Review of “The Response of the Amazon Ecosystem to the Photosynthetically Active Radiation Fields: Integrating Impacts of Biomass Burning Aerosol and Clouds in the NASA GEOS ESM”

There has been a growing number of studies looking at the potential impact on land carbon uptake from increased availability of diffused radiation associated with aerosol particles. Although conceptually simple, this effect is hard to quantify accurately as complex couplings between different components of the Earth system are at play. An Earth System Modelling (ESM) approach appears as a natural framework for this kind of problem, yet only a handful of studies using ESMs has been published so far. This submission from Bian et al. is therefore timely as they used the results from simulations performed with the NASA GEOS-ESM to analyze the impact of biomass burning aerosols on the Amazon rainforest gross primary productivity. The diffuse light fertilization effect from aerosols is not only uncertain, but it is also buffered by clouds as those compete with aerosols for radiation. This is in a way similar to pre-industrial natural aerosols controlling the amplitude of the anthropogenic aerosol radiative forcing. The role of clouds on the aerosol diffuse light fertilization effect has not been properly explored before and Bian et al. provide a novel quantification for this modulating effect.

The Amazon experiences fires every year, and the resulting biomass burning aerosols, together 23 with cloud particles, influence the penetration of sunlight through the atmosphere, increasing the 24 ratio of diffuse to direct photosynthetically active radiation (PAR) reaching the vegetation 25 canopy and thereby potentially increasing ecosystem productivity. In this study, we use the 26 NASA Goddard Earth Observing System (GEOS) model running with coupled aerosol, cloud, 27 radiation, and ecosystem modules to investigate the impact of Amazon biomass burning aerosols 28 on ecosystem productivity, as well as the role of the Amazon's clouds in tempering the impact. 29 The study focuses on a seven-year period (2010-2016) during which the Amazon experienced a 30 variety of dynamic environments (e.g., La Niña, normal years, and El Niño). The radiative 31 impacts of biomass burning aerosols on ecosystem productivity-call here the aerosol light 32 fertilizer effect-are found to increase Amazonian Gross Primary Production (GPP) by 2.6% via 33 a 3.8% increase in diffuse PAR (DFPAR) despite a 5.4% decrease in direct PAR (DRPAR) on 34 multiyear average. On a monthly basis, this increase in GPP can be as large as 9.9% (occurring 35 in August 2010). Consequently, the net primary production (NPP) in the Amazon is increased by 36 1.5%, or ~92 TgCyr -1 -equivalent to ~37% of the carbon lost due to Amazon fires over the seven 37 years considered. Clouds, however, strongly regulate the effectiveness of the aerosol light 38 fertilizer effect. The efficiency of the fertilizer effect is highest for cloud-free conditions and 39 linearly decreases with increasing cloud amount until the cloud fraction reaches ~0.8, at which 40 point the aerosol-influenced light changes from being a stimulator to an inhibitor of plant 41 growth. Nevertheless, interannual changes in the overall strength of the aerosol light fertilizer 42 effect are primarily controlled by the large interannual changes in biomass burning aerosols 43 rather than by changes in cloudiness during the studied period. 44 45 46 GEOS ESM that includes coupling between aerosol, cloud, radiation, and ecosystem processes. 139 To our knowledge, only one other study has used an ESM to investigate such fire impacts across 140 Amazonia (Malavelle et al., 2019), and as noted above, this study did not address the ability of 141 Amazon clouds to temper the BBaer impacts. Accordingly, our study is the first ESM-based 142 study to investigate the BBaer light fertilizer effect within a range of interannual Amazon cloud 143 levels. Together our objectives provide a full and comprehensive study of BBaer light fertilizer 144 effect in a context of potential Amazon dry season atmospheric conditions. 145 146 It is necessary to point out, however, that our study focuses only on the impact of Amazon 147 biomass burning aerosol. We do not consider the radiative impacts of other potentially important 148 aerosols. These other aerosol types have been examined in various observational studies (e.g., 149 Cirino  We mostly rely on the GoAmazon ("Green Ocean Amazon") field campaign 258 (http://campaign.arm.gov/goamazon2014/) for in situ site-level aerosol surface observations and 259 local-area vertical distribution measurements used to assess the model OA concentrations. 260 GoAmazon is an integrated field campaign conducted in the central Amazon Basin ( (AERONET) sun photometer network (http://aeronet.gsfc.nasa.gov). We also use MODIS 275 collection 6.1 level-3 AOD product (http://modis.gsfc.nasa.gov/data/dataprod/index.php), which 276 is characterized by observations with large spatial coverage. 277 278 MODIS cloud products (https://modis-atmosphere.gsfc.nasa.gov/data/dataprod/), specifically 279 total cloud fraction and cloud optical depth in liquid and ice particles, were used to evaluate the 280 model cloud simulation. We use the cloud data from MODIS collection 6.1 MYD08_D3, a level-281 3 1°×1° global gridded monthly joint product derived from the MODIS level-2 pixel level 282 products. MODIS level 2 cloud fraction is produced by the infrared retrieval methods during 283 both day and night at a 5×5 1-km-pixel resolution. Level 2 cloud optical thickness used in this 284 study is derived using the MODIS visible and near-infrared channel radiances from the Aqua 285 platform. 286 287 The satellite-derived Clouds and the Earth's Radiant Energy System product CERES-EBAF was 288 used to evaluate the GEOS simulation of radiation fields. CERES-EBAF retrieves surface 289 downward shortwave radiation (RSFC) using cloud information from more recent satellite data 290 (MODIS, CERES, CloudSat and CALIPSO) and aerosol fields from AERONET/MODIS 291 validation-based estimates (Kato et al., 2013). This global product is provided at a 1°×1° 292 horizontal resolution and covers the years 2000-2015 for both all-and clear-sky conditions. The 293 multiyear RSFC products provide both a regional and a time evolution view of radiation over 294 Amazonia. 295 Two observations-based GPP products were used to evaluate the GEOS ecosystem simulations. 296 Through upscaling using machine learning methods (Jung et al., 2020), the FluxCom GPP 297 product provides globally distributed eddy-covariance-based estimates of carbon fluxes between 298 the biosphere and the atmosphere. FluxSat GPP is estimated with models that use satellite data 299 (e.g., MODIS reflectances and solar-induced fluorescence (SIF)) within a simplified light-use 300 efficiency framework (Joiner et al., 2018). We used monthly GPP for August through October of 301 2010-2015 in this study. 302

Experiment setup 303
All experiments were run with the coupled atmosphere and land components of the NASA 304 GEOS ESM system discussed above.
Check atmospheric BB aerosol impact on plants via radiation fields during 2010-2016 Pair 2 callaer Standard all, w/ AERbb emission fixed at 2010 Check how clouds adjust the above impact We also designed a pair of experiments (callaer and cnobbaer, hereafter referred to as "pair2") to 349 address the sensitivity of the BBaer light fertilizer effect to the presence of the Amazon dry  The simulated tracer fields are compared with measurements over the Amazon in Figures 1 and  373 2. Figure 1 shows results for surface OA concentration, surface CO concentration, and the OA 374 concentration vertical profile. We focus primarily on the OA evaluation since we are interested 375 in biomass burning aerosols from fires. Figure 1a shows the comparison of surface daily OA 376 concentration between the model simulation and the GoAmazon measurements at Manaus, 377 Brazil, in 2014 (The location is indicated in Figure 2c with an open-diamond). The simulated OA 378 broadly captures the seasonal trend in OA concentrations measured at the ARM site, but it is 379 lower than observed OA values by ~24% during Sept-Oct and ~ 30% annually. For the period of 380 interest, the model simulates a large fire signal in August that is not seen in the measurements. 381 However, this strong August biomass burning signal does show up in the GoAmazon CO 382 measurements ( Figure 1b). Generally, it is challenging for a model to capture an aerosol plume, 383 particularly one from biomass burning, at the right time and location due to the aerosols' high 384 spatial inhomogeneity and short lifetime. 385 386

388
When compared with aircraft G-1 measurements over a ~2°×2° region around the center of 389 Manaus during the biomass burning season (Sept. 6 -Oct. 4, 2014) (Figure 1c), the simulated 390 vertical OA concentrations underestimate the measurements in the free troposphere but 391 overestimate them in the boundary layer, although they overlap within their standard deviations 392 for all altitudes. Here the model data have been sampled spatially and temporally along the G-1 393 flight paths. This surface OA overestimation by the model seems to contradict the model's 394 underestimation seen in Figure 1a, indicating again that capturing aerosols at the right times and 395 locations is a challenge. 396 397 Figure 2 shows the AOD (550nm) and SSA (440nm) comparison at a specific station and over 398 South America. We consider AERONET observational data at Alta_Floresta, which is located 399 close to the central Amazon fires (The location is marked in Figure 2c as a filled-in circle). The scattering. However, it is puzzling to observe the extremely low measured SSA in the beginning 405 of August given that the AOD is still low then, as shown in Figure 2a. Regionally over the 406 Amazon region, defined throughout the study as the land area within 80°W-30°W, 25°S-5°N, the 407 model-simulated AOD (0.22 in Figure 2d) during the biomass burning season generally agrees 408 with MODIS satellite retrievals (0.21 in Figure 2c). A simulated high bias is seen over the east 409 Amazon; however, though this region is in our area of interest, the bias should have only a minor 410 impact on our study given that the area is relatively bare, with little vegetation coverage. 411 412 The accurate simulation of cloud fields is also important for our study. In Figure 3     Ecosystems could still respond positively to increasing BBAOD even if the incident diffuse 494 radiation diminishes below its peak value, though for some value of BBAOD, the reduction in 495 total radiation will be large enough to overwhelm the impact of increased diffuse radiation, and 496 plant photosynthesis will be lower than that for clean sky conditions. 497 498

499
The Amazon dry season is characterized by high biomass burning aerosol loading combined with 500 low cloud cover, a good match to obtain more diffuse radiation without the loss of too much total 501 radiation. However, as we have pointed out, cloud impacts on radiation typically dominate those 502 of aerosols. To examine this, we repeated the radiation model calculations after adding, at the top 503 of the aerosol layer, a cloud layer with a cloud fraction of 1.0 and a cloud optical depth (COD) of 504 10, which is close to the mean liquid cloud COD over the Amazon dry season (Figure 3). The 505 impact on Rdir and Rdiff is quite large (Figure 6b). Without BBaer, the clouds already fill the sky 506 with abundant diffuse light that can reach the surface (i.e., CIdiff > 50%), while almost shutting 507 down the direct light (i.e., CIdir < 1%). Accordingly, for full cloud coverage, a clean sky (i.e., no 508 aerosols) would provide the best conditions for plant growth. When fires start, the diffuse light 509 declines rapidly, reducing the potential for plant growth. At BBAOD ~ 3 the two curves look 510 similar, that is essentially no radiation at the surface. 511 The simple examples in Figure 6 illustrate the complicated responses of direct and diffuse light 512 to the presence of aerosol and cloud. Measurements indicate that plant growth peaks for a 513 clearness index (CI, defined as CIdir+CIdiff) of about 0.4-0.7 for some forest ecosystems (Butt 514 et al., 2010, Letts and Lafleur, 2005). This CI range translates, based on Figure 6, to a BBAOD 515 range of about 0.3~1.5 in clear sky and 0~0.5 in cloudy-sky conditions. 516 517

How the ecosystem responds to the BBaer light fertilizer effect 518
We first examine the two pair1 experiments by taking a close look at the time series of aerosol, 519 cloud, radiation, and ecosystem responses generated at a selected site (54°W, 15°S) during Aug-520 Oct 2010 (Figure 7) (site location marked in Figure 8), with the aim of extending the general 521 522

523
understanding gained in section 3.2 to a real case study at a single site in the Amazon. This is an 524 interesting site and period, showing a large DFPAR change (Figure 8f) and providing a wide 525 variety of conditions for study -the sky alternates between clear and cloudy conditions in 526 August, is relatively clear in September, and is relatively cloudy in October, and the biomass 527 burning aerosols increase in August, peak in September, and diminish greatly in early October 528 (Figure 7). During August-September, when the atmosphere experiences biomass burning 529 pollution, the allaer (with BBAOD light fertilizer) and nobbaer (without BBAOD light fertilizer) 530 results differ significantly: DRPAR for allaer (solid line) lies below that for nobbaer (dotted-531 line), while DFPAR and GPP for allaer are generally higher than those for nobbaer. In October, 532 the sky is almost clean (i.e., low BBaer), leading to very similar results for DRPAR, DFPAR, 533 and GPP between the two experiments. Looking closer, we see that the changes of DRPAR, 534 DFPAR, and GPP between allaer and nobbaer are more prominent when the atmosphere has low 535 cloudiness and high aerosol (e.g., at the end of August), confirming both that BBaer does 536 transform some of the direct light at the surface into diffuse light and that plants are more 537 efficient in their use of diffuse light. When both cloudiness and aerosols are high (e.g., at the end 538 of September), the influence of aerosols is overwhelmed by clouds, and the impact of the 539 aerosols on radiation and the ecosystem becomes secondary .  540  541  542  543  544  545  546  547  548  549  550  551  552  553  554  555  556  557  558  559  560  561  562  563  564  565  566  567  568  569  570  571  572  We now evaluate BB aerosol impacts on radiation and ecosystem fields over the Amazon during 574 August 2010, when the aerosol has its largest impact. Figure 8 shows the simulated Amazon 575 DRPAR, DFPAR, and GPP fields from the two experiments comprising pair1 (nobbaer and 576 allaer). The distribution of DRPAR shows a clear spatial gradient, with low values in the 577 northwest and high values in the southeast, and the spatial pattern of DFPAR shows the reverse 578 pattern. These features are primarily controlled by the cloud distribution (Figure 3). Comparing 579 the nobbaer and allaer results by calculating field relative change (i.e., (allaer-nobbaer)/allaer), 580 we find that BBaer decreases DRPAR by 16% and increases DFPAR by 10% over the Amazon 581 region, with maximum local changes of up to -50% for DRPAR and 25% for DFPAR. 582 Interestingly, these maxima are not co-located, though the spatial patterns of perturbations do 583 agree with each other. The mismatch in the locations of the maxima in the difference fields 584 implies a nonlinear response of direct and diffuse light to aerosol and cloud particles (see section 585 3.2). In response to the inclusion of BBaer, the Amazon GPP increases by 10%. That is, the 586 increase in GPP stemming from the increase in the diffuse light fraction overwhelms a potential 587 reduction in GPP from a reduction of total PAR. 588 589 We also examine the multi-year (2010-2016) BBaer impacts on net primary production (NPP), 590 that is, the rate at which carbon is accumulated (GPP) in excess of autotrophic respiration. In 591 essence, NPP can be considered a proxy for the net plant sink of atmospheric carbon. Figure 9  592 shows monthly and long-term averaged NPP over the Amazon Basin from the two experiments 593 comprising pair1. The monthly change of NPP (i.e., dNPP = NPP(allaer) -NPP(nobbaer)) is 594 shown in the figure as a green line. Each year, during the August-September period when BBaer 595 is high and cloudiness is low over the Amazon, BBaer is seen to enhance NPP. The percentage 596 difference of annually-averaged NPP (dNPP/NPP(nobbaer)*100) in % is 4.2, 0.06, 1.9, 0.5, 1.3, 597 1.9, and 1.0 for the seven studied years. That means the BBaer-induced NPP increases range 598 from 5 TgC yr -1 or 0.06% (2011) to 278 TgC yr -1 or 4.2% (2010), with a seven-year average of 599 92 TgC or 1.5%. This is equivalent to storing 92TgC annually within the Amazon ecosystem 600 during the studied period. 601  (cloud cover 0.1-0.3, 0.3-0.6 and >0.6), and all-sky conditions based on gridded daily cloud 624 cover over the Amazon region. Figure  aerosols. The correlation presumably stems from the fact that biomass burning aerosols increase 636 the diffuse PAR reaching the canopy (dotted pink line) although they decrease the total PAR 637 (dotted purple line) via decreasing direct PAR (Table S1a). This aerosol-radiation-GPP 638 relationship is seen to vary with cloud amount, with clouds acting to reduce the aerosol impact; 639 both the diffuse radiation and the GPP show larger changes with BBAOD under clear sky 640 conditions. The overall (i.e., all-sky) aerosol impact on dGPP is similar to that for a cloud 641 coverage of 0.3-0.6, presumably because the averaged cloud coverage over the Amazon during 642 the studied period is roughly in that range. 643 644 Figure 10 and Table S1e show that on an interannual (dry season) basis, the aerosol light 645 fertilizer effect differed the most between 2010 and 2011 (i.e., the dGPP was 8.   conditions for every day of every year starting in 2011. Here we study the sensitivity of the 660 aerosol light fertilizer effect to a unit change of BBAOD. That is, on a daily basis, the sensitivity 661 of a variable X to a change in the biomass burning AOD is calculated as: ddX/dBBAOD = 662 ((dX)1-(dX)2)/(BBAOD1-BBAOD2). Here, the X represents GPP, DRPAR, and DFPAR, and the 663 subscripts 1 and 2 represent the pair1 or pair2 experiment, respectively. 664 665 ddX/dBBAOD is computed on a gridded daily basis over August-September of 2011-2016. 666 The calculations are then catalogued according to daily cloud cover fraction -we combine 667 the results within each of 10 cloud fraction bins (0-0.1, 0.1-0.2, …, 0.9-1.0). To examine the 668 maximum impact of interannual cloud change during our study period, the binned 669 ddX/dBBAOD vs. CLDFRC relationship is also computed separately from daily (August-670 September) values in 2013 and from corresponding daily values in 2015, as these are the 671 years for which monthly cloud cover is around the maximum (0.44) and minimum (0.35), 672 respectively ( Figure 10 and table S1e). 673 674 Figure 11 shows the results. An almost linear relationship is seen between the ddX/dBBAOD 675 values and cloud cover fraction. BB aerosols increase GPP in clear sky conditions (e.g., 29.6 676 kgm -2 s -1 ) but decrease it under full cloudiness conditions (e.g., -5.8 kgm -2 s -1 ). The cloud fraction 677 at which BB aerosol switches from stimulating to inhibiting plant growth occurs at ~0.8. Cloud 678 conditions thus not only affect strongly the strength of the aerosol light fertilizer effect but can 679 also change the fundamental direction of the effect. The lines produced for the three different 680 study periods are fairly similar, indicating that the relationship of ddX/dBBAOD to CLDFRC is 681 fairly stable within the range of cloud cover seen over the Amazon during the period of interest. 682