This study employs a globally validated mechanistic ecohydrological model to simulate the ecohydrological responses to variations in solar radiation. A notable aspect of the author’s approach is the avoidance of the simplistic method of merely adjusting incoming shortwave radiation. Instead, the study utilizes climate sensitivities derived from CMIP6 scenario simulations as inputs for the ecohydrological model. This innovative experimental design accounts for more realistic climate feedbacks.

I find the analysis framework to be novel. While I have no major comments, I offer a few minor comments:

• Line 65, could you elaborate on how the term “pure” solar radiation change is defined? Given that the G1 simulation introduces some degree of land-atmospheric feedback at the local scale, it would be beneficial to clarify the meaning of “pure.”
• Line 90 requires further explanation. Why is it impossible?
• Line 96 raises a question regarding the relevance of reporting climate sensitivities for the third case, as it seems tangential to the ecohydrological model. Conversely, why didn’t you simulate the ecohydrological response to long-term climate feedback?
• Line 101, does the G1 experiment solely modify surface solar radiation? I seek additional clarification on why most induced changes in climate variables are directly associated with radiation changes in this experiment (Line 117). What distinguishes G1 from the method mentioned in Line 80 (i.e., the first way)?
• Line 118 calls for an explanation of the model selection process.
• In Table 1, it’s noted that only one model (IPSL-CM6A-LR) is used to calculate short-term climate sensitivity and SRnc. When comparing ecohydrological responses between SRsc and SRnc, does this imply the exclusive use of this model? If not, how was the structural uncertainty among different models addressed? Moreover, including a line to detail the calculation of climate sensitivity based on the four experiments would be helpful.
• Lines 132 and 133 necessitate further elaboration on the rationale for selecting the first decade and the last 50 years for computation.
• Line 158, the term “no climate feedback” might be misleading since this scenario does involve feedback.
• Line 176, it would be interesting to know why leaf area index was not considered as a vegetation variable.
• There seem to be typographical errors at Lines 251 (fig. 3g?) and 254 (Fig. 3a?). Please verify the correct figure references.

The study develops a methodology to investigate the sensitivity of ecosystem response to changes in solar radiation, modified through geoengineering. This geoengineering link is very unclear both space and time scales-- whether or not actual geo-engineering implemented in a model, and what short- and long-term sensitivities are. The introduction should give a more thorough explanation of the problem, and the methods section should be more clear. Perhaps each of the four sensitivity approaches can be written in a separate paragraph.

The results show that ecosystem response is relatively small, some fraction of the changes in ET, which is understandable. Although this sort of muted response is expected for the given sensitivities in the model forcing of 115 sitess, but I am unsettled with the way climate sensitivity is introduced based on the mean annual changes as reported in Figure 1. The future climate will bring a lot more variability across the globe, and that variability would potentially have a wider range of sensitivities than the sensitivity calculated using the global averages. In addition, sensitivities may also very more significantly when at least growing and non-growing season sensitivities are separated, as opposed to using sensitivities derived using global mean annual outputs. It is also unclear how climate sensitivities were implemented in each of the 115 sites. Are you doing something like bias correction in a time series obtained from a GCM somehow. Anyhow I detailed some of these issues below. Could be an interesting study after making all these issues clear.

More comments:

What is the scale of you modification of the solar radiation, local or global, because your simulations seem to focus on local hydrometeorology but when you are discussing those 4 scenarios you are mixing scales, local versus global and regional. Please give example of this solar radiation manipulation methods and their corresponding scales. At the end of the day I was not sure if any of your GCM runs actually incorporated ang geo-engineering methods.. Or are you just grinding through the data to obtain delta Rn versus Delta other weather variables. The paper repeatedly uses short-term climate feedback and long-term climate feedback without describing neither. Is this like you implement a solar manipulation and look at the surface variables over and short and long durations later.

Why would you want to separate the effects of radiation change on ecohydrology from everything else anyway, what is the type of GE this may be suitable for. If you only play with radiation, then that would be like a greenhouse experiment where one can manipulate radiation and keep temperature the way they wish, but there is no reason to test something that is not realistic. Besides, models are generally linear to Radiation, so the response can be predictable by some excel calculations.

Material and Methods lines 80 – 90:

Four options (O) were outlined for different geoengineering solution scales for manipulating radiation, but without what type of GE that might be, so that makes the alternatives hard to picture, below are my interpretations:

O1. Just change radiation, all else being constant, suitable for local scale: I cannot picture what type of geoengineering solution this might, after all these GE solutions target large areas.

O2. “..include short-term climate feedback, in which solar radiation changes lead to a modification of other climate variables..” Isn’t this a fancy way of saying we can use a reginal climate model to downscale GCM outputs.. Then the question is how you would implement GE within your RCM. But I don’t this this was done at all.

O3. “ … all long-term climate impacts introduced by initial modification of solar radiation…” This can be practically done by modifying your local (let’s say atmospheric chemistry) and feeding your RCM into the GCM.. but how was this implemented?

O4. ”.. the fourth scenario is one in which solar radiation effects are isolated from global temperature changes by perturbing two variables, usually radiation is reduced, and CO2 is increased to preserve the global scale mean temperature..” I am completely lost about the relevance of this.. what sort of geoengineering is this, what scale is this done, at the global scale we are already doing this experiment by putting more CO2 in the atmosphere. Can you give more insights on this?

The study ends up using O2 and O4 and looks at the impacts of this local hydrology, but also climate sensitivities of O3. I am not sure how this climate sensitivities of O3 was looked at, is the solar modification done globally, which model was this in the list?

To put this in the language I understand, seems like O2 is essentially using GCM output in some fashion so that we can identify the ecosystem response to Delta_Solar.

Lines 95-100: are those hydromet-sensitivities calculated from GCMs from the locations of those 115 sites? It would be very tricky to downscales from GCMs to those locations, how was this done? I presume those different CO2 levels were also included in the simulations, or maybe not, it depends what these Geoengineering scenarios are. Without more guidance on geoengineering solutions, it is difficult to judge the study.

Lines 125: I see that the sensitivity is actually calculated using the annual values. Is there evidence that these relationships obtained from annual values represent the seasonal changes.  shouldn’t this be done perhaps using a moving-average that cover the seasonal trends between solar-precip-temperature etc., especially considering the fact that you have an hourly model.  Or if not, how do you justify? I presume your ecohydrology model response would be most sensitive to the “growing season” however it is naturally different across the 115 locations.. Again here, I would like to go back and think about how the initial downscaling was made, that also would have major implications on the local weather variables, perhaps not so much the solar.

Lines 192-195: what is the basis and evidence for this, your analysis is global and I am not seeing any citations to support for this claim.

Figure 1. As expected annual trends seems fairly muted.

Line 225. How did you compute the climate sensitivities at those 115 locations, again wouldn’t this require some sort of GCM downscaling?