|I find the author’s refocus to model improvement language is helpful. I think the analysis and results are sound and interesting and makes the paper worth publishing. I still think the paper would be stronger with rephrasing of the introduction and motivation to place the study into the context of the field. I think the current motivation is a bit weak and may be doing a disservice to the methods and analysis (which are very nice). I have a few more recommendations to consider before the work is published. I expanded some of my thoughts a bit more to help the authors find where my confusion lies. I hope the recommendations are helpful.|
1) I am still a bit lost about the purpose of the paper, likely because there are multiple interacting components of the goals. It seems there are (1) very broad goals related to generally improving NASA LIS parametrizations and (2) there are very specific and narrow goals about studying Dry Chaco water balance and effects of deforestation. Both are fine by themselves, but these are currently mixed together in this paper (for example in Lines 85-90) and not easy to follow throughout the paper. This mixing seems to confound each analysis. The general goal (1) is limited by the fact that the Dry Chaco appears to be a not well-behaving location (has large land use change). Therefore, it could be a strange location to make general LSM improvements and assess LSM differences as opposed to a different location that does not have land use change confounding factors. The specific goals (2) initially appear confounded by the general goals because the authors do not mention that the deforestation effect and vegetation parameters would be fully isolated (which they are in the methods) from investigations with SHPs.
There are a lot of changes being made: vegetation parameters, SHPs, changing LSMs, etc. The authors either need to focus on one (an extreme change would be to separate the broad and specific goals into different papers). Or they need to clearly partition each goal in the introduction and what the current problem is in the field with each. The paper ends with a list of conclusions on Line 655 that don’t feel very organized in addressing the multifaceted goals. I recommend acknowledging these differences in the broad and narrow goals and be very clear why the Dry Chaco is the best testbed to evaluate all of these interacting goals.
These ideas should be clearer in lines 85-90 and throughout, including the conclusions. Lines 1-5 in the abstract are close in doing that: the authors start with the big picture to make the overall LIS improvement and narrow into the specific deforestation issues. This is mostly good. But I am still left wondering whether the Dry Chaco is the best location to study the overall water balance and test overall LSM performance if it is undergoing major land use change that could confound tests with those parameters.
2) I still find limited context in the introduction about the state of the knowledge in the model performance. For example, the statement starting on line 29 leaves the reader wondering: how well do LSMs currently represent land surface processes and where are the knowledge gaps and model limitations? It is a nice statement to set the rest of the introduction up. But afterwards, the wording is vague and sounds a bit obvious: better parameterizations will make the model better. There is great discussion in the Discussion section on the state of the field, so it seems the motivation is there. The authors should be more explicit up front about this. I think a new paragraph needs to address, for example: have previous studies showed specific issues with SHP parameters that caused large biases/RMSE? What specifically happened in previous studies when proper vegetation change and phenology was not included? These are just examples, but more direct motivating statements answering such questions would strengthen the motivation of the paper.
Line 9: large regional and long term differences in what?
Line 72-75: Since I am not very clear on the goals, I am a bit lost with this hypothesis. First of all, the hypothesis here seems obvious – that it would have to be true. Aren’t the authors, in part, trying to evaluate only the vegetation to evaluate its effect in representing the deforestation? Then separately evaluate the soil component to see how this additionally improves it? So the hypothesis would be something about the fact that improving vegetation parameterizations greatly improves model improvement in the Dry Chaco over effect of SHP updates.
Line 78: Wouldn’t one want to evaluate only vegetation parameter enhancements to isolate the the enhancement with deforestation? On the other hand, wouldn’t one want to evaluate improvement in soil parameters in a region where no deforestation is occurring? I would make it clear here that the methods here isolate the effects of each. It almost sounds like the study runs the risk of making multiple changes at once and not being able to attribute which changes are truly enhancing the model – which is not the case here.
Line 85: It still isn’t clear to me: is the Dry Chaco the best region to evaluate all changes being made (different LSMs, SHPs, and vegetation parameters)? It seems clear to evaluate the effect of deforestation here. But there are a bunch of other goals that make me thing the Dry Chaco is not the best for. Wouldn’t one want to test the effect of different SHPs and LSMs at a site that does not have a strong land use change signal? Please clearly list all the reasons why the Dry Chaco was chosen for this multifaceted study (probably in Section 2.1) – the deforestation is good for some parts, but seems counterproductive for other parts of the study.
Line 91: It would be helpful to specifically mention what each of these datasets is being used for In each of subsections of Section 2. For example, is the data used for model input or validation?
Lines 313-318: Efficiency curves are a holistic and meaningful way to compare between models. However, the language in these lines makes it unclear what the goals are and what the reader is looking for in Figs 8 and 9. It seems the authors desire in-situ data to see if the updated SHPs represent in-situ behavior better, but it is not available. Line 316 then says that a comparison is being made, but it needs to be clearer what the authors are testing. Is it to find which LSM performance seems “better”? Is it to evaluate the hydrologic behavior itself that the models provide some consensus on and how updated SHPs change this behavior?
Line 481: This 4.3.1 section can seem distracting at first. At this point of the paper, I am not sure why one needs to know whether MERRA2 precipitation compares to in-situ measurements in the context of the study goals. One finds out the purpose in line 603. It should be explained earlier why this analysis is being done and how a P bias can confound the results.
Line 646: Is the study about the Dry Chaco or the LSMs? This statement seems to be focused more on the Dry Chaco hydrology, but the study is mainly focused on LSMs and not Dry Chaco itself (the main findings bullets in the conclusion here all focus on the LSM parametrizations and not Dry Chaco)
Line 654: As I mentioned earlier, the conclusions appear to be a bit unorganized with respect to which of the objectives each is addressing. If the goals are more clearly structured and outlined, it will be easier to follow how each conclusion addressed each goal.