
      <ref id="bib1.bibx45"><label>Pieterse(2013)</label><mixed-citation> Pieterse, G.: Modelling the global tropospheric molecular hydrogen cycle, Ph.D. thesis, Utrecht University, ISBN 978-94-6191-559-7, 2013.</mixed-citation></ref>
      <ref id="bib1.bibx46"><label>Pieterse et al.(2011)Pieterse, Krol, Batenburg, Steele, Krummel, Langenfelds, and Röckmann</label><mixed-citation>Pieterse, G., Krol, M. C., Batenburg, A. M., Steele, L. P., Krummel, P. B., Langenfelds, R. L., and Röckmann, T.: Global modelling of H2 mixing ratios and isotopic compositions with the TM5 model, Atmos. Chem. Phys., 11, 7001–7026, <ext-link xlink:href="https://doi.org/10.5194/acp-11-7001-2011" ext-link-type="DOI">10.5194/acp-11-7001-2011</ext-link>, 2011.</mixed-citation></ref>
      <ref id="bib1.bibx47"><label>Pieterse et al.(2013)Pieterse, Krol, Batenburg, Brenninkmeijer, Popa, O'Doherty, Grant, Steele, Krummel, Langenfelds, Wang, Vermeulen, Schmidt, Yver, Jordan, Engel, Fisher, Lowry, Nisbet, and Röckmann</label><mixed-citation>Pieterse, G., Krol, M., Batenburg, A., Brenninkmeijer, C., Popa, E., O'Doherty, S., Grant, A., Steele, L., Krummel, P., Langenfelds, R., Wang, H., Vermeulen, A., Schmidt, M., Yver, C., Jordan, A., Engel, A., Fisher, R., Lowry, D., Nisbet, E., and Röckmann, T.: Reassessing the variability in atmospheric H2 using the two-way nested TM5 model, J. Geophys. Res., <ext-link xlink:href="https://doi.org/10.1002/jgrd.50204" ext-link-type="DOI">10.1002/jgrd.50204</ext-link>, 2013.</mixed-citation></ref>
      <ref id="bib1.bibx48"><label>Pison et al.(2009)Pison, Bousquet, Chevallier, Szopa, and Hauglustaine</label><mixed-citation>Pison, I., Bousquet, P., Chevallier, F., Szopa, S., and Hauglustaine, D.: Multi-species inversion of CH<sub>4</sub>, CO and H<sub>2</sub> emissions from surface measurements, Atmos. Chem. Phys., 9, 5281–5297, <ext-link xlink:href="https://doi.org/10.5194/acp-9-5281-2009" ext-link-type="DOI">10.5194/acp-9-5281-2009</ext-link>, 2009.</mixed-citation></ref>
      <ref id="bib1.bibx49"><label>Prather et al.(2012)Prather, Holmes, and Hsu</label><mixed-citation>Prather, M. J., Holmes, C. D., and Hsu, J.: Reactive greenhouse gas scenarios: Systematic exploration of uncertainties and the role of atmospheric chemistry, Geophys. Res. Lett., 39, 2012GL051440, <ext-link xlink:href="https://doi.org/10.1029/2012GL051440" ext-link-type="DOI">10.1029/2012GL051440</ext-link>, 2012.</mixed-citation></ref>
      <ref id="bib1.bibx50"><label>Price et al.(2007)Price, Jaeglé, Rice, Quay, Novelli, and Gammon</label><mixed-citation>Price, H., Jaeglé, L., Rice, A., Quay, P., Novelli, P. C., and Gammon, R.: Global budget of molecular hydrogen and its deuterium content: Constraints from ground station, cruise, and aircraft observations, J. Geophys. Res., 112, D22108, <ext-link xlink:href="https://doi.org/10.1029/2006JD008152" ext-link-type="DOI">10.1029/2006JD008152</ext-link>, 2007.</mixed-citation></ref>
      <ref id="bib1.bibx51"><label>Prinn et al.(2025)Prinn, Weiss, Arduini, Choi, Engel, Fraser, Ganesan, Harth, Hermansen, Kim, Krummel, Loh, Lunder, Maione, Manning, Mitrevski, Mühle, O'Doherty, Park, Pitt, Reimann, Rigby, Saito, Salameh, Schmidt, Simmonds, Stanley, Stavert, Steele, Vollmer, Wagenhäuser, Wang, Wenger, Western, Yao, Young, Zhou, and Zhu</label><mixed-citation>Prinn, R., Weiss, R., Arduini, J., Choi, H., Engel, A., Fraser, P., Ganesan, A., Harth, C., Hermansen, O., Kim, J., Krummel, P., Loh, Z., Lunder, C., Maione, M., Manning, A., Mitrevski, B., Mühle, J., O'Doherty, S., Park, S., Pitt, J., Reimann, S., Rigby, M., Saito, T., Salameh, P., Schmidt, R., Simmonds, P., Stanley, K., Stavert, A., Steele, P., Vollmer, M., Wagenhäuser, T., Wang, H., Wenger, A., Western, L., Yao, B., Young, D., Zhou, L., and Zhu, L.: The dataset of in-situ measurements of chemically and radiatively important atmospheric gases from the Advanced Global Atmospheric Gas Experiment (AGAGE) and affiliated stations, NASA Langley Research Center (LaRC) Data Host Facility (DHF),  <ext-link xlink:href="https://doi.org/10.60718/75D7-QE84" ext-link-type="DOI">10.60718/75D7-QE84</ext-link>, 2025.</mixed-citation></ref>
      <ref id="bib1.bibx52"><label>Pétron et al.(2024)Pétron, Crotwell, Mund, Crotwell, Mefford, Thoning, Hall, Kitzis, Madronich, Moglia, Neff, Wolter, Jordan, Krummel, Langenfelds, and Patterson</label><mixed-citation>Pétron, G., Crotwell, A. M., Mund, J., Crotwell, M., Mefford, T., Thoning, K., Hall, B., Kitzis, D., Madronich, M., Moglia, E., Neff, D., Wolter, S., Jordan, A., Krummel, P., Langenfelds, R., and Patterson, J.: Atmospheric H2 observations from the NOAA Cooperative Global Air Sampling Network, Atmos. Meas. Tech., 17, 4803–4823, <ext-link xlink:href="https://doi.org/10.5194/amt-17-4803-2024" ext-link-type="DOI">10.5194/amt-17-4803-2024</ext-link>, 2024.</mixed-citation></ref>
      <ref id="bib1.bibx53"><label>Rahn et al.(2003)Rahn, Eiler, Boering, Wennberg, McCarthy, Tyler, Schauffler, Donnelly, and Atlas</label><mixed-citation>Rahn, T., Eiler, J. M., Boering, K. A., Wennberg, P. O., McCarthy, M. C., Tyler, S., Schauffler, S., Donnelly, S., and Atlas, E.: Extreme deuterium enrichment in stratospheric hydrogen and the global atmospheric budget of H2, Nature, 424, 918–921, <ext-link xlink:href="https://doi.org/10.1038/nature01917" ext-link-type="DOI">10.1038/nature01917</ext-link>, 2003.</mixed-citation></ref>
      <ref id="bib1.bibx54"><label>Rhee et al.(2006)Rhee, Brenninkmeijer, and Röckmann</label><mixed-citation>Rhee, T. S., Brenninkmeijer, C. A. M., and Röckmann, T.: The overwhelming role of soils in the global atmospheric hydrogen cycle, Atmos. Chem. Phys., 6, 1611–1625, <ext-link xlink:href="https://doi.org/10.5194/acp-6-1611-2006" ext-link-type="DOI">10.5194/acp-6-1611-2006</ext-link>, 2006.</mixed-citation></ref>
      <ref id="bib1.bibx55"><label>Rodell et al.(2004)Rodell, Houser, Jambor, Gottschalck, Mitchell, Meng, Arsenault, Cosgrove, Radakovich, Bosilovich, Entin, Walker, Lohmann, and Toll</label><mixed-citation>Rodell, M., Houser, P. R., Jambor, U., Gottschalck, J., Mitchell, K., Meng, C.-J., Arsenault, K., Cosgrove, B., Radakovich, J., Bosilovich, M., Entin, J. K., Walker, J. P., Lohmann, D., and Toll, D.: The Global Land Data Assimilation System, B. Am. Meteor. Soc., 85, 381–394, <ext-link xlink:href="https://doi.org/10.1175/BAMS-85-3-381" ext-link-type="DOI">10.1175/BAMS-85-3-381</ext-link>, 2004.</mixed-citation></ref>
      <ref id="bib1.bibx56"><label>Röckmann et al.(2003)Röckmann, Rhee, and Engel</label><mixed-citation>Röckmann, T., Rhee, T. S., and Engel, A.: Heavy hydrogen in the stratosphere, Atmos. Chem. Phys., 3, 2015–2023, <ext-link xlink:href="https://doi.org/10.5194/acp-3-2015-2003" ext-link-type="DOI">10.5194/acp-3-2015-2003</ext-link>, 2003.</mixed-citation></ref>
      <ref id="bib1.bibx57"><label>Sand et al.(2023)Sand, Skeie, Sandstad, Krishnan, Myhre, Bryant, Derwent, Hauglustaine, Paulot, Prather, and Stevenson</label><mixed-citation>Sand, M., Skeie, R. B., Sandstad, M., Krishnan, S., Myhre, G., Bryant, H., Derwent, R., Hauglustaine, D., Paulot, F., Prather, M., and Stevenson, D.: A multi-model assessment of the Global Warming Potential of hydrogen, Commun. Earth  Environ., 4, 1–12, <ext-link xlink:href="https://doi.org/10.1038/s43247-023-00857-8" ext-link-type="DOI">10.1038/s43247-023-00857-8</ext-link>,  2023.</mixed-citation></ref>
      <ref id="bib1.bibx58"><label>Sand et al.(2026)Sand, Ocko, Sun, Skeie, Hauglustaine, Warwick, Stevenson, Paulot, Myhre, Archibald, Sandstad, Krishnan, and Prather</label><mixed-citation>Sand, M., Ocko, I. B., Sun, T., Skeie, R. B., Hauglustaine, D. A., Warwick, N. J., Stevenson, D. S., Paulot, F., Myhre, G., Archibald, A. T., Sandstad, M., Krishnan, S., and Prather, M.: Climate impacts of hydrogen emissions, Nat. Rev. Earth Environ., <ext-link xlink:href="https://doi.org/10.1038/s43017-026-00792-0" ext-link-type="DOI">10.1038/s43017-026-00792-0</ext-link>, 2026.</mixed-citation></ref>
      <ref id="bib1.bibx59"><label>Sander et al.(2006)Sander, Friedl, Ravishankara, Golden, Kolb, Kurylo, Molina, Moortgat, Keller-Rudek, Finlayson-Pitts, Wine, Huie, and Orkin</label><mixed-citation>Sander, S. P., Friedl, R. R., Ravishankara, A. R., Golden, D. M., Kolb, C. E., Kurylo, M. J., Molina, M. J., Moortgat, G. K., Keller-Rudek, H., Finlayson-Pitts, B. J., Wine, P. H., Huie, R. E., and Orkin, V. L.: Chemical Kinetics and Photochemical Data for Use in Atmospheric Studies: Evaluation No. 15, Tech. Rep. JPL Publication 06-2, Jet Propulsion Laboratory, California Institute of Technology, <uri>https://science.jpl.nasa.gov/documents/1493/JPL_15_AllInOne.pdf</uri> (last access: 28 May 2026), 2006.</mixed-citation></ref>
      <ref id="bib1.bibx60"><label>Sanderson et al.(2003)Sanderson, Collins, Derwent, and Johnson</label><mixed-citation>Sanderson, M. G., Collins, W. J., Derwent, R. G., and Johnson, C. E.: Simulation of Global Hydrogen Levels Using a Lagrangian Three-Dimensional Model, J. Atmos. Chem., 46, 15–28, <ext-link xlink:href="https://doi.org/10.1023/A:1024824223232" ext-link-type="DOI">10.1023/A:1024824223232</ext-link>, 2003.</mixed-citation></ref>
      <ref id="bib1.bibx61"><label>Smith-Downey et al.(2006)Smith-Downey, Randerson, and Eiler</label><mixed-citation>Smith-Downey, N. V., Randerson, J. T., and Eiler, J. M.: Temperature and moisture dependence of soil H<sub>2</sub> uptake measured in the laboratory, Geophys. Res. Lett., 33, 2006GL026749, <ext-link xlink:href="https://doi.org/10.1029/2006GL026749" ext-link-type="DOI">10.1029/2006GL026749</ext-link>, 2006.</mixed-citation></ref>
      <ref id="bib1.bibx62"><label>Soulie et al.(2024)Soulie, Granier, Darras, Zilbermann, Doumbia, Guevara, Jalkanen, Keita, Liousse, Crippa, Guizzardi, Hoesly, and Smith</label><mixed-citation>Soulie, A., Granier, C., Darras, S., Zilbermann, N., Doumbia, T., Guevara, M., Jalkanen, J.-P., Keita, S., Liousse, C., Crippa, M., Guizzardi, D., Hoesly, R., and Smith, S. J.: Global anthropogenic emissions (CAMS-GLOB-ANT) for the Copernicus Atmosphere Monitoring Service simulations of air quality forecasts and reanalyses, Earth Syst. Sci. Data, 16, 2261–2279, <ext-link xlink:href="https://doi.org/10.5194/essd-16-2261-2024" ext-link-type="DOI">10.5194/essd-16-2261-2024</ext-link>, 2024.</mixed-citation></ref>
      <ref id="bib1.bibx63"><label>Stone et al.(2026)Stone, Chen, Solomon, Western, Krummel, Pétron, Mühle, and O'Doherty</label><mixed-citation>Stone, K., Chen, C., Solomon, S., Western, L. M., Krummel, P. B., Pétron, G., Mühle, J., and O'Doherty, S.: Constraining the atmospheric hydrogen oxidation and soil sink seasonal cycles using HFC-152a, Atmos. Chem. Phys., 26, 10241–10253, <ext-link xlink:href="https://doi.org/10.5194/acp-26-10241-2026" ext-link-type="DOI">10.5194/acp-26-10241-2026</ext-link>, 2026.</mixed-citation></ref>
      <ref id="bib1.bibx64"><label>Stroo(2026)</label><mixed-citation>Stroo, F.:  Posterior and Prior fluxes and model code for the 2003–2023 H2 inversions, Zenodo [data set], <ext-link xlink:href="https://doi.org/10.5281/zenodo.20427212" ext-link-type="DOI">10.5281/zenodo.20427212</ext-link>, 2026.</mixed-citation></ref>
      <ref id="bib1.bibx65"><label>Sweeney et al.(2015)Sweeney, Karion, Wolter, Newberger, Guenther, Higgs, Andrews, Lang, Neff, Dlugokencky, Miller, Montzka, Miller, Masarie, Biraud, Novelli, Crotwell, Crotwell, Thoning, and Tans</label><mixed-citation>Sweeney, C., Karion, A., Wolter, S., Newberger, T., Guenther, D., Higgs, J. A., Andrews, A. E., Lang, P. M., Neff, D., Dlugokencky, E., Miller, J. B., Montzka, S. A., Miller, B. R., Masarie, K. A., Biraud, S. C., Novelli, P. C., Crotwell, M., Crotwell, A. M., Thoning, K., and Tans, P. P.: Seasonal climatology of CO<sub>2</sub> across North America from aircraft measurements in the NOAA/ESRL Global Greenhouse Gas Reference Network, J. Geophys. Res.-Atmos., 120, 5155–5190, <ext-link xlink:href="https://doi.org/10.1002/2014JD022591" ext-link-type="DOI">10.1002/2014JD022591</ext-link>, 2015.</mixed-citation></ref>
      <ref id="bib1.bibx66"><label>Tardito Chaudhri and Stevenson(2025)</label><mixed-citation>Tardito Chaudhri, A. K. and Stevenson, D. S.: Soil deposition of atmospheric hydrogen constrained using planetary-scale observations, Atmos. Chem. Phys., 25, 7369–7385, <ext-link xlink:href="https://doi.org/10.5194/acp-25-7369-2025" ext-link-type="DOI">10.5194/acp-25-7369-2025</ext-link>, 2025.</mixed-citation></ref>
      <ref id="bib1.bibx67"><label>Thompson et al.(2022)Thompson, Wofsy, Prather, Newman, Hanisco, Ryerson, Fahey, Apel, Brock, Brune, Froyd, Katich, Nicely, Peischl, Ray, Veres, Wang, Allen, Asher, Bian, Blake, Bourgeois, Budney, Bui, Butler, Campuzano-Jost, Chang, Chin, Commane, Correa, Crounse, Daube, Dibb, DiGangi, Diskin, Dollner, Elkins, Fiore, Flynn, Guo, Hall, Hannun, Hills, Hintsa, Hodzic, Hornbrook, Huey, Jimenez, Keeling, Kim, Kupc, Lacey, Lait, Lamarque, Liu, McKain, Meinardi, Miller, Montzka, Moore, Morgan, Murphy, Murray, Nault, Neuman, Nguyen, Gonzalez, Rollins, Rosenlof, Sargent, Schill, Schwarz, Clair, Steenrod, Stephens, Strahan, Strode, Sweeney, Thames, Ullmann, Wagner, Weber, Weinzierl, Wennberg, Williamson, Wolfe, and Zeng</label><mixed-citation>Thompson, C. R., Wofsy, S. C., Prather, M. J., Newman, P. A., Hanisco, T. F., Ryerson, T. B., Fahey, D. W., Apel, E. C., Brock, C. A., Brune, W. H., Froyd, K., Katich, J. M., Nicely, J. M., Peischl, J., Ray, E., Veres, P. R., Wang, S., Allen, H. M., Asher, E., Bian, H., Blake, D., Bourgeois, I., Budney, J., Bui, T. P., Butler, A., Campuzano-Jost, P., Chang, C., Chin, M., Commane, R., Correa, G., Crounse, J. D., Daube, B., Dibb, J. E., DiGangi, J. P., Diskin, G. S., Dollner, M., Elkins, J. W., Fiore, A. M., Flynn, C. M., Guo, H., Hall, S. R., Hannun, R. A., Hills, A., Hintsa, E. J., Hodzic, A., Hornbrook, R. S., Huey, L. G., Jimenez, J. L., Keeling, R. F., Kim, M. J., Kupc, A., Lacey, F., Lait, L. R., Lamarque, J.-F., Liu, J., McKain, K., Meinardi, S., Miller, D. O., Montzka, S. A., Moore, F. L., Morgan, E. J., Murphy, D. M., Murray, L. T., Nault, B. A., Neuman, J. A., Nguyen, L., Gonzalez, Y., Rollins, A., Rosenlof, K., Sargent, M., Schill, G., Schwarz, J. P., Clair, J. M. S., Steenrod, S. D., Stephens, B. B., Strahan, S. E., Strode, S. A., Sweeney, C., Thames, A. B., Ullmann, K., Wagner, N., Weber, R., Weinzierl, B., Wennberg, P. O., Williamson, C. J., Wolfe, G. M., and Zeng, L.: The NASA Atmospheric Tomography (ATom) Mission: Imaging the Chemistry of the Global Atmosphere, B. Am. Meteor. Soc., 103, E761–E790, <ext-link xlink:href="https://doi.org/10.1175/BAMS-D-20-0315.1" ext-link-type="DOI">10.1175/BAMS-D-20-0315.1</ext-link>, 2022.</mixed-citation></ref>
      <ref id="bib1.bibx68"><label>Thoning et al.(1989)Thoning, Tans, and Komhyr</label><mixed-citation>Thoning, K. W., Tans, P. P., and Komhyr, W. D.: Atmospheric carbon dioxide at Mauna Loa Observatory: 2. Analysis of the NOAA GMCC data, 1974–1985, J. Geophys. Res.-Atmos., 94, 8549–8565, <ext-link xlink:href="https://doi.org/10.1029/JD094iD06p08549" ext-link-type="DOI">10.1029/JD094iD06p08549</ext-link>, 1989.</mixed-citation></ref>
      <ref id="bib1.bibx69"><label>Tolk et al.(2011)Tolk, Dolman, Meesters, and Peters</label><mixed-citation>Tolk, L. F., Dolman, A. J., Meesters, A. G. C. A., and Peters, W.: A comparison of different inverse carbon flux estimation approaches for application on a regional domain, Atmos. Chem. Phys., 11, 10349–10365, <ext-link xlink:href="https://doi.org/10.5194/acp-11-10349-2011" ext-link-type="DOI">10.5194/acp-11-10349-2011</ext-link>, 2011.</mixed-citation></ref>
      <ref id="bib1.bibx70"><label>Trapani et al.(2025)Trapani, Marocco, Gandiglio, and Santarelli</label><mixed-citation>Trapani, D., Marocco, P., Gandiglio, M., and Santarelli, M.: Hydrogen leakages across the supply chain: Current estimates and future scenarios, Int. J. Hydrogen Energ., 145, 1084–1095, <ext-link xlink:href="https://doi.org/10.1016/j.ijhydene.2025.06.103" ext-link-type="DOI">10.1016/j.ijhydene.2025.06.103</ext-link>, 2025.</mixed-citation></ref>
      <ref id="bib1.bibx71"><label>Van Der Laan-Luijkx et al.(2017)Van Der Laan-Luijkx, Van Der Velde, Van Der Veen, Tsuruta, Stanislawska, Babenhauserheide, Zhang, Liu, He, Chen, Masarie, Krol, and Peters</label><mixed-citation>van der Laan-Luijkx, I. T., van der Velde, I. R., van der Veen, E., Tsuruta, A., Stanislawska, K., Babenhauserheide, A., Zhang, H. F., Liu, Y., He, W., Chen, H., Masarie, K. A., Krol, M. C., and Peters, W.: The CarbonTracker Data Assimilation Shell (CTDAS) v1.0: implementation and global carbon balance 2001–2015, Geosci. Model Dev., 10, 2785–2800, <ext-link xlink:href="https://doi.org/10.5194/gmd-10-2785-2017" ext-link-type="DOI">10.5194/gmd-10-2785-2017</ext-link>, 2017.</mixed-citation></ref>
      <ref id="bib1.bibx72"><label>Van Der Werf et al.(2025)Van Der Werf, Randerson, Van Wees, Chen, Giglio, Hall, Roland, Mu, Binte Shahid, Barsanti, Yokelson, and Morton</label><mixed-citation>Van Der Werf, G. R., Randerson, J. T., Van Wees, D., Chen, Y., Giglio, L., Hall, J., Roland, V., Mu, M., Binte Shahid, S., Barsanti, K. C., Yokelson, R., and Morton, D. C.: Landscape fire emissions from the 5th version of the Global Fire Emissions Database (GFED5), Sci. Data, 12, 1870, <ext-link xlink:href="https://doi.org/10.1038/s41597-025-06127-w" ext-link-type="DOI">10.1038/s41597-025-06127-w</ext-link>, 2025.</mixed-citation></ref>
      <ref id="bib1.bibx73"><label>Vollmer et al.(2010)Vollmer, Walter, Bond, Soltic, and Röckmann</label><mixed-citation>Vollmer, M. K., Walter, S., Bond, S. W., Soltic, P., and Röckmann, T.: Molecular hydrogen (H2) emissions and their isotopic signatures (H/D) from a motor vehicle: implications on atmospheric H2, Atmos. Chem. Phys., 10, 5707–5718, <ext-link xlink:href="https://doi.org/10.5194/acp-10-5707-2010" ext-link-type="DOI">10.5194/acp-10-5707-2010</ext-link>, 2010.</mixed-citation></ref>
      <ref id="bib1.bibx74"><label>Walter et al.(2016)Walter, Kock, Steinhoff, Fiedler, Fietzek, Kaiser, Krol, Popa, Chen, Tanhua, and Röckmann</label><mixed-citation>Walter, S., Kock, A., Steinhoff, T., Fiedler, B., Fietzek, P., Kaiser, J., Krol, M., Popa, M. E., Chen, Q., Tanhua, T., and Röckmann, T.: Isotopic evidence for biogenic molecular hydrogen production in the Atlantic Ocean, Biogeosciences, 13, 323–340, <ext-link xlink:href="https://doi.org/10.5194/bg-13-323-2016" ext-link-type="DOI">10.5194/bg-13-323-2016</ext-link>, 2016.</mixed-citation></ref>
      <ref id="bib1.bibx75"><label>Williams and Van Noije(2008)</label><mixed-citation> Williams, J. and Van Noije, T.: On the upgrading of the modified Carbon Bond Mechanism IV for use in global Chemistry Transport Models, Scientific Report WR-08-02, KNMI, 2008.</mixed-citation></ref>
      <ref id="bib1.bibx76"><label>Williams et al.(2017)Williams, Boersma, Le Sager, and Verstraeten</label><mixed-citation>Williams, J. E., Boersma, K. F., Le Sager, P., and Verstraeten, W. W.: The high-resolution version of TM5-MP for optimized satellite retrievals: description and validation, Geosci. Model Dev., 10, 721–750, <ext-link xlink:href="https://doi.org/10.5194/gmd-10-721-2017" ext-link-type="DOI">10.5194/gmd-10-721-2017</ext-link>, 2017.</mixed-citation></ref>
      <ref id="bib1.bibx77"><label>Wofsy et al.(2017)Wofsy, Daube, Jimenez-Pizarro, Kort, Pittman, Park, Commane, Xiang, Santoni, Jacob, Fisher, Pickett-Heaps, Wang, Wecht, Wang, Stephens, Shertz, Watt, Romashkin, Campos, Haggerty, Cooper, Rogers, Beaton, Hendershot, Elkins, Fahey, Gao, Schwarz, Moore, Montzka, Perring, Hurst, Miller, Sweeney, Oltmans, Hintsa, Nance, Dutton, Watts, Spackman, Rosenlof, Ray, Hall, Zondlo, Diao, Keeling, Bent, Atlas, Lueb, and (Deceased)</label><mixed-citation>Wofsy, S., Daube, B., Jimenez-Pizarro, R., Kort, E., Pittman, J., Park, S., Commane, R., Xiang, B., Santoni, G., Jacob, D., Fisher, J., Pickett-Heaps, C., Wang, H., Wecht, K., Wang, Q., Stephens, B., Shertz, S., Watt, A., Romashkin, P., Campos, T., Haggerty, J., Cooper, W., Rogers, D., Beaton, S., Hendershot, R., Elkins, J., Fahey, D., Gao, R., Schwarz, J., Moore, F., Montzka, S., Perring, A., Hurst, D., Miller, B., Sweeney, C., Oltmans, S., Hintsa, E., Nance, D., Dutton, G., Watts, L., Spackman, J., Rosenlof, K., Ray, E., Hall, B., Zondlo, M., Diao, M., Keeling, R., Bent, J., Atlas, E., Lueb, R., and (Deceased), M. J. M.: HIPPO Combined Discrete Flask and GC Sample GHG, Halocarbon, and Hydrocarbon Data, Version 1.0, NSF NCAR Earth Observing Laboratory, <ext-link xlink:href="https://doi.org/10.3334/CDIAC/HIPPO_012" ext-link-type="DOI">10.3334/CDIAC/HIPPO_012</ext-link>,  2017.</mixed-citation></ref>
      <ref id="bib1.bibx78"><label>Wofsy et al.(2021)Wofsy, Afshar, Allen, Apel, Asher, Barletta, Bent, Bian, Biggs, Blake, Blake, Bourgeois, Brock, Brune, Budney, Bui, Butler, Campuzano-Jost, Chang, Chin, Commane, Correa, Crounse, Cullis, Daube, Day, Dean-Day, Dibb, DiGangi, Diskin, Dollner, Elkins, Erdesz, Fiore, Flynn, Froyd, Gesler, Hall, Hanisco, Hannun, Hills, Hintsa, Hoffman, Hornbrook, Huey, Hughes, Jimenez, Johnson, Katich, Keeling, Kim, Kupc, Lait, McKain, Mclaughlin, Meinardi, Miller, Montzka, Moore, Morgan, Murphy, Murray, Nault, Neuman, Newman, Nicely, Pan, Paplawsky, Peischl, Prather, Price, Ray, Reeves, Richardson, Rollins, Rosenlof, Ryerson, Scheuer, Schill, Schroder, Schwarz, St.Clair, Steenrod, Stephens, Strode, Sweeney, Tanner, Teng, Thames, Thompson, Ullmann, Veres, Wagner, Watt, Weber, Weinzierl, Wennberg, Williamson, Wilson, Wolfe, Woods, Zeng, and Vieznor</label><mixed-citation>Wofsy, S., Afshar, S., Allen, H., Apel, E., Asher, E., Barletta, B., Bent, J., Bian, H., Biggs, B., Blake, D., Blake, N., Bourgeois, I., Brock, C., Brune, W., Budney, J., Bui, T., Butler, A., Campuzano-Jost, P., Chang, C., Chin, M., Commane, R., Correa, G., Crounse, J., Cullis, P. D., Daube, B., Day, D., Dean-Day, J., Dibb, J., DiGangi, J., Diskin, G., Dollner, M., Elkins, J., Erdesz, F., Fiore, A., Flynn, C., Froyd, K., Gesler, D., Hall, S., Hanisco, T., Hannun, R., Hills, A., Hintsa, E., Hoffman, A., Hornbrook, R., Huey, L., Hughes, S., Jimenez, J., Johnson, B., Katich, J., Keeling, R., Kim, M., Kupc, A., Lait, L., McKain, K., Mclaughlin, R., Meinardi, S., Miller, D., Montzka, S., Moore, F., Morgan, E., Murphy, D., Murray, L., Nault, B., Neuman, J., Newman, P., Nicely, J., Pan, X., Paplawsky, W., Peischl, J., Prather, M., Price, D., Ray, E., Reeves, J., Richardson, M., Rollins, A., Rosenlof, K., Ryerson, T., Scheuer, E., Schill, G., Schroder, J., Schwarz, J., St.Clair, J., Steenrod, S., Stephens, B., Strode, S., Sweeney, C., Tanner, D., Teng, A., Thames, A., Thompson, C., Ullmann, K., Veres, P., Wagner, N., Watt, A., Weber, R., Weinzierl, B., Wennberg, P., Williamson, C., Wilson, J., Wolfe, G., Woods, C., Zeng, L., and Vieznor, N.: Atmospheric Tomography Mission (ATom)ATom: Merged Atmospheric Chemistry, Trace Gases, and Aerosols, Version 2, <ext-link xlink:href="https://doi.org/10.3334/ORNLDAAC/1925" ext-link-type="DOI">10.3334/ORNLDAAC/1925</ext-link>,  ORNL Distributed Active Archive Center,  2021.</mixed-citation></ref>
      <ref id="bib1.bibx79"><label>Wofsy(2011)</label><mixed-citation>Wofsy, S. C.: HIAPER Pole-to-Pole Observations (HIPPO): fine-grained, global-scale measurements of climatically important atmospheric gases and aerosols, Philos. T. Roy. Soc. A, 369, 2073–2086, <ext-link xlink:href="https://doi.org/10.1098/rsta.2010.0313" ext-link-type="DOI">10.1098/rsta.2010.0313</ext-link>, 2011. </mixed-citation></ref>
      <ref id="bib1.bibx80"><label>Xiao et al.(2007)Xiao, Prinn, Simmonds, Steele, Novelli, Huang, Langenfelds, O'Doherty, Krummel, Fraser, Porter, Weiss, Salameh, and Wang</label><mixed-citation>Xiao, X., Prinn, R. G., Simmonds, P. G., Steele, L. P., Novelli, P. C., Huang, J., Langenfelds, R. L., O'Doherty, S., Krummel, P. B., Fraser, P. J., Porter, L. W., Weiss, R. F., Salameh, P., and Wang, R. H. J.: Optimal estimation of the soil uptake rate of molecular hydrogen from the Advanced Global Atmospheric Gases Experiment and other measurements, J. Geophys. Res., 112, D07303, <ext-link xlink:href="https://doi.org/10.1029/2006JD007241" ext-link-type="DOI">10.1029/2006JD007241</ext-link>, 2007.</mixed-citation></ref>
      <ref id="bib1.bibx81"><label>Yang et al.(2025)Yang, Jacob, Lin, Dang, Bates, East, Travis, Pendergrass, and Murray</label><mixed-citation>Yang, L. H., Jacob, D. J., Lin, H., Dang, R., Bates, K. H., East, J. D., Travis, K. R., Pendergrass, D. C., and Murray, L. T.: Assessment of Hydrogen's Climate Impact Is Affected by Model OH Biases, Geophys. Res. Lett., 52, e2024GL112445, <ext-link xlink:href="https://doi.org/10.1029/2024GL112445" ext-link-type="DOI">10.1029/2024GL112445</ext-link>, 2025.</mixed-citation></ref>
      <ref id="bib1.bibx82"><label>Yarwood et al.(2005)Yarwood, Rao, Way, Yocke, Whitten, and Reyes</label><mixed-citation> Yarwood, G., Rao, S., Way, R., Yocke, M., Whitten, G. Z., and Reyes, S.: Updates to the Carbon Bond chemical mechanism: CB05, Tech. rep., U.S. Environmental Protection Agency, EPA-HQ-OAR-2010-0799-1062, 2005.</mixed-citation></ref>
      <ref id="bib1.bibx83"><label>Yashiro et al.(2011)Yashiro, Sudo, Yonemura, and Takigawa</label><mixed-citation>Yashiro, H., Sudo, K., Yonemura, S., and Takigawa, M.: The impact of soil uptake on the global distribution of molecular hydrogen: chemical transport model simulation, Atmos. Chem. Phys., 11, 6701–6719, <ext-link xlink:href="https://doi.org/10.5194/acp-11-6701-2011" ext-link-type="DOI">10.5194/acp-11-6701-2011</ext-link>, 2011.</mixed-citation></ref>
      <ref id="bib1.bibx84"><label>Yver et al.(2011)Yver, Pison, Fortems-Cheiney, Schmidt, Chevallier, Ramonet, Jordan, Søvde, Engel, Fisher, Lowry, Nisbet, Levin, Hammer, Necki, Bartyzel, Reimann, Vollmer, Steinbacher, Aalto, Maione, Arduini, O'Doherty, Grant, Sturges, Forster, Lunder, Privalov, Paramonova, Werner, and Bousquet</label><mixed-citation>Yver, C. E., Pison, I. C., Fortems-Cheiney, A., Schmidt, M., Chevallier, F., Ramonet, M., Jordan, A., Søvde, O. A., Engel, A., Fisher, R. E., Lowry, D., Nisbet, E. G., Levin, I., Hammer, S., Necki, J., Bartyzel, J., Reimann, S., Vollmer, M. K., Steinbacher, M., Aalto, T., Maione, M., Arduini, J., O'Doherty, S., Grant, A., Sturges, W. T., Forster, G. L., Lunder, C. R., Privalov, V., Paramonova, N., Werner, A., and Bousquet, P.: A new estimation of the recent tropospheric molecular hydrogen budget using atmospheric observations and variational inversion, Atmos. Chem. Phys., 11, 3375–3392, <ext-link xlink:href="https://doi.org/10.5194/acp-11-3375-2011" ext-link-type="DOI">10.5194/acp-11-3375-2011</ext-link>, 2011.</mixed-citation></ref>

  </ref-list></back>
    <!--<article-title-html>Constraining the hydrogen soil sink and photochemical source: insights from atmospheric H<sub>2</sub> inversions (2003–2023)</article-title-html>
<abstract-html/>
<ref-html id="bib1.bib1"><label>Apadula et al.(2025)Apadula, Arnold, Bergamaschi, Biermann, Chen,
Colomb, Conil, Couret, Cristofanelli, De Mazière, Delmotte, Di Iorio,
Emmenegger, Forster, Frumau, Harris, Haszpra, Hatakka, Heliasz, Heltai,
Hensen, Hermansen, Hoheisel, Kneuer, Komínková, Kubistin, Larmanou,
Laurent, Laurila, Lehner, Lehtinen, Leskinen, Leuenberger, Levula, Lindauer,
Lopez, Lund Myhre, Lunder, Mammarella, Manca, Manning, Marek, Marklund,
Meinhardt, Miettinen, Molnár, Montaguti, Mölder, Müller-Williams,
O'Doherty, Ottosson-Löfvenius, Piacentino, Pichon, Pitt, Platt,
Plaß-Dülmer, Ramonet, Rivas-Soriano, Roulet, Scheeren, Schmidt, Schumacher,
Sferlazzo, Sha, Smith, Stanley, Steinbacher, Sørensen, Trisolino, Vítková,
Ylisirniö, Yver-Kwok, Zazzeri, Zwerschke, di Sarra, ICOS ATC-Laboratoires
des Sciences du Climat et de L'Environnement (LSCE), France, ICOS Central
Radiocarbon Laboratory (CRL), Germany, ICOS Flask And Calibration
Laboratory (FCL), Germany, ICOS-CAL-CRL, and
ICOS-CAL-FCL</label><mixed-citation>
      
Apadula, F., Arnold, S., Bergamaschi, P., Biermann, T., Chen, H., Colomb, A.,
Conil, S., Couret, C., Cristofanelli, P., De Mazière, M., Delmotte, M.,
Di Iorio, T., Emmenegger, L., Forster, G., Frumau, A., Harris, E., Haszpra,
L., Hatakka, J., Heliasz, M., Heltai, D., Hensen, A., Hermansen, O.,
Hoheisel, A., Kneuer, T., Komínková, K., Kubistin, D., Larmanou, E.,
Laurent, O., Laurila, T., Lehner, I., Lehtinen, K., Leskinen, A.,
Leuenberger, M., Levula, J., Lindauer, M., Lopez, M., Lund Myhre, C., Lunder,
C., Mammarella, I., Manca, G., Manning, A., Marek, M. V., Marklund, P.,
Meinhardt, F., Miettinen, P., Molnár, M., Montaguti, S., Mölder, M.,
Müller-Williams, J., O'Doherty, S., Ottosson-Löfvenius, M., Piacentino, S.,
Pichon, J.-M., Pitt, J., Platt, S. M., Plaß-Dülmer, C., Ramonet, M.,
Rivas-Soriano, P., Roulet, Y.-A., Scheeren, B., Schmidt, M., Schumacher, M.,
Sferlazzo, D., Sha, M. K., Smith, P., Stanley, K., Steinbacher, M.,
Sørensen, L. L., Trisolino, P., Vítková, G., Ylisirniö, A., Yver-Kwok,
C., Zazzeri, G., Zwerschke, E., di Sarra, A., ICOS ATC-Laboratoires des
Sciences du Climat et de L'Environnement (LSCE), France, ICOS Central
Radiocarbon Laboratory (CRL), Germany, ICOS Flask And Calibration
Laboratory (FCL), Germany, ICOS-CAL-CRL, and ICOS-CAL-FCL: ICOS
Atmosphere Release 2025-1 of Level 2 Greenhouse Gas Mole
Fractions of CO2, CH4, N2O, CO, meteorology and 14CO2, and flask
samples analysed for CO2, CH4, N2O, CO, H2, SF6, delta 13C
CO2, delta18OCO2, delta O2N2 and 14C,   ICOS ATC time series,ICOS ATC Flask time series Version Number: 1.0, <a href="https://doi.org/10.18160/PP29-9CNZ" target="_blank">https://doi.org/10.18160/PP29-9CNZ</a>,
2025.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib2"><label>Aronson et al.(2019)</label><mixed-citation>
      
Aronson, E. L., Dierick, D., Botthoff, J. K., Oberbauer, S., Zelikova, T. J.,
Harmon, T. C., Rundel, P., Johnson, R. F., Swanson, A. C., Pinto‐Tomás,
A. A., Artavia-León, A., Matarrita-Carranza, B., and Allen, M. F.:
ENSO-Influenced Drought Drives Methane Flux Dynamics in a
Tropical Wet Forest Soil, J. Geophys. Res.-Biogeo., 124, 2267–2276, <a href="https://doi.org/10.1029/2018JG004832" target="_blank">https://doi.org/10.1029/2018JG004832</a>, 2019.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib3"><label>Barnston et al.(1997)Barnston, Chelliah, and
Goldenberg</label><mixed-citation>
      
Barnston, A. G., Chelliah, M., and Goldenberg, S. B.: Documentation of a highly
ENSO-related sst region in the equatorial pacific: Research note,
Atmosphere-Ocean, 35, 367–383, <a href="https://doi.org/10.1080/07055900.1997.9649597" target="_blank">https://doi.org/10.1080/07055900.1997.9649597</a>, 1997.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib4"><label>Batenburg et al.(2016)Batenburg, Schuck, Baker, Zahn, Brenninkmeijer,
and Roeckmann</label><mixed-citation>
      
Batenburg, A., Schuck, T., Baker, A., Zahn, A., Brenninkmeijer, C., and
Roeckmann, T.: H2 mixing ratio and isotope data from CARIBIC project,
figshare, <a href="https://doi.org/10.6084/M9.FIGSHARE.2068992.V1" target="_blank">https://doi.org/10.6084/M9.FIGSHARE.2068992.V1</a>,  2016.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib5"><label>Batenburg et al.(2011)Batenburg, Walter, Pieterse, Levin, Schmidt,
Jordan, Hammer, Yver, and Röckmann</label><mixed-citation>
      
Batenburg, A. M., Walter, S., Pieterse, G., Levin, I., Schmidt, M., Jordan, A., Hammer, S., Yver, C., and Röckmann, T.: Temporal and spatial variability of the stable isotopic composition of atmospheric molecular hydrogen: observations at six EUROHYDROS stations, Atmos. Chem. Phys., 11, 6985–6999, <a href="https://doi.org/10.5194/acp-11-6985-2011" target="_blank">https://doi.org/10.5194/acp-11-6985-2011</a>, 2011.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib6"><label>Bertagni et al.(2021)Bertagni, Paulot, and
Porporato</label><mixed-citation>
      
Bertagni, M. B., Paulot, F., and Porporato, A.: Moisture Fluctuations
Modulate Abiotic and Biotic Limitations of H2 Soil Uptake,
Global Biogeochem. Cy., 35, <a href="https://doi.org/10.1029/2021GB006987" target="_blank">https://doi.org/10.1029/2021GB006987</a>, 2021.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib7"><label>Bousquet et al.(2011)Bousquet, Yver, Pison, Li, Fortems,
Hauglustaine, Szopa, Rayner, Novelli, Langenfelds, Steele, Ramonet, Schmidt,
Foster, Morfopoulos, and Ciais</label><mixed-citation>
      
Bousquet, P., Yver, C., Pison, I., Li, Y. S., Fortems, A., Hauglustaine, D.,
Szopa, S., Rayner, P. J., Novelli, P., Langenfelds, R., Steele, P., Ramonet,
M., Schmidt, M., Foster, P., Morfopoulos, C., and Ciais, P.: A
three-dimensional synthesis inversion of the molecular hydrogen cycle:
Sources and sinks budget and implications for the soil uptake, J.
Geophys. Res., 116, D01302, <a href="https://doi.org/10.1029/2010JD014599" target="_blank">https://doi.org/10.1029/2010JD014599</a>, 2011.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib8"><label>Chen et al.(2015)Chen, Popa, Batenburg, and
Röckmann</label><mixed-citation>
      
Chen, Q., Popa, M. E., Batenburg, A. M., and Röckmann, T.: Isotopic signatures of production and uptake of H2 by soil, Atmos. Chem. Phys., 15, 13003–13021, <a href="https://doi.org/10.5194/acp-15-13003-2015" target="_blank">https://doi.org/10.5194/acp-15-13003-2015</a>, 2015.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib9"><label>Ciais et al.(2022)Ciais, Bastos, Chevallier, Lauerwald, Poulter,
Canadell, Hugelius, Jackson, Jain, Jones, Kondo, Luijkx, Patra, Peters,
Pongratz, Petrescu, Piao, Qiu, Von Randow, Regnier, Saunois, Scholes,
Shvidenko, Tian, Yang, Wang, and Zheng</label><mixed-citation>
      
Ciais, P., Bastos, A., Chevallier, F., Lauerwald, R., Poulter, B., Canadell, J. G., Hugelius, G., Jackson, R. B., Jain, A., Jones, M., Kondo, M., Luijkx, I. T., Patra, P. K., Peters, W., Pongratz, J., Petrescu, A. M. R., Piao, S., Qiu, C., Von Randow, C., Regnier, P., Saunois, M., Scholes, R., Shvidenko, A., Tian, H., Yang, H., Wang, X., and Zheng, B.: Definitions and methods to estimate regional land carbon fluxes for the second phase of the REgional Carbon Cycle Assessment and Processes Project (RECCAP-2), Geosci. Model Dev., 15, 1289–1316, <a href="https://doi.org/10.5194/gmd-15-1289-2022" target="_blank">https://doi.org/10.5194/gmd-15-1289-2022</a>, 2022.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib10"><label>Conrad(1996)</label><mixed-citation>
      
Conrad, R.: Soil microorganisms as controllers of atmospheric trace gases
(H2, CO, CH4, OCS, N2O, and NO), Microbiol. Rev., 60,
609–640, <a href="https://doi.org/10.1128/mr.60.4.609-640.1996" target="_blank">https://doi.org/10.1128/mr.60.4.609-640.1996</a>, 1996.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib11"><label>Conrad et al.(1989)Conrad, Mayer, and Wüst</label><mixed-citation>
      
Conrad, R., Mayer, H.-P., and Wüst, M.: Temporal change of gas metabolism by
hydrogen-syntrophic methanogenic bacterial associations in anoxic paddy soil,
FEMS Microbiol. Lett., 62, 265–273,
<a href="https://doi.org/10.1111/j.1574-6968.1989.tb03701.x" target="_blank">https://doi.org/10.1111/j.1574-6968.1989.tb03701.x</a>, 1989.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib12"><label>Conte et al.(2019)Conte, Szopa, Séférian, and
Bopp</label><mixed-citation>
      
Conte, L., Szopa, S., Séférian, R., and Bopp, L.: The oceanic cycle of carbon monoxide and its emissions to the atmosphere, Biogeosciences, 16, 881–902, <a href="https://doi.org/10.5194/bg-16-881-2019" target="_blank">https://doi.org/10.5194/bg-16-881-2019</a>, 2019.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib13"><label>Cowan et al.(2025)Cowan, Roberts, Hanlon, Bezanger, Toteva, Tweedie,
Yeung, Deshpande, Levy, Skiba, Nemitz, and Drewer</label><mixed-citation>
      
Cowan, N., Roberts, T., Hanlon, M., Bezanger, A., Toteva, G., Tweedie, A., Yeung, K., Deshpande, A., Levy, P., Skiba, U., Nemitz, E., and Drewer, J.: Quantifying the soil sink of atmospheric hydrogen: a full year of field measurements from grassland and forest soils in the UK, Biogeosciences, 22, 3449–3461, <a href="https://doi.org/10.5194/bg-22-3449-2025" target="_blank">https://doi.org/10.5194/bg-22-3449-2025</a>, 2025.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib14"><label>De Kleijne et al.(2024)De Kleijne, Huijbregts, Knobloch, Van Zelm,
Hilbers, De Coninck, and Hanssen</label><mixed-citation>
      
De Kleijne, K., Huijbregts, M. A. J., Knobloch, F., Van Zelm, R., Hilbers,
J. P., De Coninck, H., and Hanssen, S. V.: Worldwide greenhouse gas emissions
of green hydrogen production and transport, Nature Energy,
<a href="https://doi.org/10.1038/s41560-024-01563-1" target="_blank">https://doi.org/10.1038/s41560-024-01563-1</a>, 2024.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib15"><label>Ehhalt and Rohrer(2009)</label><mixed-citation>
      
Ehhalt, D. H. and Rohrer, F.: The tropospheric cycle of H2: a critical
review, Tellus B, 61, 500–535,
<a href="https://doi.org/10.1111/j.1600-0889.2009.00416.x" target="_blank">https://doi.org/10.1111/j.1600-0889.2009.00416.x</a>, 2009.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib16"><label>Ehhalt and Rohrer(2011)</label><mixed-citation>
      
Ehhalt, D. H. and Rohrer, F.: The dependence of soil H<sub>2</sub> uptake
on temperature and moisture: a reanalysis of laboratory data, Tellus B, 63, 1040,
<a href="https://doi.org/10.1111/j.1600-0889.2011.00581.x" target="_blank">https://doi.org/10.1111/j.1600-0889.2011.00581.x</a>, 2011.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib17"><label>Feng et al.(2020)Feng, Smith, Braun, Crippa, Gidden, Hoesly, Klimont,
Van Marle, Van Den Berg, and Van Der Werf</label><mixed-citation>
      
Feng, L., Smith, S. J., Braun, C., Crippa, M., Gidden, M. J., Hoesly, R., Klimont, Z., van Marle, M., van den Berg, M., and van der Werf, G. R.: The generation of gridded emissions data for CMIP6, Geosci. Model Dev., 13, 461–482, <a href="https://doi.org/10.5194/gmd-13-461-2020" target="_blank">https://doi.org/10.5194/gmd-13-461-2020</a>, 2020.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib18"><label>Gibson et al.(2024)Gibson, Eby, and Jaggi</label><mixed-citation>
      
Gibson, J., Eby, P., and Jaggi, A.: Natural isotope fingerprinting of produced
hydrogen and its potential applications to the hydrogen economy,
Int. J. Hydrogen Energ., 66, 468–478,
<a href="https://doi.org/10.1016/j.ijhydene.2024.04.077" target="_blank">https://doi.org/10.1016/j.ijhydene.2024.04.077</a>, 2024.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib19"><label>Gros et al.(2009)Gros, Peeken, Bluhm, Zöllner, Sarda-Esteve, and
Bonsang</label><mixed-citation>
      
Gros, V., Peeken, I., Bluhm, K., Zöllner, E., Sarda-Esteve, R., and Bonsang,
B.: Carbon monoxide emissions by phytoplankton: evidence from laboratory
experiments, Environ. Chem., 6, 369, <a href="https://doi.org/10.1071/EN09020" target="_blank">https://doi.org/10.1071/EN09020</a>, 2009.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib20"><label>Guenther et al.(2012)Guenther, Jiang, Heald, Sakulyanontvittaya,
Duhl, Emmons, and Wang</label><mixed-citation>
      
Guenther, A. B., Jiang, X., Heald, C. L., Sakulyanontvittaya, T., Duhl, T., Emmons, L. K., and Wang, X.: The Model of Emissions of Gases and Aerosols from Nature version 2.1 (MEGAN2.1): an extended and updated framework for modeling biogenic emissions, Geosci. Model Dev., 5, 1471–1492, <a href="https://doi.org/10.5194/gmd-5-1471-2012" target="_blank">https://doi.org/10.5194/gmd-5-1471-2012</a>, 2012.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib21"><label>Hall et al.(2018)Hall, Ullmann, Prather, Flynn, Murray, Fiore,
Correa, Strode, Steenrod, Lamarque, Guth, Josse, Flemming, Huijnen, Abraham,
and Archibald</label><mixed-citation>
      
Hall, S. R., Ullmann, K., Prather, M. J., Flynn, C. M., Murray, L. T., Fiore, A. M., Correa, G., Strode, S. A., Steenrod, S. D., Lamarque, J.-F., Guth, J., Josse, B., Flemming, J., Huijnen, V., Abraham, N. L., and Archibald, A. T.: Cloud impacts on photochemistry: building a climatology of photolysis rates from the Atmospheric Tomography mission, Atmos. Chem. Phys., 18, 16809–16828, <a href="https://doi.org/10.5194/acp-18-16809-2018" target="_blank">https://doi.org/10.5194/acp-18-16809-2018</a>, 2018.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib22"><label>Hanisco et al.(2019)Hanisco, Hannun, St.Clair, and
Wolfe</label><mixed-citation>
      
Hanisco, T., Hannun, R., St.Clair, J., and Wolfe, G.: Atmospheric Tomography
Mission (ATom)ATom: L2 Measurements of In Situ Airborne
Formaldehyde (ISAF), ORNL Distributed Active Archive Center, <a href="https://doi.org/10.3334/ORNLDAAC/1730" target="_blank">https://doi.org/10.3334/ORNLDAAC/1730</a>, 2019.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib23"><label>Hauglustaine and Ehhalt(2002)</label><mixed-citation>
      
Hauglustaine, D. A. and Ehhalt, D. H.: A three-dimensional model of molecular
hydrogen in the troposphere: TROPOSPHERIC MOLECULAR HYDROGEN, J. Geophys. Res.-Atmos., 107, ACH 4-1–ACH 4-16,
<a href="https://doi.org/10.1029/2001JD001156" target="_blank">https://doi.org/10.1029/2001JD001156</a>, 2002.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib24"><label>Hersbach et al.(2020)</label><mixed-citation>
      
Hersbach, H., Bell, B., Berrisford, P., Hirahara, S., Horányi, A.,
Muñoz-Sabater, J., Nicolas, J., Peubey, C., Radu, R., Schepers, D.,
Simmons, A., Soci, C., Abdalla, S., Abellan, X., Balsamo, G., Bechtold, P.,
Biavati, G., Bidlot, J., Bonavita, M., De Chiara, G., Dahlgren, P., Dee, D.,
Diamantakis, M., Dragani, R., Flemming, J., Forbes, R., Fuentes, M., Geer,
A., Haimberger, L., Healy, S., Hogan, R. J., Hólm, E., Janisková, M.,
Keeley, S., Laloyaux, P., Lopez, P., Lupu, C., Radnoti, G., De Rosnay, P.,
Rozum, I., Vamborg, F., Villaume, S., and Thépaut, J.: The ERA5 global
reanalysis, Q. J. Roy. Meteor. Soc., 146,
1999–2049, <a href="https://doi.org/10.1002/qj.3803" target="_blank">https://doi.org/10.1002/qj.3803</a>, 2020.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib25"><label>Hoesly et al.(2018)Hoesly, Smith, Feng, Klimont, Janssens-Maenhout,
Pitkanen, Seibert, Vu, Andres, Bolt, Bond, Dawidowski, Kholod, Kurokawa, Li,
Liu, Lu, Moura, O'Rourke, and Zhang</label><mixed-citation>
      
Hoesly, R. M., Smith, S. J., Feng, L., Klimont, Z., Janssens-Maenhout, G., Pitkanen, T., Seibert, J. J., Vu, L., Andres, R. J., Bolt, R. M., Bond, T. C., Dawidowski, L., Kholod, N., Kurokawa, J.-I., Li, M., Liu, L., Lu, Z., Moura, M. C. P., O'Rourke, P. R., and Zhang, Q.: Historical (1750–2014) anthropogenic emissions of reactive gases and aerosols from the Community Emissions Data System (CEDS), Geosci. Model Dev., 11, 369–408, <a href="https://doi.org/10.5194/gmd-11-369-2018" target="_blank">https://doi.org/10.5194/gmd-11-369-2018</a>, 2018.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib26"><label>IEA(2019)</label><mixed-citation>
      
IEA: The Future of Hydrogen, Tech. rep., International Energy Agency,
<a href="https://www.iea.org/reports/the-future-of-hydrogen" target="_blank"/> (last access: 28 May 2026), 2019.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib27"><label>IEA(2023)</label><mixed-citation>
      
IEA: Global Hydrogen Review 2023, Tech. rep., International Energy Agency,
<a href="https://www.iea.org/reports/global-hydrogen-review-2023" target="_blank"/> (last access: 28 May 2026),
2023.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib28"><label>Inness et al.(2019)Inness, Ades, Agustí-Panareda, Barré,
Benedictow, Blechschmidt, Dominguez, Engelen, Eskes, Flemming, Huijnen,
Jones, Kipling, Massart, Parrington, Peuch, Razinger, Remy, Schulz, and
Suttie</label><mixed-citation>
      
Inness, A., Ades, M., Agustí-Panareda, A., Barré, J., Benedictow, A., Blechschmidt, A.-M., Dominguez, J. J., Engelen, R., Eskes, H., Flemming, J., Huijnen, V., Jones, L., Kipling, Z., Massart, S., Parrington, M., Peuch, V.-H., Razinger, M., Remy, S., Schulz, M., and Suttie, M.: The CAMS reanalysis of atmospheric composition, Atmos. Chem. Phys., 19, 3515–3556, <a href="https://doi.org/10.5194/acp-19-3515-2019" target="_blank">https://doi.org/10.5194/acp-19-3515-2019</a>, 2019.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib29"><label>Jordan and Steinberg(2011)</label><mixed-citation>
      
Jordan, A. and Steinberg, B.: Calibration of atmospheric hydrogen measurements, Atmos. Meas. Tech., 4, 509–521, <a href="https://doi.org/10.5194/amt-4-509-2011" target="_blank">https://doi.org/10.5194/amt-4-509-2011</a>, 2011.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib30"><label>Jordan et al.(2024)Jordan, Moossen, Rothe, Brand, Heimann, Zaehle,
and Rödenbeck</label><mixed-citation>
      
Jordan, A., Moossen, H., Rothe, M., Brand, W., Heimann, M., Zaehle, S., and
Rödenbeck, C.: Atmospheric flask sampling program of MPI-BGC, version
2024.1,  Edmond, <a href="https://doi.org/10.17617/3.S3KHMS" target="_blank">https://doi.org/10.17617/3.S3KHMS</a>, 2024.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib31"><label>Krol et al.(2005)Krol, Houweling, Bregman, Van Den Broek, Segers,
Van Velthoven, Peters, Dentener, and Bergamaschi</label><mixed-citation>
      
Krol, M., Houweling, S., Bregman, B., van den Broek, M., Segers, A., van Velthoven, P., Peters, W., Dentener, F., and Bergamaschi, P.: The two-way nested global chemistry-transport zoom model TM5: algorithm and applications, Atmos. Chem. Phys., 5, 417–432, <a href="https://doi.org/10.5194/acp-5-417-2005" target="_blank">https://doi.org/10.5194/acp-5-417-2005</a>, 2005.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib32"><label>Krummel et al.(2025)Krummel, Langenfelds, Guerette, and
Loh</label><mixed-citation>
      
Krummel, P., Langenfelds, R., Guerette, E., and Loh, Z.: Atmospheric H2 at
sites from the Commonwealth Scientific and Industrial Research
Organisation (CSIRO) global monitoring network, World Data Centre for Greenhouse Gases,
<a href="https://doi.org/10.50849/WDCGG_H2_ALL_2025" target="_blank">https://doi.org/10.50849/WDCGG_H2_ALL_2025</a>, 2025.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib33"><label>McCarthy(2004)</label><mixed-citation>
      
McCarthy, M. C.: The hydrogen isotopic composition of water vapor entering the
stratosphere inferred from high-precision measurements of CHD3 and HD,
J. Geophys. Res., 109, D07304, <a href="https://doi.org/10.1029/2003JD004003" target="_blank">https://doi.org/10.1029/2003JD004003</a>,
2004.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib34"><label>McKain et al.(2022)McKain, Sweeney, Crotwell, Crotwell, Higgs,
Legard, Madronich, Miller, Moglia, Mund, Neff, Wolter, and NOAA Global
Monitoring Laboratory</label><mixed-citation>
      
McKain, K., Sweeney, C., Crotwell, A., Crotwell, M., Higgs, J., Legard, T.,
Madronich, M., Miller, J., Moglia, E., Mund, J., Neff, D., Wolter, S., and
NOAA Global Monitoring Laboratory: NOAA Global Greenhouse Gas
Reference Network Flask-Air PFP Sample Measurements of
CO<sub>2</sub>, CH<sub>4</sub>, CO, N<sub>2</sub>O,
H<sub>2</sub>, SF<sub>6</sub> and isotopic ratios collected from
aircraft vertical profiles, NOAA Global Monitoring Laboratory, <a href="https://doi.org/10.15138/39HR-9N34" target="_blank">https://doi.org/10.15138/39HR-9N34</a>, 2022.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib35"><label>Montzka et al.(2019)Montzka, Moore, and
Sweeney</label><mixed-citation>
      
Montzka, S., Moore, F., and Sweeney, C.: Atmospheric Tomography Mission
(ATom)ATom: L2 Measurements from the Programmable Flask Package
(PFP) Whole Air Sampler, ORNL Distributed Active Archive Center, <a href="https://doi.org/10.3334/ORNLDAAC/1746" target="_blank">https://doi.org/10.3334/ORNLDAAC/1746</a>,  2019.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib36"><label>Naik et al.(2013)Naik, Voulgarakis, Fiore, Horowitz, Lamarque, Lin,
Prather, Young, Bergmann, Cameron-Smith, Cionni, Collins, Dalsøren, Doherty,
Eyring, Faluvegi, Folberth, Josse, Lee, MacKenzie, Nagashima, Van Noije,
Plummer, Righi, Rumbold, Skeie, Shindell, Stevenson, Strode, Sudo, Szopa, and
Zeng</label><mixed-citation>
      
Naik, V., Voulgarakis, A., Fiore, A. M., Horowitz, L. W., Lamarque, J.-F., Lin, M., Prather, M. J., Young, P. J., Bergmann, D., Cameron-Smith, P. J., Cionni, I., Collins, W. J., Dalsøren, S. B., Doherty, R., Eyring, V., Faluvegi, G., Folberth, G. A., Josse, B., Lee, Y. H., MacKenzie, I. A., Nagashima, T., van Noije, T. P. C., Plummer, D. A., Righi, M., Rumbold, S. T., Skeie, R., Shindell, D. T., Stevenson, D. S., Strode, S., Sudo, K., Szopa, S., and Zeng, G.: Preindustrial to present-day changes in tropospheric hydroxyl radical and methane lifetime from the Atmospheric Chemistry and Climate Model Intercomparison Project (ACCMIP), Atmos. Chem. Phys., 13, 5277–5298, <a href="https://doi.org/10.5194/acp-13-5277-2013" target="_blank">https://doi.org/10.5194/acp-13-5277-2013</a>, 2013.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib37"><label>Naus et al.(2021)Naus, Montzka, Patra, and
Krol</label><mixed-citation>
      
Naus, S., Montzka, S. A., Patra, P. K., and Krol, M. C.: A three-dimensional-model inversion of methyl chloroform to constrain the atmospheric oxidative capacity, Atmos. Chem. Phys., 21, 4809–4824, <a href="https://doi.org/10.5194/acp-21-4809-2021" target="_blank">https://doi.org/10.5194/acp-21-4809-2021</a>, 2021.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib38"><label>NOAA Global Monitoring
Laboratory(2026)</label><mixed-citation>
      
NOAA Global Monitoring Laboratory: Curve Fitting Methods Applied to
Time Series (CCGCRV),
<a href="https://gml.noaa.gov/ccgg/mbl/crvfit/" target="_blank"/> (last access: 28 May 2026), 2026.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib39"><label>Novelli et al.(1999)Novelli, Lang, Masarie, Hurst, Myers, and
Elkins</label><mixed-citation>
      
Novelli, P., Lang, P., Masarie, K., Hurst, D., Myers, R., and Elkins, J.:
Molecular hydrogen in the troposphere: Global distribution and budget,
J. Geophys. Res., 104, 30427–30444,
<a href="https://doi.org/10.1029/1999JD900788" target="_blank">https://doi.org/10.1029/1999JD900788</a>, 1999.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib40"><label>Ocko and Hamburg(2022)</label><mixed-citation>
      
Ocko, I. B. and Hamburg, S. P.: Climate consequences of hydrogen emissions, Atmos. Chem. Phys., 22, 9349–9368, <a href="https://doi.org/10.5194/acp-22-9349-2022" target="_blank">https://doi.org/10.5194/acp-22-9349-2022</a>, 2022.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib41"><label>Ouyang et al.(2025)Ouyang, Jackson, Saunois, Canadell, Zhao,
Morfopoulos, Krummel, Patra, Peters, Dennison, Gasser, Archibald, Arora,
Baudoin, Chandra, Ciais, Davis, Feron, Guo, Hauglustaine, Jones, Jones, Kato,
Kennedy, Knauer, Lienert, Lombardozzi, Melton, Nabel, O’Sullivan, Pétron,
Poulter, Rogelj, Sandoval Calle, Smith, Suntharalingam, Tian, Wang, and
Wiltshire</label><mixed-citation>
      
Ouyang, Z., Jackson, R. B., Saunois, M., Canadell, J. G., Zhao, Y.,
Morfopoulos, C., Krummel, P. B., Patra, P. K., Peters, G. P., Dennison, F.,
Gasser, T., Archibald, A. T., Arora, V., Baudoin, G., Chandra, N., Ciais, P.,
Davis, S. J., Feron, S., Guo, F., Hauglustaine, D., Jones, C. D., Jones,
M. W., Kato, E., Kennedy, D., Knauer, J., Lienert, S., Lombardozzi, D.,
Melton, J. R., Nabel, J. E., O'Sullivan, M., Pétron, G., Poulter, B.,
Rogelj, J., Sandoval Calle, D., Smith, P., Suntharalingam, P., Tian, H.,
Wang, C., and Wiltshire, A.: The global hydrogen budget, Nature, 648,
616–624, <a href="https://doi.org/10.1038/s41586-025-09806-1" target="_blank">https://doi.org/10.1038/s41586-025-09806-1</a>, 2025.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib42"><label>Patra et al.(2021)Patra, Krol, Prinn, Takigawa, Mühle, Montzka, Lal,
Yamashita, Naus, Chandra, Weiss, Krummel, Fraser, O'Doherty, and
Elkins</label><mixed-citation>
      
Patra, P. K., Krol, M. C., Prinn, R. G., Takigawa, M., Mühle, J., Montzka,
S. A., Lal, S., Yamashita, Y., Naus, S., Chandra, N., Weiss, R. F., Krummel,
P. B., Fraser, P. J., O'Doherty, S., and Elkins, J. W.: Methyl Chloroform
Continues to Constrain the Hydroxyl (OH) Variability in the
Troposphere, J. Geophys. Res.-Atmos., 126,
e2020JD033862, <a href="https://doi.org/10.1029/2020JD033862" target="_blank">https://doi.org/10.1029/2020JD033862</a>, 2021.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib43"><label>Paulot et al.(2021)Paulot, Paynter, Naik, Malyshev, Menzel, and
Horowitz</label><mixed-citation>
      
Paulot, F., Paynter, D., Naik, V., Malyshev, S., Menzel, R., and Horowitz,
L. W.: Global modeling of hydrogen using GFDL-AM4.1: Sensitivity of
soil removal and radiative forcing, Int. J. Hydrogen Energ.,
46, 13446–13460, <a href="https://doi.org/10.1016/j.ijhydene.2021.01.088" target="_blank">https://doi.org/10.1016/j.ijhydene.2021.01.088</a>, 2021.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib44"><label>Percival and Sander(2026)</label><mixed-citation>
      
Percival, C. J. and Sander, S. P.: Chemical Kinetics and Photochemical
Data for Use in Atmospheric Studies Evaluation Number 20, Tech.
rep., Jet Propulsion Laboratory, California Institute of Technology,
<a href="https://doi.org/10.48577/JPL.KFQ0TZ" target="_blank">https://doi.org/10.48577/JPL.KFQ0TZ</a>, 2026.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib45"><label>Pieterse(2013)</label><mixed-citation>
      
Pieterse, G.: Modelling the global tropospheric molecular hydrogen cycle, Ph.D.
thesis, Utrecht University, ISBN 978-94-6191-559-7,
2013.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib46"><label>Pieterse et al.(2011)Pieterse, Krol, Batenburg, Steele, Krummel,
Langenfelds, and Röckmann</label><mixed-citation>
      
Pieterse, G., Krol, M. C., Batenburg, A. M., Steele, L. P., Krummel, P. B., Langenfelds, R. L., and Röckmann, T.: Global modelling of H2 mixing ratios and isotopic compositions with the TM5 model, Atmos. Chem. Phys., 11, 7001–7026, <a href="https://doi.org/10.5194/acp-11-7001-2011" target="_blank">https://doi.org/10.5194/acp-11-7001-2011</a>, 2011.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib47"><label>Pieterse et al.(2013)Pieterse, Krol, Batenburg, Brenninkmeijer, Popa,
O'Doherty, Grant, Steele, Krummel, Langenfelds, Wang, Vermeulen, Schmidt,
Yver, Jordan, Engel, Fisher, Lowry, Nisbet, and
Röckmann</label><mixed-citation>
      
Pieterse, G., Krol, M., Batenburg, A., Brenninkmeijer, C., Popa, E., O'Doherty,
S., Grant, A., Steele, L., Krummel, P., Langenfelds, R., Wang, H., Vermeulen,
A., Schmidt, M., Yver, C., Jordan, A., Engel, A., Fisher, R., Lowry, D.,
Nisbet, E., and Röckmann, T.: Reassessing the variability in atmospheric
H2 using the two-way nested TM5 model, J. Geophys. Res.,
<a href="https://doi.org/10.1002/jgrd.50204" target="_blank">https://doi.org/10.1002/jgrd.50204</a>, 2013.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib48"><label>Pison et al.(2009)Pison, Bousquet, Chevallier, Szopa, and
Hauglustaine</label><mixed-citation>
      
Pison, I., Bousquet, P., Chevallier, F., Szopa, S., and Hauglustaine, D.: Multi-species inversion of CH<sub>4</sub>, CO and H<sub>2</sub> emissions from surface measurements, Atmos. Chem. Phys., 9, 5281–5297, <a href="https://doi.org/10.5194/acp-9-5281-2009" target="_blank">https://doi.org/10.5194/acp-9-5281-2009</a>, 2009.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib49"><label>Prather et al.(2012)Prather, Holmes, and Hsu</label><mixed-citation>
      
Prather, M. J., Holmes, C. D., and Hsu, J.: Reactive greenhouse gas scenarios:
Systematic exploration of uncertainties and the role of atmospheric
chemistry, Geophys. Res. Lett., 39, 2012GL051440,
<a href="https://doi.org/10.1029/2012GL051440" target="_blank">https://doi.org/10.1029/2012GL051440</a>, 2012.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib50"><label>Price et al.(2007)Price, Jaeglé, Rice, Quay, Novelli, and
Gammon</label><mixed-citation>
      
Price, H., Jaeglé, L., Rice, A., Quay, P., Novelli, P. C., and Gammon, R.:
Global budget of molecular hydrogen and its deuterium content: Constraints
from ground station, cruise, and aircraft observations, J.
Geophys. Res., 112, D22108, <a href="https://doi.org/10.1029/2006JD008152" target="_blank">https://doi.org/10.1029/2006JD008152</a>, 2007.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib51"><label>Prinn et al.(2025)Prinn, Weiss, Arduini, Choi, Engel, Fraser,
Ganesan, Harth, Hermansen, Kim, Krummel, Loh, Lunder, Maione, Manning,
Mitrevski, Mühle, O'Doherty, Park, Pitt, Reimann, Rigby, Saito, Salameh,
Schmidt, Simmonds, Stanley, Stavert, Steele, Vollmer, Wagenhäuser, Wang,
Wenger, Western, Yao, Young, Zhou, and Zhu</label><mixed-citation>
      
Prinn, R., Weiss, R., Arduini, J., Choi, H., Engel, A., Fraser, P., Ganesan,
A., Harth, C., Hermansen, O., Kim, J., Krummel, P., Loh, Z., Lunder, C.,
Maione, M., Manning, A., Mitrevski, B., Mühle, J., O'Doherty, S., Park, S.,
Pitt, J., Reimann, S., Rigby, M., Saito, T., Salameh, P., Schmidt, R.,
Simmonds, P., Stanley, K., Stavert, A., Steele, P., Vollmer, M.,
Wagenhäuser, T., Wang, H., Wenger, A., Western, L., Yao, B., Young, D.,
Zhou, L., and Zhu, L.: The dataset of in-situ measurements of chemically and
radiatively important atmospheric gases from the Advanced Global
Atmospheric Gas Experiment (AGAGE) and affiliated stations,
NASA Langley Research Center (LaRC) Data Host Facility (DHF),  <a href="https://doi.org/10.60718/75D7-QE84" target="_blank">https://doi.org/10.60718/75D7-QE84</a>, 2025.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib52"><label>Pétron et al.(2024)Pétron, Crotwell, Mund, Crotwell, Mefford,
Thoning, Hall, Kitzis, Madronich, Moglia, Neff, Wolter, Jordan, Krummel,
Langenfelds, and Patterson</label><mixed-citation>
      
Pétron, G., Crotwell, A. M., Mund, J., Crotwell, M., Mefford, T., Thoning, K., Hall, B., Kitzis, D., Madronich, M., Moglia, E., Neff, D., Wolter, S., Jordan, A., Krummel, P., Langenfelds, R., and Patterson, J.: Atmospheric H2 observations from the NOAA Cooperative Global Air Sampling Network, Atmos. Meas. Tech., 17, 4803–4823, <a href="https://doi.org/10.5194/amt-17-4803-2024" target="_blank">https://doi.org/10.5194/amt-17-4803-2024</a>, 2024.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib53"><label>Rahn et al.(2003)Rahn, Eiler, Boering, Wennberg, McCarthy, Tyler,
Schauffler, Donnelly, and Atlas</label><mixed-citation>
      
Rahn, T., Eiler, J. M., Boering, K. A., Wennberg, P. O., McCarthy, M. C.,
Tyler, S., Schauffler, S., Donnelly, S., and Atlas, E.: Extreme deuterium
enrichment in stratospheric hydrogen and the global atmospheric budget of
H2, Nature, 424, 918–921, <a href="https://doi.org/10.1038/nature01917" target="_blank">https://doi.org/10.1038/nature01917</a>, 2003.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib54"><label>Rhee et al.(2006)Rhee, Brenninkmeijer, and
Röckmann</label><mixed-citation>
      
Rhee, T. S., Brenninkmeijer, C. A. M., and Röckmann, T.: The overwhelming role of soils in the global atmospheric hydrogen cycle, Atmos. Chem. Phys., 6, 1611–1625, <a href="https://doi.org/10.5194/acp-6-1611-2006" target="_blank">https://doi.org/10.5194/acp-6-1611-2006</a>, 2006.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib55"><label>Rodell et al.(2004)Rodell, Houser, Jambor, Gottschalck, Mitchell,
Meng, Arsenault, Cosgrove, Radakovich, Bosilovich, Entin, Walker, Lohmann,
and Toll</label><mixed-citation>
      
Rodell, M., Houser, P. R., Jambor, U., Gottschalck, J., Mitchell, K., Meng,
C.-J., Arsenault, K., Cosgrove, B., Radakovich, J., Bosilovich, M., Entin,
J. K., Walker, J. P., Lohmann, D., and Toll, D.: The Global Land Data
Assimilation System, B. Am. Meteor. Soc., 85,
381–394, <a href="https://doi.org/10.1175/BAMS-85-3-381" target="_blank">https://doi.org/10.1175/BAMS-85-3-381</a>, 2004.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib56"><label>Röckmann et al.(2003)Röckmann, Rhee, and
Engel</label><mixed-citation>
      
Röckmann, T., Rhee, T. S., and Engel, A.: Heavy hydrogen in the stratosphere, Atmos. Chem. Phys., 3, 2015–2023, <a href="https://doi.org/10.5194/acp-3-2015-2003" target="_blank">https://doi.org/10.5194/acp-3-2015-2003</a>, 2003.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib57"><label>Sand et al.(2023)Sand, Skeie, Sandstad, Krishnan, Myhre, Bryant,
Derwent, Hauglustaine, Paulot, Prather, and
Stevenson</label><mixed-citation>
      
Sand, M., Skeie, R. B., Sandstad, M., Krishnan, S., Myhre, G., Bryant, H.,
Derwent, R., Hauglustaine, D., Paulot, F., Prather, M., and Stevenson, D.: A
multi-model assessment of the Global Warming Potential of hydrogen,
Commun. Earth  Environ., 4, 1–12,
<a href="https://doi.org/10.1038/s43247-023-00857-8" target="_blank">https://doi.org/10.1038/s43247-023-00857-8</a>,  2023.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib58"><label>Sand et al.(2026)Sand, Ocko, Sun, Skeie, Hauglustaine, Warwick,
Stevenson, Paulot, Myhre, Archibald, Sandstad, Krishnan, and
Prather</label><mixed-citation>
      
Sand, M., Ocko, I. B., Sun, T., Skeie, R. B., Hauglustaine, D. A., Warwick,
N. J., Stevenson, D. S., Paulot, F., Myhre, G., Archibald, A. T., Sandstad,
M., Krishnan, S., and Prather, M.: Climate impacts of hydrogen emissions,
Nat. Rev. Earth Environ., <a href="https://doi.org/10.1038/s43017-026-00792-0" target="_blank">https://doi.org/10.1038/s43017-026-00792-0</a>, 2026.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib59"><label>Sander et al.(2006)Sander, Friedl, Ravishankara, Golden, Kolb,
Kurylo, Molina, Moortgat, Keller-Rudek, Finlayson-Pitts, Wine, Huie, and
Orkin</label><mixed-citation>
      
Sander, S. P., Friedl, R. R., Ravishankara, A. R., Golden, D. M., Kolb, C. E.,
Kurylo, M. J., Molina, M. J., Moortgat, G. K., Keller-Rudek, H.,
Finlayson-Pitts, B. J., Wine, P. H., Huie, R. E., and Orkin, V. L.: Chemical
Kinetics and Photochemical Data for Use in Atmospheric Studies:
Evaluation No. 15, Tech. Rep. JPL Publication 06-2, Jet Propulsion
Laboratory, California Institute of Technology,
<a href="https://science.jpl.nasa.gov/documents/1493/JPL_15_AllInOne.pdf" target="_blank"/> (last access: 28 May 2026), 2006.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib60"><label>Sanderson et al.(2003)Sanderson, Collins, Derwent, and
Johnson</label><mixed-citation>
      
Sanderson, M. G., Collins, W. J., Derwent, R. G., and Johnson, C. E.:
Simulation of Global Hydrogen Levels Using a Lagrangian
Three-Dimensional Model, J. Atmos. Chem., 46, 15–28,
<a href="https://doi.org/10.1023/A:1024824223232" target="_blank">https://doi.org/10.1023/A:1024824223232</a>, 2003.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib61"><label>Smith-Downey et al.(2006)Smith-Downey, Randerson, and
Eiler</label><mixed-citation>
      
Smith-Downey, N. V., Randerson, J. T., and Eiler, J. M.: Temperature and
moisture dependence of soil H<sub>2</sub> uptake measured in the
laboratory, Geophys. Res. Lett., 33, 2006GL026749,
<a href="https://doi.org/10.1029/2006GL026749" target="_blank">https://doi.org/10.1029/2006GL026749</a>, 2006.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib62"><label>Soulie et al.(2024)Soulie, Granier, Darras, Zilbermann, Doumbia,
Guevara, Jalkanen, Keita, Liousse, Crippa, Guizzardi, Hoesly, and
Smith</label><mixed-citation>
      
Soulie, A., Granier, C., Darras, S., Zilbermann, N., Doumbia, T., Guevara, M., Jalkanen, J.-P., Keita, S., Liousse, C., Crippa, M., Guizzardi, D., Hoesly, R., and Smith, S. J.: Global anthropogenic emissions (CAMS-GLOB-ANT) for the Copernicus Atmosphere Monitoring Service simulations of air quality forecasts and reanalyses, Earth Syst. Sci. Data, 16, 2261–2279, <a href="https://doi.org/10.5194/essd-16-2261-2024" target="_blank">https://doi.org/10.5194/essd-16-2261-2024</a>, 2024.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib63"><label>Stone et al.(2026)Stone, Chen, Solomon, Western, Krummel, Pétron,
Mühle, and O'Doherty</label><mixed-citation>
      
Stone, K., Chen, C., Solomon, S., Western, L. M., Krummel, P. B., Pétron, G., Mühle, J., and O'Doherty, S.: Constraining the atmospheric hydrogen oxidation and soil sink seasonal cycles using HFC-152a, Atmos. Chem. Phys., 26, 10241–10253, <a href="https://doi.org/10.5194/acp-26-10241-2026" target="_blank">https://doi.org/10.5194/acp-26-10241-2026</a>, 2026.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib64"><label>Stroo(2026)</label><mixed-citation>
      
Stroo, F.:  Posterior and Prior fluxes and model code for the 2003–2023 H2 inversions, Zenodo [data set], <a href="https://doi.org/10.5281/zenodo.20427212" target="_blank">https://doi.org/10.5281/zenodo.20427212</a>, 2026.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib65"><label>Sweeney et al.(2015)Sweeney, Karion, Wolter, Newberger, Guenther,
Higgs, Andrews, Lang, Neff, Dlugokencky, Miller, Montzka, Miller, Masarie,
Biraud, Novelli, Crotwell, Crotwell, Thoning, and
Tans</label><mixed-citation>
      
Sweeney, C., Karion, A., Wolter, S., Newberger, T., Guenther, D., Higgs, J. A.,
Andrews, A. E., Lang, P. M., Neff, D., Dlugokencky, E., Miller, J. B.,
Montzka, S. A., Miller, B. R., Masarie, K. A., Biraud, S. C., Novelli, P. C.,
Crotwell, M., Crotwell, A. M., Thoning, K., and Tans, P. P.: Seasonal
climatology of CO<sub>2</sub> across North America from aircraft
measurements in the NOAA/ESRL Global Greenhouse Gas Reference
Network, J. Geophys. Res.-Atmos., 120, 5155–5190,
<a href="https://doi.org/10.1002/2014JD022591" target="_blank">https://doi.org/10.1002/2014JD022591</a>, 2015.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib66"><label>Tardito Chaudhri and Stevenson(2025)</label><mixed-citation>
      
Tardito Chaudhri, A. K. and Stevenson, D. S.: Soil deposition of atmospheric hydrogen constrained using planetary-scale observations, Atmos. Chem. Phys., 25, 7369–7385, <a href="https://doi.org/10.5194/acp-25-7369-2025" target="_blank">https://doi.org/10.5194/acp-25-7369-2025</a>, 2025.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib67"><label>Thompson et al.(2022)Thompson, Wofsy, Prather, Newman, Hanisco,
Ryerson, Fahey, Apel, Brock, Brune, Froyd, Katich, Nicely, Peischl, Ray,
Veres, Wang, Allen, Asher, Bian, Blake, Bourgeois, Budney, Bui, Butler,
Campuzano-Jost, Chang, Chin, Commane, Correa, Crounse, Daube, Dibb, DiGangi,
Diskin, Dollner, Elkins, Fiore, Flynn, Guo, Hall, Hannun, Hills, Hintsa,
Hodzic, Hornbrook, Huey, Jimenez, Keeling, Kim, Kupc, Lacey, Lait, Lamarque,
Liu, McKain, Meinardi, Miller, Montzka, Moore, Morgan, Murphy, Murray, Nault,
Neuman, Nguyen, Gonzalez, Rollins, Rosenlof, Sargent, Schill, Schwarz, Clair,
Steenrod, Stephens, Strahan, Strode, Sweeney, Thames, Ullmann, Wagner, Weber,
Weinzierl, Wennberg, Williamson, Wolfe, and Zeng</label><mixed-citation>
      
Thompson, C. R., Wofsy, S. C., Prather, M. J., Newman, P. A., Hanisco, T. F.,
Ryerson, T. B., Fahey, D. W., Apel, E. C., Brock, C. A., Brune, W. H., Froyd,
K., Katich, J. M., Nicely, J. M., Peischl, J., Ray, E., Veres, P. R., Wang,
S., Allen, H. M., Asher, E., Bian, H., Blake, D., Bourgeois, I., Budney, J.,
Bui, T. P., Butler, A., Campuzano-Jost, P., Chang, C., Chin, M., Commane, R.,
Correa, G., Crounse, J. D., Daube, B., Dibb, J. E., DiGangi, J. P., Diskin,
G. S., Dollner, M., Elkins, J. W., Fiore, A. M., Flynn, C. M., Guo, H., Hall,
S. R., Hannun, R. A., Hills, A., Hintsa, E. J., Hodzic, A., Hornbrook, R. S.,
Huey, L. G., Jimenez, J. L., Keeling, R. F., Kim, M. J., Kupc, A., Lacey, F.,
Lait, L. R., Lamarque, J.-F., Liu, J., McKain, K., Meinardi, S., Miller,
D. O., Montzka, S. A., Moore, F. L., Morgan, E. J., Murphy, D. M., Murray,
L. T., Nault, B. A., Neuman, J. A., Nguyen, L., Gonzalez, Y., Rollins, A.,
Rosenlof, K., Sargent, M., Schill, G., Schwarz, J. P., Clair, J. M. S.,
Steenrod, S. D., Stephens, B. B., Strahan, S. E., Strode, S. A., Sweeney, C.,
Thames, A. B., Ullmann, K., Wagner, N., Weber, R., Weinzierl, B., Wennberg,
P. O., Williamson, C. J., Wolfe, G. M., and Zeng, L.: The NASA
Atmospheric Tomography (ATom) Mission: Imaging the Chemistry of
the Global Atmosphere, B. Am. Meteor. Soc.,
103, E761–E790, <a href="https://doi.org/10.1175/BAMS-D-20-0315.1" target="_blank">https://doi.org/10.1175/BAMS-D-20-0315.1</a>, 2022.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib68"><label>Thoning et al.(1989)Thoning, Tans, and
Komhyr</label><mixed-citation>
      
Thoning, K. W., Tans, P. P., and Komhyr, W. D.: Atmospheric carbon dioxide at
Mauna Loa Observatory: 2. Analysis of the NOAA GMCC data,
1974–1985, J. Geophys. Res.-Atmos., 94, 8549–8565,
<a href="https://doi.org/10.1029/JD094iD06p08549" target="_blank">https://doi.org/10.1029/JD094iD06p08549</a>, 1989.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib69"><label>Tolk et al.(2011)Tolk, Dolman, Meesters, and
Peters</label><mixed-citation>
      
Tolk, L. F., Dolman, A. J., Meesters, A. G. C. A., and Peters, W.: A comparison of different inverse carbon flux estimation approaches for application on a regional domain, Atmos. Chem. Phys., 11, 10349–10365, <a href="https://doi.org/10.5194/acp-11-10349-2011" target="_blank">https://doi.org/10.5194/acp-11-10349-2011</a>, 2011.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib70"><label>Trapani et al.(2025)Trapani, Marocco, Gandiglio, and
Santarelli</label><mixed-citation>
      
Trapani, D., Marocco, P., Gandiglio, M., and Santarelli, M.: Hydrogen leakages
across the supply chain: Current estimates and future scenarios,
Int. J. Hydrogen Energ., 145, 1084–1095,
<a href="https://doi.org/10.1016/j.ijhydene.2025.06.103" target="_blank">https://doi.org/10.1016/j.ijhydene.2025.06.103</a>, 2025.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib71"><label>Van Der Laan-Luijkx et al.(2017)Van Der Laan-Luijkx, Van Der Velde,
Van Der Veen, Tsuruta, Stanislawska, Babenhauserheide, Zhang, Liu, He, Chen,
Masarie, Krol, and Peters</label><mixed-citation>
      
van der Laan-Luijkx, I. T., van der Velde, I. R., van der Veen, E., Tsuruta, A., Stanislawska, K., Babenhauserheide, A., Zhang, H. F., Liu, Y., He, W., Chen, H., Masarie, K. A., Krol, M. C., and Peters, W.: The CarbonTracker Data Assimilation Shell (CTDAS) v1.0: implementation and global carbon balance 2001–2015, Geosci. Model Dev., 10, 2785–2800, <a href="https://doi.org/10.5194/gmd-10-2785-2017" target="_blank">https://doi.org/10.5194/gmd-10-2785-2017</a>, 2017.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib72"><label>Van Der Werf et al.(2025)Van Der Werf, Randerson, Van Wees, Chen,
Giglio, Hall, Roland, Mu, Binte Shahid, Barsanti, Yokelson, and
Morton</label><mixed-citation>
      
Van Der Werf, G. R., Randerson, J. T., Van Wees, D., Chen, Y., Giglio, L.,
Hall, J., Roland, V., Mu, M., Binte Shahid, S., Barsanti, K. C., Yokelson,
R., and Morton, D. C.: Landscape fire emissions from the 5th version of the
Global Fire Emissions Database (GFED5), Sci. Data, 12, 1870,
<a href="https://doi.org/10.1038/s41597-025-06127-w" target="_blank">https://doi.org/10.1038/s41597-025-06127-w</a>, 2025.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib73"><label>Vollmer et al.(2010)Vollmer, Walter, Bond, Soltic, and
Röckmann</label><mixed-citation>
      
Vollmer, M. K., Walter, S., Bond, S. W., Soltic, P., and Röckmann, T.: Molecular hydrogen (H2) emissions and their isotopic signatures (H/D) from a motor vehicle: implications on atmospheric H2, Atmos. Chem. Phys., 10, 5707–5718, <a href="https://doi.org/10.5194/acp-10-5707-2010" target="_blank">https://doi.org/10.5194/acp-10-5707-2010</a>, 2010.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib74"><label>Walter et al.(2016)Walter, Kock, Steinhoff, Fiedler, Fietzek, Kaiser,
Krol, Popa, Chen, Tanhua, and Röckmann</label><mixed-citation>
      
Walter, S., Kock, A., Steinhoff, T., Fiedler, B., Fietzek, P., Kaiser, J., Krol, M., Popa, M. E., Chen, Q., Tanhua, T., and Röckmann, T.: Isotopic evidence for biogenic molecular hydrogen production in the Atlantic Ocean, Biogeosciences, 13, 323–340, <a href="https://doi.org/10.5194/bg-13-323-2016" target="_blank">https://doi.org/10.5194/bg-13-323-2016</a>, 2016.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib75"><label>Williams and Van Noije(2008)</label><mixed-citation>
      
Williams, J. and Van Noije, T.: On the upgrading of the modified Carbon
Bond Mechanism IV for use in global Chemistry Transport Models,
Scientific Report WR-08-02, KNMI, 2008.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib76"><label>Williams et al.(2017)Williams, Boersma, Le Sager, and
Verstraeten</label><mixed-citation>
      
Williams, J. E., Boersma, K. F., Le Sager, P., and Verstraeten, W. W.: The high-resolution version of TM5-MP for optimized satellite retrievals: description and validation, Geosci. Model Dev., 10, 721–750, <a href="https://doi.org/10.5194/gmd-10-721-2017" target="_blank">https://doi.org/10.5194/gmd-10-721-2017</a>, 2017.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib77"><label>Wofsy et al.(2017)Wofsy, Daube, Jimenez-Pizarro, Kort, Pittman, Park,
Commane, Xiang, Santoni, Jacob, Fisher, Pickett-Heaps, Wang, Wecht, Wang,
Stephens, Shertz, Watt, Romashkin, Campos, Haggerty, Cooper, Rogers, Beaton,
Hendershot, Elkins, Fahey, Gao, Schwarz, Moore, Montzka, Perring, Hurst,
Miller, Sweeney, Oltmans, Hintsa, Nance, Dutton, Watts, Spackman, Rosenlof,
Ray, Hall, Zondlo, Diao, Keeling, Bent, Atlas, Lueb, and
(Deceased)</label><mixed-citation>
      
Wofsy, S., Daube, B., Jimenez-Pizarro, R., Kort, E., Pittman, J., Park, S.,
Commane, R., Xiang, B., Santoni, G., Jacob, D., Fisher, J., Pickett-Heaps,
C., Wang, H., Wecht, K., Wang, Q., Stephens, B., Shertz, S., Watt, A.,
Romashkin, P., Campos, T., Haggerty, J., Cooper, W., Rogers, D., Beaton, S.,
Hendershot, R., Elkins, J., Fahey, D., Gao, R., Schwarz, J., Moore, F.,
Montzka, S., Perring, A., Hurst, D., Miller, B., Sweeney, C., Oltmans, S.,
Hintsa, E., Nance, D., Dutton, G., Watts, L., Spackman, J., Rosenlof, K.,
Ray, E., Hall, B., Zondlo, M., Diao, M., Keeling, R., Bent, J., Atlas, E.,
Lueb, R., and (Deceased), M. J. M.: HIPPO Combined Discrete Flask and
GC Sample GHG, Halocarbon, and Hydrocarbon Data, Version 1.0,
NSF NCAR Earth Observing Laboratory, <a href="https://doi.org/10.3334/CDIAC/HIPPO_012" target="_blank">https://doi.org/10.3334/CDIAC/HIPPO_012</a>,  2017.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib78"><label>Wofsy et al.(2021)Wofsy, Afshar, Allen, Apel, Asher, Barletta, Bent,
Bian, Biggs, Blake, Blake, Bourgeois, Brock, Brune, Budney, Bui, Butler,
Campuzano-Jost, Chang, Chin, Commane, Correa, Crounse, Cullis, Daube, Day,
Dean-Day, Dibb, DiGangi, Diskin, Dollner, Elkins, Erdesz, Fiore, Flynn,
Froyd, Gesler, Hall, Hanisco, Hannun, Hills, Hintsa, Hoffman, Hornbrook,
Huey, Hughes, Jimenez, Johnson, Katich, Keeling, Kim, Kupc, Lait, McKain,
Mclaughlin, Meinardi, Miller, Montzka, Moore, Morgan, Murphy, Murray, Nault,
Neuman, Newman, Nicely, Pan, Paplawsky, Peischl, Prather, Price, Ray, Reeves,
Richardson, Rollins, Rosenlof, Ryerson, Scheuer, Schill, Schroder, Schwarz,
St.Clair, Steenrod, Stephens, Strode, Sweeney, Tanner, Teng, Thames,
Thompson, Ullmann, Veres, Wagner, Watt, Weber, Weinzierl, Wennberg,
Williamson, Wilson, Wolfe, Woods, Zeng, and Vieznor</label><mixed-citation>
      
Wofsy, S., Afshar, S., Allen, H., Apel, E., Asher, E., Barletta, B., Bent, J.,
Bian, H., Biggs, B., Blake, D., Blake, N., Bourgeois, I., Brock, C., Brune,
W., Budney, J., Bui, T., Butler, A., Campuzano-Jost, P., Chang, C., Chin, M.,
Commane, R., Correa, G., Crounse, J., Cullis, P. D., Daube, B., Day, D.,
Dean-Day, J., Dibb, J., DiGangi, J., Diskin, G., Dollner, M., Elkins, J.,
Erdesz, F., Fiore, A., Flynn, C., Froyd, K., Gesler, D., Hall, S., Hanisco,
T., Hannun, R., Hills, A., Hintsa, E., Hoffman, A., Hornbrook, R., Huey, L.,
Hughes, S., Jimenez, J., Johnson, B., Katich, J., Keeling, R., Kim, M., Kupc,
A., Lait, L., McKain, K., Mclaughlin, R., Meinardi, S., Miller, D., Montzka,
S., Moore, F., Morgan, E., Murphy, D., Murray, L., Nault, B., Neuman, J.,
Newman, P., Nicely, J., Pan, X., Paplawsky, W., Peischl, J., Prather, M.,
Price, D., Ray, E., Reeves, J., Richardson, M., Rollins, A., Rosenlof, K.,
Ryerson, T., Scheuer, E., Schill, G., Schroder, J., Schwarz, J., St.Clair,
J., Steenrod, S., Stephens, B., Strode, S., Sweeney, C., Tanner, D., Teng,
A., Thames, A., Thompson, C., Ullmann, K., Veres, P., Wagner, N., Watt, A.,
Weber, R., Weinzierl, B., Wennberg, P., Williamson, C., Wilson, J., Wolfe,
G., Woods, C., Zeng, L., and Vieznor, N.: Atmospheric Tomography Mission
(ATom)ATom: Merged Atmospheric Chemistry, Trace Gases, and
Aerosols, Version 2, <a href="https://doi.org/10.3334/ORNLDAAC/1925" target="_blank">https://doi.org/10.3334/ORNLDAAC/1925</a>,  ORNL Distributed Active Archive Center,  2021.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib79"><label>Wofsy(2011)</label><mixed-citation>
      
Wofsy, S. C.: HIAPER Pole-to-Pole Observations (HIPPO): fine-grained,
global-scale measurements of climatically important atmospheric gases and
aerosols, Philos. T. Roy. Soc. A, 369, 2073–2086,
<a href="https://doi.org/10.1098/rsta.2010.0313" target="_blank">https://doi.org/10.1098/rsta.2010.0313</a>, 2011.


    </mixed-citation></ref-html>
<ref-html id="bib1.bib80"><label>Xiao et al.(2007)Xiao, Prinn, Simmonds, Steele, Novelli, Huang,
Langenfelds, O'Doherty, Krummel, Fraser, Porter, Weiss, Salameh, and
Wang</label><mixed-citation>
      
Xiao, X., Prinn, R. G., Simmonds, P. G., Steele, L. P., Novelli, P. C., Huang,
J., Langenfelds, R. L., O'Doherty, S., Krummel, P. B., Fraser, P. J., Porter,
L. W., Weiss, R. F., Salameh, P., and Wang, R. H. J.: Optimal estimation of
the soil uptake rate of molecular hydrogen from the Advanced Global
Atmospheric Gases Experiment and other measurements, J.
Geophys. Res., 112, D07303, <a href="https://doi.org/10.1029/2006JD007241" target="_blank">https://doi.org/10.1029/2006JD007241</a>, 2007.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib81"><label>Yang et al.(2025)Yang, Jacob, Lin, Dang, Bates, East, Travis,
Pendergrass, and Murray</label><mixed-citation>
      
Yang, L. H., Jacob, D. J., Lin, H., Dang, R., Bates, K. H., East, J. D.,
Travis, K. R., Pendergrass, D. C., and Murray, L. T.: Assessment of
Hydrogen's Climate Impact Is Affected by Model OH Biases,
Geophys. Res. Lett., 52, e2024GL112445,
<a href="https://doi.org/10.1029/2024GL112445" target="_blank">https://doi.org/10.1029/2024GL112445</a>, 2025.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib82"><label>Yarwood et al.(2005)Yarwood, Rao, Way, Yocke, Whitten, and
Reyes</label><mixed-citation>
      
Yarwood, G., Rao, S., Way, R., Yocke, M., Whitten, G. Z., and Reyes, S.:
Updates to the Carbon Bond chemical mechanism: CB05, Tech. rep., U.S.
Environmental Protection Agency, EPA-HQ-OAR-2010-0799-1062, 2005.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib83"><label>Yashiro et al.(2011)Yashiro, Sudo, Yonemura, and
Takigawa</label><mixed-citation>
      
Yashiro, H., Sudo, K., Yonemura, S., and Takigawa, M.: The impact of soil uptake on the global distribution of molecular hydrogen: chemical transport model simulation, Atmos. Chem. Phys., 11, 6701–6719, <a href="https://doi.org/10.5194/acp-11-6701-2011" target="_blank">https://doi.org/10.5194/acp-11-6701-2011</a>, 2011.

    </mixed-citation></ref-html>
<ref-html id="bib1.bib84"><label>Yver et al.(2011)Yver, Pison, Fortems-Cheiney, Schmidt, Chevallier,
Ramonet, Jordan, Søvde, Engel, Fisher, Lowry, Nisbet, Levin, Hammer, Necki,
Bartyzel, Reimann, Vollmer, Steinbacher, Aalto, Maione, Arduini, O'Doherty,
Grant, Sturges, Forster, Lunder, Privalov, Paramonova, Werner, and
Bousquet</label><mixed-citation>
      
Yver, C. E., Pison, I. C., Fortems-Cheiney, A., Schmidt, M., Chevallier, F., Ramonet, M., Jordan, A., Søvde, O. A., Engel, A., Fisher, R. E., Lowry, D., Nisbet, E. G., Levin, I., Hammer, S., Necki, J., Bartyzel, J., Reimann, S., Vollmer, M. K., Steinbacher, M., Aalto, T., Maione, M., Arduini, J., O'Doherty, S., Grant, A., Sturges, W. T., Forster, G. L., Lunder, C. R., Privalov, V., Paramonova, N., Werner, A., and Bousquet, P.: A new estimation of the recent tropospheric molecular hydrogen budget using atmospheric observations and variational inversion, Atmos. Chem. Phys., 11, 3375–3392, <a href="https://doi.org/10.5194/acp-11-3375-2011" target="_blank">https://doi.org/10.5194/acp-11-3375-2011</a>, 2011.

    </mixed-citation></ref-html>--></article>
