Dannenberg et al. develop a neural network to predict GPP, ET, and NEE at FLUXNET sites using satellite retrievals of several environmental drivers from several different observing frequencies (optical/thermal from MODIS and microwave from SMAP). DrylANNd is able to predict the GPP and ET seasonal cycle, spatial variability, and, to a lesser degree, their interannual variability. The predictions of NEE are weaker due to satellites not being able to observe respiration. Overall, I find this to be a nice advance and hope this lays the foundation for follow-up studies. The study is very thorough and well-motivated. I support its publication with consideration of points below. With future applications in mind, I encourage the authors to consider several points below as well as some methodological clarifications. Nice work!

-Andrew Feldman

Overall/Major Comments

1) What are the desired use cases of DrylANNd? It is a named model, which indicates a future application as the authors briefly mention for a global study in line 117. Machine learning approaches like this require careful calibration and validation, which the authors have done well here. However, if the conditions change to a different region or globally, what needs to change about the inputs as the predictors and predicted variables? Can we rely on the few dryland locations in the Western US to predict other regions when there may be different rainfall seasonality and vegetation types (i.e. African and Australian drylands) or do we need to train the model in each different defined region? Are we restricted to certain datasets to serve as the GPP and ET independent variables?

 I recommend laying a framework for applications in the discussion by providing more concrete recommendations on how to apply DrylANNd and points about pitfalls that may come about applying DrylANNd at larger spatial scales or other, related to the questions here. I know Section 4.3 may have been an attempt to do this in trying to improve the model overall with new datasets, but I think the authors can expand on that section with regard to these questions and maybe put a more positive outlook on it. Specifically, I recommend being clearer about how DrylANNd can be applied. Next, maybe give a big picture roadmap such as discussing how we may not have as reliable of observation-only data from satellites as we have from FLUXNET to use to train the model on ET or GPP. Therefore, we are constrained to using the model regionally where FLUXNET is available. Perhaps a SIF product (or other) can be used as a predicted variable elsewhere (like Australia) where there are not widespread, publicly available flux tower observations.

2) I want to caution that there may be a drawback in using soil temperature from SMAP L4 as predictor here, especially with regard to the desire to use remote sensing observations to train DrylANNd. The SMAP L4 retrievals are outputs from a land surface model assimilation (see the Reichle et al. 2019 study referenced in the submitted manuscript). While the soil moisture output is highly a function of remote sensing from SMAP’s brightness temperatures (especially the 0-5cm product), the soil temperature is likely not as highly influenced by the SMAP observations. Historically, we input soil temperature data from a GMAO model in the process of retrieving L3 SMAP soil moisture – we don’t go the other way around to estimate soil temperature. Microwave brightness temperature is a function of physical soil temperature, but more strongly associated with moisture on the surface, and is thus (at least not to my knowledge) not necessarily influencing the soil temperature outputs as heavily in the assimilation. Perhaps the L4 soil temperature output is less of an “effective” remote sensing parameter. I don’t think we have good evidence otherwise, though I would be happy for this claim to be refuted which may require a closer look through the literature on assimilating L-band brightness temperature into land surface models. As a consequence, I think SMAP could be overestimated in its ability to explain GPP, NEE, and ET in Figure 8. Since the study’s goal is to explain these variables with different observation frequencies from remote sensing instruments, I am not sure the SMAP soil temperature is as appropriate here as the other variables and recommend the MODIS LST or raw infrared data instead.

3) Given that the Western US has seen some unprecedented climatic behavior in the past two decades and especially in the past two years, does this create an issue training DrylANNd on stronger dry response anomalies over 2015-2021? It certainly will be a limitation in applications of predicting future ET and GPP (with regard to my point #1 above).

4) I think some mention of how spatial scale mismatch between datasets has an influence on results is important. For example, the flux towers have a fetch of <1000m. However, some of the remote sensing products have much larger native resolutions here, which could lead to problems with spatial mismatch of data allowing spatial heterogeneity errors to creep in to the prediction performance estimates. This may be motivation to demonstrate the method entirely with flux tower data at FLUXNET sites and see if similar results occur. I leave that up to the authors to try.

Line-specific comments

-L30: Wonderfully written introduction

-L65-90: What about effects of biases from soil color contrast and thus soil contamination on the visible signal?

-L135: Is the gap filling necessary where the NN approach cannot be used on irregularly sampled data? Such gap filling methods could bias a predictive approach if a functional form is used to gap fill (for example, a look up table that may be based on model assumptions). A noisy insertion could eliminate issues of model assumptions becoming imprinted in the prediction model. Maybe gap filling is not very common in the available time series? What percentage of the different time series are gap filled? 

-L173: Note that SMAP products do not retrieve soil temperature, though there are some nuances about the assimilation process in L4. See the major point above.

-L225: what is the “holdout model?”

-L228: Are only 7 data points being used in the interannual timescale prediction? It seems 7 data points from all sites are normalized by taking out their mean and aggregated with other sites to increase sample size (as in Fig. 7). Please clarify in the text.

-L233: Are these months averaged in all cases for the warm season or is the max used in the case of visible/NDVI like it was for individual months (as stated in line 185)?

-L234: A word of caution that SMAP went into safe mode in summer 2019 which led to 1-2 months of loss of data. This is unfortunate because leaving this year in the analysis could bias predictions with biased means. Taking 2019 out removes samples from an already short time series. I encourage the authors to assess the consequences of removing 2019. 

-L246: By the end of the methods, I have not gotten a picture of precisely what the inputs and outputs are. Are the predictors always from remote sensing and the predicted, independent variables are always from FLUXNET? Table 1 does help, but it may help further to add to table 1 that FLUXNET ET, GPP, and NEE are the independent/predicted variables.  

-L293: It might be helpful to mention that time and space are mixed on the left panels of Fig 7 where the spatial patterns might be dominating the good performance there. Only temporal patterns are shown on the right panels.

-Fig 8: Can the authors indicate in table 1 or elsewhere which variables are grouped into VI only, LST, and SMAP as corresponding to Fig 8?

Review for “Upscaling dryland carbon and water fluxes with artificial neural networks of optical, thermal, and microwave satellite remote sensing” by Dannenberg et al.

Dannenberg et al. present an approach for estimating dryland GPP, NEE, and ET by training an artificial neural network (ANN) with remote sensing signals (optical vegetation indices, thermal observations, and microwave soil moisture/temperature). The study is novel, scientifically sound, well written and within the scope of Biogeosciences. I would recommend this paper for publication but have a few revisions I think should be addressed, mainly around paper presentation and clarification on methodology.

Minor Concerns:

• The structure of the introduction and methods have some overlapping material. For example, the fourth paragraph of the introduction in lines 63-72 mentions that plant physiological responses are not necessarily reflected in optical signals, but this paragraph doesn’t make the connection between optical VI’s that are sensitive to greenness specifically. Discussion of ‘greeness’-based metrics failing comes later in the methods section in lines 141-155 but I think it would be useful to draw the connection earlier in the introduction. In addition, the same paragraph in lines 63-72 says “microwave, thermal, and visible wavelengths can capture complementary information about plant and ecosystem stress that is unattainable from optical VIs alone”. An explanation as to WHY these indices are useful is available in the methods but could be moved further to the introduction.
• The final paragraph of the introduction could be rephrased to make the hypothesis/study aim clearer. Specifically, the first sentence states, “Here, we develop and test an approach for data-driven prediction of a full suite of carbon and water fluxes that are specially adapted for drylands using…” but I think this can be much stronger to highlight the value of the study. Something along the lines of, “We aim to improve the prediction of GPP, NEE, and ET based on remotely sensed metrics by using…”
• Somewhere in the methods should include the number of test/train data points used.
• The final paragraph of the methods discusses the authors approach for testing the importance of predictor variables. Has this approach been used in other studies? Some validation of this approach or references for more information would be useful.
• The color palette of figures could be adjusted to follow more a ‘intuitive’ color scheme e.g. dark green for ENF – this is not critical but might help with figure readability.

Line edits:

Line 37: intensity of water limitation feels like awkward phrasing

Line 53: It might make more sense to move this like to the end of the last paragraph so someone scanning the paper could easily find “First, Second, Third” in the three paragraphs talking about the unique nature of drylands.

Line 54: It might be nice to define mesic

Line 59: “the effects of soil moisture stress…” but it’s the effects of ALL soil moisture right?

Lines 53-60: I found this paragraph a little difficult to follow as several sentences are quite long. I think it would be worth revisiting for clarity.

Line 67: Satellite-based estimates of fPAR should still be fine, it’s just that the plants aren’t responding to the increase in light by being more photosynthetically active. I would rephrase this.

Line 88: ‘however’ is unnecessary

Line 90: can be more specific with ‘uniqueness’

Line 91: ‘other places and other types of ecosystems’ seems redundant

Line 94: ‘for example’ is unnecessary

Line 97-100: I would rephrase to put the emphasis on the finding of the study, not the author, and just present the citation at the end.

Lines 113-117: References to sections might be useful

Line 117: ‘global-scale estimates’ – of ecosystem fluxes?

Line 185: ‘compositing’ is confusing and maybe incorrect?

Line 192: this statement deserves a citation

Line 194: ‘… predictions of multiple variables.’ Deserves a citation

Line 210: here could be a good place to include the number of test/train data points

Line 328: ‘Interestingly’ is unnecessary

Line 333: ‘However’ is unnecessary

Line 340: ‘modeling’ feels like the wrong term to use here – I think predicting or estimating would be more accurate since modeling implies process based (to me).

Line 403: ‘thermal data’ – it might be better to say LST here?

Figure 2: I think it would be useful to say what the input variables are in the figure (not just the outputs)

Figure 3: the + indicator is a bit difficult to see/compare with the bars – it might be easier to see in black or a different shape.

Figures 5, 6: I think it would be useful to indicate on the figures somewhere which sites fall under which land cover classification category

Figure 7: It’s unclear to me what the lines in a and c are

Figure 8: Do the lines connecting the scatter points represent anything? If not I would remove