I am submitting an unsolicited comment on, "Direct observations of NO_x emissions over the San Joaquin Valley using airborne flux measurements during RECAP-CA 2021 field campaign" (https://doi.org/10.5194/acp-2023-3, by Zhu et al.

Having considerable experience with the topic at hand (ariborne NO_x flux estimates from the San Joaquin Valley), I am concerned by the authors' lack of reference to the considerable volume of prior work that has occurred before them on this very topic.  However, mine is not only a complaint about the professional cordiality of such an oversight, it has direct technical bearing on the accuracy of the work. 

The main concern with the work involves the use of a model to determine the boundary layer depths in the Southern San Joaquin Valley, a numerically challenging parameter given the complex topograpy and unconstrained irrigation patterns on the valley floor (Bianco et al, 2011; Faloona et al., 2020; Alexander et al., 2022; Caputi et al., 2022).

All of these works (and other references below) show that average summertime boundary layer heights in the region are about 500-600 m agl, which means that the lowest flight leg (~300m) would be at or above z/z_i ~ 0.5.  Figure 2 of the submitted manuscript clearly shows that the authors are using boundary layer depths (from the HRRR model) that are 1,200 m or greater, which is twice as deep as the typicall boundary layer in the Southern San Joaquin Valley.  This dramatically alters their analysis, and should not be overlooked. Alexander et al. (2022), in particular, shows how WRF model boundary layer heights are often several hundred meters deeper than observed in this region. 

At issue is the NO_x flux divergence across the depth of the boundary layer because the measurements at the lowest aircraft flight altitude have to be extrapolated to the surface. In Line 142 the authors erroneously state that, “The vertical divergence is a result of multiple processes, including net in-situ production or loss, storage, and horizontal advection.” I believe this statement to be misleading.  These are all terms in the mean scalar (NO_x) budget equation, but one term does not necessarily result from the action of the other terms directly. These terms are *not* in the governing equation for a scalar flux divergence, which can be derived.  For example, net photochemical production will not “cause” a flux divergence if it is evenly distributed within the ABL, or if the time scale of production is much longer than the turbulent mixing (Damköhler number ≪1).

From the perspective of the boundary layer dynamics, the flux divergence is to first order the consequence of a surface flux (what is trying to be measured) and the entrainment flux. The latter can be estimated from the knowledge of the jump in [NO_x] across the top of the boundary layer and the many reports of midday entrainment now established in the literature (Karl et al., 2013; Trousdell et al., 2016; Trousdell et al., 2019; Caputi et al., 2022).  For this reason, I would recommend showing profiles of mean flight data for water vapor, potential temperature, NO_x, and turbulent kinetic energy (or <w²>) to determine where the boundary layer top is, and then to estimate the entrainment flux of NO_x, which will be positive due to dilution. Because NO_x is a scalar similar to water vapor because they both have a source at the surface and a dilution flux at the top, it could very well mean that the flux divergence is minimal and the extrapolation will not lead to such a large change in flux from that measured by the aircraft. 

Extrapolating Figure 2 up to z_i shows a downward flux of about -0.5 mg/m²/hr (indicating a source of NO_x above the top of the ABL), which is further evidence that the z_i being used is erroneous. Moreover, the exclusion of 3 out of the 7 flux data points (top one and bottom two in Fig. 2) in fitting the line to get the flux divergence is not rationalized and appears arbitrary and incorrect.   

At the very least I believe consideration of this one very important point about the boundary layer depths should be addressed before proceeding to publication. Nevertheless, I will put my other comments in a separate section of this review.

References that the authors did not cite to pertinent work I was involved in include:

1.)  Trousdell, Justin F., Stephen A. Conley, Andy Post, and Ian C. Faloona. Observing entrainment mixing, photochemical ozone production, and regional methane emissions by aircraft using a simple mixed-layer framework. Atmospheric Chemistry and Physics, 16(24): 15433-15450, 2016.

www.atmos-chem-phys.net/16/1/2016/.

2.)  Trousdell, J. F., Caputi, D., Smoot, J., Conley, S. A., and Faloona, I. C.: Photochemical Production of Ozone and Emissions of NO_x and CH₄ in the San Joaquin Valley, Atmos. Chem. Phys., 19, 10697–10716, https://www.atmos-chem-phys.net/19/10697/2019/, 2019. 

3.)  Faloona, I.C., S. Chiao, A. Eiserloh, R.J. Alvarez II, G. Kirgis, A.O. Langford, C.J. Senff, D. Caputi, A. Hu, L. T. Iraci, E. L. Yates, J. E. Marrero, J.-M. Ryoo, S. Conley, S. Tanrikulu, J. Xu, and T. Kuwayama, The California Baseline Ozone Transport Study (CABOTS), Bulletin of American Meteorological Society, v101, n4, 2020: E427–E445.

https://doi.org/10.1175/BAMS-D-18-0302.1

4.)  Alexander, G. Aaron, Heather A. Holmes, Xia Sun, Dani Caputi, Ian C. Faloona, and Holly J. Oldroyd. "Simulating land-atmosphere coupling in the Central Valley, California: Investigating soil moisture impacts on boundary layer properties." Agricultural and Forest Meteorology 317 (2022): 108898.  https://doi.org/10.1016/j.agrformet.2022.108898

5.)  Caputi, D., Trousdell, J., S. Mehrotra, S. Conley, G. A. Alexander, H. J. Oldroyd, and I. Faloona, Entrainment Rates and Their Synoptic Dependence on Wind Speed Aloft in California's Central Valley, Boundary-Layer Meteorology, 2022.   https://link.springer.com/article/10.1007/s10546-022-00770-1

6.)  Pan, K., and I. Faloona, The Impacts of Wildfires on Ozone Production and Boundary Layer Dynamics in California’s Central Valley, Atmospheric Chemistry & Physics, 22, 9681–9702, 2022.  https://doi.org/10.5194/acp-22-9681-2022

and older references on the boundary layer dynamics of the San Joaquin Valley that were overlooked:

7.)  Bianco, L., I. V. Djalalova, C. W. King, and J. M. Wilczak. "Diurnal evolution and annual variability of boundary-layer height and its correlation to other meteorological variables in California’s Central Valley." Boundary-layer meteorology 140 (2011): 491-511.  https://link.springer.com/article/10.1007/s10546-011-9622-4

Other Specific Technical Comments:

We believe the authors do a disservice to the ongoing scientific debate by simply stating the results of Almaraz et al. (2018) and then stating that these are contradicted by a different study by Guo et al. (2020). Further evidence supporting the hypothesis of much larger NO_x emissions in the SJV than modeled by Guo et al. (2020) have been adduced in the works of Sha et al. (2021), Wang et al. (2021), and Luo et al. (2022). 

We also feel like there was a missed opportunity in comparing similar soil NO_x flux estimates, such as those presented in the exact same area by Trousdell et al. (2019) (their Fig. 3) but in units of kg N ha^-1 yr^-1.  Other urban emissions to compare to include those found in Karl et al. (2023) and Vaughan et al., (2021), but in units of nmol m^-2 s^-1.

Line 127: What are you considering large-scale here?  The cospectrum of Fig. 3S seems to indicate peaks near 2.2 km, and a secondary peak near 8.5 km, so why filter at 2 km? 

Line 136:  The Högström (1988) similarity functions apply only to the surface layer winds, not throughout the depth of the entire boundary layer. 

Line 185/189:  The undersampling of turbulent eddies would lead to a systematic error, but it should not lead to a sign change (downward flux). Unless you are suggesting that some eddies are carrying downward fluxes, different than other sizes?  Similarly, there is no reason to believe that incomplete sampling would lead to random large spikes in upward fluxes.  Again, this is a function of instrument noise, or actually surface regions of high fluxes, not systematic errors due to frequency sampling limitations.   

Line 199:  “due to higher altitude above the ground relative to the boundary layer height” is a statement that is difficult to understand.  Are the winds stronger “at the foothills”? This is data that you should have measured.  Are the “foothills” flight tracks at higher altitude than the others? 

Thank you for considering my opinions about this study,

Ian Faloona

Additional References

Tong Sha et al., “Impacts of Soil NO_x Emission on O₃ Air Quality in Rural California”, Env. Sci. & Tech., 2021, 55, 7113−7122.

Yi Wang et al., “Improved modelling of soil NO_x emissions in a high temperature agricultural region: role of background emissions on NO₂ trend over the US”, Env. Res. Lett., 2021, 16,  084061.

Karl, T., et al., “High urban NO_x triggers a substantial chemical downward flux of ozone”, Sci. Adv., 2023, 9, eadd2365.

Lina Luo et al., “Integrated Modeling of U.S. Agricultural Soil Emissions of Reactive Nitrogen and Associated Impacts on Air Pollution, Health, and Climate”, Env. Sci. & Tech., 2022, 56, 9265−9276.