Capturing the Impact of the 2018 European Drought and Heat across Different Vegetation Types Using OCO-2 Solar-Induced Fluorescence

Round 1

Reviewer 1 Report

The SIF is highly sensitive to the GPP of vegetation and is worth to deep study. This study explored the spatiotemporal variation of SIF and its response to extreme drought of 2018. There are some issues the authors should pay attention.

1. The SPEI data is too coarse compare to other datasets. Moreover, different timescales could get different SPEI value. The SPEI is a very important base dataset for this study, and would highly influence the results. Therefore, I suggest the authors to build their own SPEI dataset based on OCO-2 meteorological data to decrease the uncertainty of drought area about different vegetations.
2. How do you define drought area? Because, the SPEI varies during the whole year, and the drought area should be varying along the SPEI.
3. There are several spatial resolutions you used in this study. What is the final resolution you used to do the data analysis, and how to combinate these datasets?
4. L211. Why do you choose a 5-day running average? Did you test other lags of average?
5. Why don’t you exclude days during heatwave or drought to obtain the normal SIF during 2015 - 2017 to compare with those anomaly values in 2018
6. The uncertainty analysis should be added in the Discussion.
7. Figure 1. The method to get fig1(c) should be descripted in caption.
8. Figure 3,4 shows the temporal trade and response of SIF, VPD, etc to drought. However, no spatial analysis descripts in the manuscript. May be the authors could add analysis the spatial distribution of SIF under the severe drought condition.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

This paper analyzes OCO-2 and other data during the 2018 European heatwave and drought as compared with previous 3 years mean. The main conclusion is that OCO-2 has a quicker response and possible higher sensitivity to drought compared to MODIS fPAR and NDVI and that therefore there is added value of SIF for studying impact of drought on vegetation. The paper fits the scope of Remote Sensing. However, there are several issues that should be addressed before acceptable for publication. The paper seems rushed to a conclusion trying to get flashy result without detailed analysis of available data sets. Major revisions recommended before publication as there are unsubstantiated claims made. These might or might not hold up if further investigation is made but it remains to be seen.

I do not understand why the fPAR and NDVI data are given in appendix when they are a main part of the conclusions. Data are described in section 2.2.2, so why not show in main part of manuscript. The data should be shown together for easier comparison since this is a main conclusion. Yet more confusing is some NDVI shown in appendix and some in supplement. Of all data sets, NDVI is the least noisy. fPAR looks very noisy and not clear why it would look so different from NDVI. These data sets do not appear sampled as OCO-2. Description of MODIS should state collection 6 not version 6. MODIS is 8 day composite for fPAR and 16 day for NDVI which does not match up with OCO-2. OCO-2 data have more points per unit time and are noisier. The 2 peaks for broadleaved forest do not show in NDVI. So NDVI data have been more smoothed out. Therefore comparison appears not apples to apples as stated in lines 216-217. This is important and may invalidate the claim made on line 352.

Agricultural area shows clear cut drought effect at day 150 for both SIF and NDVI. Broadleaved and mixed forest does not show a clear effect for any data set except perhaps at end of time series (both SIF and fPAR are noisy and even with uncertainties shown not a clear effect). Only coniferous forest shows effect at about day 170 in drought area for SIF while NDVI and fPAR show later effects. Statement in abstract seems misleading as only true for coniferous. Is it not interesting that SIF shows similar response as NDVI in agricultural areas?

A comparison between GSIF and OCO-2 is shown in appendix but GSIF never described. It is downscaled 8 day GOME-2 SIF product. As paper objective seems to be to show SIF has faster response, why not use this dataset as well as it would have more complete coverage and better comparison with MODIS. In that case, also why not use TROPOMI for 2018 compared with 2 year baseline of 2019 and 2020 that are available. TROPOMI has more complete coverage and could show more spatial detail as could GSIF. 

Authors say SIF values affected by VZA, yet no difference is shown in mean SIF in table A1. More explanation needed. Are the angle differences not large enough, but how to reconcile with references? Table A1 does not show this or I miss something. Besides that showing that they have the same mean does not necessarily mean it is OK (observations taken over different days with repeat cycle of 16 days so why would we expect to be same since conditions change in 16 days).  Only one reference given on angle issue but there are many more.

Line 161, it says OCO-2 at 1:30 better captures sensitivity of fluorescence yield to water stress which is higher in afternoon then gives reference. This sounds like reference has shown this. Reference is to theoretical study of GEO sensor and does not show this, so sentence is misleading. To my knowledge no publication showed that SIF afternoon observations have more sensitivity to fluorescence yield.

Discussion on lines 368-369 is non-sensical. Longer periods of data are available with GSIF data set used in Appendix. If point is to highlight what OCO-2 can add, then detailed comparison should be made between these two data sets. If OCO-2 shows better performance than other available data sets that can be quantified, this would be important result. What means SENTINEL in line 370? Sentinel is series of ESA satellites of which several can report NDVI.

What exactly means efficiently capture in line 429. Other references show spatial drought response of SIF and for different vegetation types. Results here similar to previous studies of SIF on drought and support previous studies. This should be mentioned. What about GSIF? If nothing published on GSIF for drought study, then it should be studied and compared.

Line 447 says OCO-2 SIF sensitive to both physiological and structural changes. Without more exact comparison with vegetation index this cannot be claimed, only suggested.

Typos:

Line 101: should be OCO-2

Tables not labeled correctly in appendix

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 3 Report

Please, see attached file.

Comments for author File: Comments.pdf

Author Response

Please see the attachment

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Glad to see the authors has made great improvements according to the reviewers’ comments. I hope the authors could improve the resolution of SPEI datasets, which would highly increase the value of your research. Any away, congratulations!