Interactive comment on “Potential sensitivity of photosynthesis and isoprene emission to direct radiative effects of atmospheric aerosol pollution”

Abstract. A global Earth system model is applied to quantify the impacts of direct anthropogenic aerosol effective radiative forcing on gross primary productivity (GPP) and isoprene emission. The impacts of different pollution aerosol sources (anthropogenic, biomass burning, and non-biomass burning) are investigated by performing sensitivity experiments. The model framework includes all known light and meteorological responses of photosynthesis, but uses fixed canopy structures and phenology. On a global scale, our results show that global land carbon fluxes (GPP and isoprene emission) are not sensitive to pollution aerosols, even under a global decline in surface solar radiation (direct + diffuse) by  ∼ 9 %. At a regional scale, GPP and isoprene emission show a robust but opposite sensitivity to pollution aerosols in regions where forested canopies dominate. In eastern North America and Eurasia, anthropogenic pollution aerosols (mainly from non-biomass burning sources) enhance GPP by +5–8 % on an annual average. In the northwestern Amazon Basin and central Africa, biomass burning aerosols increase GPP by +2–5 % on an annual average, with a peak in the northwestern Amazon Basin during the dry-fire season (+5–8 %). The prevailing mechanism varies across regions: light scattering dominates in eastern North America, while a reduction in direct radiation dominates in Europe and China. Aerosol-induced GPP productivity increases in the Amazon and central Africa include an additional positive feedback from reduced canopy temperatures in response to increases in canopy conductance. In Eurasia and northeastern China, anthropogenic pollution aerosols drive a decrease in isoprene emission of −2 to −12 % on an annual average. Future research needs to incorporate the indirect effects of aerosols and possible feedbacks from dynamic carbon allocation and phenology.

each section, we firstly present changes at the global-scale, afterwards we focus on changes in five key regions: eastern North America, Eurasia, north-eastern China, the north-western Amazon Basin and central Africa. We agree with Referee #2 that comparisons among key regions were not clear from the figures showing the spatial distribution of annual/seasonal changes. Hence, we added a new table in the revised manuscript (Table 5) that presents, for each of the five key regions, absolute and percentage changes in annual average surface radiation, canopy temperature, GPP and isoprene emissions. For the five key regions, absolute and percentage changes in seasonal averages are reported in the Supplementary Material (Table S3 and Table S4). In the revised manuscript, the comparison between key regions is now mostly based on results gathered in Table 5. The methodology to compute absolute and percentage differences in annual and seasonal averages over selected key regions is presented in Section 2.2 (pag. 7-8, ll. 231-237).
2. p. 25446, lines 5-6: It's unclear from Figure 4 that the eastern US shows much larger of an increase in diffuse radiation than over China for example (especially looking at panel (i)). This point seems important further on in the article, so I think it deserves further clarification. Table 5 should now help in clarifying this point. Table 5 shows a larger increase in annual average diffuse radiation over eastern North-America compared to Eurasia and north-eastern China due to all anthropogenic aerosols (pag. 13, ll. 415-420):

Authors:
"The eastern North America shows the largest increase in annual diffuse radiation due to all anthropogenic aerosols (+8.6 W m −2 ; +6.2 %), followed by north-eastern China and central Africa, which experience similar changes (∼ +7.4-7.9 W m −2 ; ∼ +5.7 %). Over the eastern North-America, the increase in diffuse radiation maximizes during boreal summer (+13.6 W m −2 ; +8.9 %), with changes that are 1.6-5.7 W m −2 (1.9-C12390 3.3 %) higher that those observed over north-eastern China and Eurasia (Table S3 in the Supplementary Material)." However, in response to non-biomass burning aerosols, eastern North-America, Eurasia and north-eastern China show similar increases in diffuse radiation (pag. 17, ll. 568-570): "In response to aerosol pollution from non-biomass burning sources Europe and China show a large decrease in annual average direct radiation (−24-26%), but a similar increase in diffuse radiation (+3-5 %) as eastern North America (Table 5).".

p. 25446, lines 8-11:
The authors state that biomass burning aerosol drive the decrease in several regions (in the range of -6 to -28 W m −2 ), but as I look at Figure  4 over the regions named, it seems to me that subtracting the industrial sources also result in decreases on the order of -6 to -12 W m −2 and larger. This seems especially true when looking at the seasonal results in Figure S6. Am I misinterpreting the plots?
Authors: Referee #2 is correct. In the revised manuscript, the insertion of Table 5 should make this point clearer and illuminate the comparison of impacts of biomass burning and industrial sources on surface radiation. In industrialized key regions (i.e., eastern North America, Eurasia and north-eastern China), industrial aerosols (nonbiomass burning aerosols) mostly drive changes in surface radiation (pag. 13, ll. 415-425). On the contrary, in biomass burning key regions (i.e., the north-western Amazon Basin and central Africa), biomass and non-biomass burning aerosols share a similar contribution to changes in surface radiation (pag. 13, ll. 425-429).

Section 3.3.1 and 3.2.1:
To pick up on this a little more, I also had some trouble with Section 3.3.1. Many of the conclusions here seemed to depend on contrasting the magnitude of certain effects over various regions. However, when I would try to corroborate the statements by consulting the Figures myself, in some cases the magnitudes C12391 didn't appear to be all that different. This might have to do with the Figures themselves, or maybe this could be improved by refocusing Section 3.2.1. In some cases, perhaps (re-?) stating some of the actual values would help.

Authors:
We agree with Referee #2 that figures showing the spatial distribution of annual/seasonal changes between the control and the sensitivity simulations do not provide a proper support to compare the magnitude of aerosol-driven impacts in the five key regions. To answer to this point and underpin discussion of results, in the revised manuscript we added Table 5 (plus Table S3 and Table S4 in the Supplementary Material) to show changes (absolute and percentage changes) in annual (seasonal in the Supplementary Material) average surface radiation, canopy temperature, GPP and isoprene emissions. In the revised manuscript, the quantitative discussion of results is now mostly based on values summarized in these tables (i.e., Table 5 in the main text, Table S3 and Table S4 in the Supplementary Material). Figure 4 how the increase in diffuse radiation over the eastern US is that much larger than over China and parts of Europe (as I mentioned above). Moreover, it's not at all convincing from Figure 5 that SAT over the eastern US is "reduced". There is a very small isolated patch of blue, but there is no hatching anywhere to denote significance, and most of the region is blank. I'm also confused as to what is "contrary" about Europe and China experiencing a strong reduction in total and direct radiation. Panel 4a and 4b show the US experiences comparable decreases in total and direct radiation as for parts of Europe, and maybe China. Maybe part of this confusion can be clarified by better summary of the results of Figure 4 in Section 3.2.1?

Authors:
We attempt to avoid confusion via Table 5, plus Table S3 and Table S4 in the Supplementary Material. As described in point 2, anthropogenic pollution aerosols drive a larger increase in annual average diffuse radiation over eastern North-America C12392 compared to Eurasia and north-eastern China (pag. 13, ll. 415-420). However, nonbiomass burning aerosols (non-BBAs) drive similar increase in diffuse radiation in eastern North-America, Eurasia and north-eastern China (pag. 17, ll. 568-570).
In terms of total and direct radiation, due to non-BBAs, Eurasia and north-eastern China undergo the largest reduction in total and direct radiation. Over Eurasia and north-eastern China, decreases in total and direct radiation maximize during boreal summer, with changes that double those observed over eastern North-America (pag. 13, ll. 423-425).
6. p. 25450, lines 22-23: I can see from Figure 4 how it might be true that the increase in diffuse radiation over the Amazon is weaker than over central Africa -but it doesn't seem that different, either. As a matter of fact, Section 3.2.1 places the two regions in the same sentence within the same range . . . So it's not clear how the statement "the Amazon basin experiences a weaker increase in diffuse radiation" can be all that significant. Again, this might be helped by better structuring Section 3.2.1 to correspond to the conclusions being made here in Section 3.3.1 (and/or by referring to exact values over specific regions, for diffuse and direct radiation separately). Likewise, the "larger cooling" experienced by the Amazon basin compared to central Africa (Figure 5) doesn't appear notable to me either. In Panel 5a, they have roughly the same amount of area that is hatched as significant. This statement seems important to their conclusions about how "cooling dominates in the Amazon basin", but as is, I think the authors need to do a better job showing that this is true.

Authors:
In the revised manuscript, we reformulate our hypothesis regarding aerosoldriven effects on tropical regions (Sec. 3.4, pag. 13-15). In the model, photosynthesis and stomatal conductance are coupled through the Farquhar-Ball-Berry approach. Direct radiative forcing (DRF)-driven increases in photosynthesis and GPP are associated with increases in canopy conductance and relative humidity via increased transpi-C12393 ration. Due to BBAs, the north-western Amazon Basin records the largest increase in transpiration efficiency and, as a corollary, the largest decrease in canopy temperature (−0.31 K; −0.10 %), which is ∼ 0.1 K larger than the decrease in canopy temperature over central . We name this as "bio-meteorological effect" since reductions in the canopy temperature observed in the north-western Amazon Basin represents a positive feedback on plant productivity (further increases) in response to the DRF-driven increases. The same bio-meteorological effect (i.e., robust decrease in canopy temperature and corresponding GPP enhancement) seems to operate also in central Africa and north-eastern China; these regions undergo additional substantial robust reductions in direct radiation. In central Africa, the analysis of seasonal changes in GPP reveals that enhancement in GPP maximizes in boreal autumn, together with decrease in canopy temperature, while reductions in direct radiation maximizes in boreal summer (pag. 16, ll. 524-529).
7. p. 25451, lines 12-16: Again, given the results that have been presented, I'm not yet convinced that the different mechanisms for each region (light scattering over Eastern US; reductions in direct radiation in Europe and China; cooling in the Amazon Basin) could have been established from the present model results alone. In my opinion, the arguments leading up to this based on the present model results alone have not been clearly developed.
Authors: Please see Responses to points (1)-(6) above. We agree with Referee #2 that the way our original ideas were presented may not have been conclusively supported by the simulations results available to us. The new Table 5 makes the key drivers and processes across regions more quantitatively apparent and transparent. Of course, more than one mechanism operates in each region, and a confounding issue is that the mechanisms are not independent of each other. Therefore, given the quantitative data available to us from the completed global simulations e.g. as presented in Table 5, we identify the predominant mechanism in each region while fully C12394 recognizing the complexity of aerosol-meteorology-vegetation interactions.
As a final note, we expect the analyses to become increasingly complex when we turn on the dynamic carbon allocation and prognostic phenology. Therefore, in our on-going project work, we are developing a standalone version of YIBs that includes a fully coupled atmospheric radiative transfer scheme, which will be applied in our future studies.  (Table 4) to present quantifiable statistics regarding model evaluation against observations: MODIS for coarse aerosol optical depth (AOD), and global FLUXNET-derived gross primary productivity.
5. Section 3.2.1: p. 25445, lines 14-16: Should the authors clarify when they say "slightly affected" or "highly sensitive" that they are referring to the relative change (%)?