This paper by Do et al. examines the differences in historical trends and controlling factors of isoprene emissions across six modelling approaches, and highlights the importance of better parameterizing the CO2 effect, as well as the global/regional PFT distribution and LULCC effects, on isoprene emissions. The study indicates that the influence of  CO2 and LULCC effects on long-term isoprene emissions surpasses climate factors, which is currently not well understood. As the increase of CO2, LULCC and temperature are happening together,  I feel this work comparing these controlling factors is quite important.

Overall, I found the quality of the work strong, with appropriate  and justified methods.  Hence, the conclusions are strongly supported by the evidence provided.  The suggestions I make below are primarily to increase clarity and conciseness but in themselves are minor things that don’t significantly question the findings of this work.

Lines 17-18: The influence of isoprene on the Earth’s radiative balance, as you very nicely described in the introduction (lines 51-56). Therefore, I suggest you rewrite this sentence by clarifying the indirect effect of isoprene on the radiative balance.

Lines 375-376: From a visual analysis, I would say that VISIT-S3(G1997) and UKESM1-0-LL(P2011) show higher emissions in the northwestern Amazon, which is interesting and in line with satellite retrievals of isoprene concentrations over this region (doi:10.1029/2021JD036181). I suggest mentioning/citing this here.

Lines 376-377: I do not see a hot spot of isoprene emission in the northern Amazon for the simulations with GFDL-ESM4(G2006). It seems there is a hot stop in the central and western Amazon, but it is a weak sign compared to other simulations. Was it just a visual analysis?

Lines 537-540: I found this confusing. Wasn’t the precipitation dataset the same for both models? Can you please explain why you have increased precipitation and reduced precipitation in the same Amazon region? Maybe I missed this point, but I think it is worth explaining it here.

Line 678: it is not usual to present a new figure/result in the discussion section. I suggest moving figure 14 to the results.

Lines 619-691: In general, this whole section seems more like results (some repetition from the results) than discussion. I suggest integrating it to the results or reducing it.

Line 714: Do you mean “both sets of modelling studies”? You have cited a few studies in the previous sentence. I suggest you rewrite this sentence for clarificaiton.

The paper egusphere-2024-2313 presents a well-designed study to evaluate historical isoprene emission trends and their controlling factors from online simulations of CMIP6 Earth System Models and one offline simulation using the VISIT dynamic vegetation model. The endeavor of this work to elucidate the historical changes of global isoprene emissions during 1850-2014 due to climate, land cover changes and CO2 concentrations is deeply appreciated. The concurrent use of VISIT facilitates efficient exploration of isoprene emission sensitivity to various driving factors. Specifically, VISIT offers a way to compare the CO2 effects on isoprene emissions that are incorporated in three of the CMIP6 models, since VISIT includes the fertilization effect of CO2, but not the direct inhibition by CO2. The main result of the model comparison is that CO2 concentrations and land cover changes are the primary drivers for isoprene emission trends in CMIP6 models that include CO2 effects. This is an important finding and emphasizes the need for accurate representation of land cover changes and CO2 effects in Earth System models. The paper is organized in a clear fashion and the presentation quality is excellent. The discussion of model uncertainties is quite comprehensive and future directions are in accordance with the findings of this work. However, the discussion should be further expanded to address the links to plant phenological models, which could support the understanding of historical and future changes in plant species composition and their emission behavior due to climate changes.

Specific Comments:

1.) Evaluation of the model sensitivities to climate variables, i.e. temperature, shortwave radiation, and precipitation, would benefit from supplementing their global and probably also regional trends during 1850-2014. Global trends of temperature for each model are provided in Table 3. At least, the global trends of shortwave radiation and precipitation should be provided in tabular form. It is also suggested to present their regional trends for regions with highest isoprene emissions or on a global map. This inclusion would also facilitate the interpretation of Figure 12. Obviously, precipitation trends are not consistent among models. On page 25, lines 538-540, it is stated that different precipitation trends across the models gave rise to different isoprene emission trends in Amazonia.

2.) Figure 6 and 8 shows for VISIT(G1997) the absence of interannual variability in the response of isoprene emissions to the drivers of land cover change and climate during 1850-1900, while the CMIP6 models display variability also during this historical period. Is this due to different land cover and the climate reanalysis data used in VISIT or are there any compensating mechanisms at work that balance the emission response? The spin-up phase should have been sufficiently long. Given the high sensitivity of isoprene emissions to temperature changes (Figure 13), the missing variability in VISIT during 1850-1900 looks like an artifact. On page, lines 478-480: “VISIT(G1997) notably exhibits lower interannual variation than the CMIP6 models.” This discrepancy should be explained by analyzing the causality for the weak response to interannual changes in climate variables.

3.) Precipitation is found to play a minor role in controlling long-term trends of isoprene emissions in most CMIP6 models. The effect of extreme climate, such as drought stress, on emissions of isoprene from the vegetation is briefly mentioned in the Introduction, but not further explored in this study. Jiang et al. (2018) developed the MEGAN3 approach of using a photosynthesis parameter and a soil wetness factor to determine the drought activity factor, which improves the simulation of isoprene emissions in non-drought and drought periods at the Missouri Ozarks AmeriFlux (MOFLUX) field site. Their global simulation underpins the importance to simulate the drought-induced response of isoprene emission accurately in Earth System models. MEGAN v2.1 considers the effects of drought using soil moisture and wilting point. Different wilting point data in MEGAN can lead to substantially different outcomes for the effect of drought. The wilting points from Chen and Dudhia (2001) that is default for MEGAN v2.1 seem to be too low, and consequently the model does not capture the 2011 or 2012 drought effect on isoprene emissions (e.g., Sindelarova et al. 2014).

4.) Changes in tree species distribution and composition in response to climate change impact the amount and composition of BVOC emissions. The isoprene emission rate varies significantly across plant species. The assignment of emission factors to certain PFT is often not unambiguous. In this regard, the use of plant-specific emission factors is expected to better reflect the impact due to changes of individual climate drivers. Notably, emission factors (emission potentials) can vary significantly even within the same genus (Karl et al., 2009; Satake et al., 2024). Dani et al. (2014) suggested that the trait of isoprenoid emission in evergreen plants can be lost during evolution in favor of more storable compounds (monoterpenes) to better cope with repeated and prolonged stress. The proportion of isoprene-emitting tropical trees appears to increase with mean annual temperature but to decrease with length of dry season (Taylor et al., 2018). The discussion in section 4.1.1 should be extended to emphasize the aspects of species composition changes and plant phenological changes due to changing climate, citing the above-mentioned literature.

Technical Corrections:

Table 1: there are missing entries for S2 and S3.

Figure 6: it would be beneficial for the reader to insert a thin black horizontal line at zero.

References:

Chen, F., and Dudhia, J.: Coupling an advanced land surface-hydrology model with the penn State–NCAR MM5 modeling system. Part I: model implementation and sensitivity, Mon. Wea. Rev., 129, 569–585, 2001.

Dani, K. G. S., Jamie, I. M., Prentice, I. C., Atwell, B. J.: Evolution of isoprene emission capacity in plants, Trends Plant Sci., 19(7), 439–446, doi:10.1016/j.tplants.2014.01.009, 2014.

Jiang, X., Guenther, A., Potosnak, M., Geron, C., Seco, R., Karl, T., Kim, S., Gu, L., and Pallardy, S.: Isoprene emission response to drought and the impact on global atmospheric chemistry, Atmos. Environ., 183, 69–83, doi:10.1016/j.atmosenv.2018.01.026, 2018.

Karl, M., Guenther, A., Köble, R., Leip, A., and Seufert, G.: A new European plant-specific emission inventory of biogenic volatile organic compounds for use in atmospheric transport models, Biogeosciences, 6, 1059–1087, doi:10.5194/bg-6-1059-2009, 2009.

Satake, A., Hagiwara, T., Nagano, A. J., Yamaguchi, N., Sekimoto, K., Shiojiri, K., and Sudo, K.: Plant molecular phenology and climate feedbacks mediated by BVOCs, Annu. Rev. Plant Biol. 2024. 75, 605–627, https:// doi.org/10.1146/annurev-arplant-060223- 032108, 2024.

Taylor, T. C., McMahon, S. M., Smith, M. N., Boyle, B., Violle, C., et al.: Isoprene emission structures tropical tree biogeography and community assembly responses to climate, New Phytol., 220, 435–446, doi:10.1111/nph.15304, 2018.