bpbond
commented
3 years ago Shoshi and Hope are staying for the summer and will be working on SRDB.
Right now, in the SRDB folder, there are lots of studies in the "Sophia_mckever" folder that I'm sorting through—lots of not very relevant studies. The 'good' (relevant) ones I move to the "New (unopened)" folder, and the not relevant ones to got "Rejected". Feel free to help, or yes if you'd like to QC Hope and Shoshi's data, that's fine too! I will Slack you all.
jinshijian
Global C revision
I am not familiar with the datasets of partitioning ratios, but these measurements are notoriously scarce and obtain a very high variable importance in the bootstrapping. I therefore would have expected a more comprehensive discussion of their representativeness and potential influence on the results. For example, to what extent are climatic gradients within vegetation types represented in this database and should they be?
Another argument presented in the case of Rs:GPP ratio explicitly given in the SRDB database or estimated from combination of SRDB with FLUXNET is to explain that there is probably no bias related to spatial extend of data since weighting (or not) the measurements by vegetation areas globally gives similar results. But it just proves that values from underrepresented vegetation type is not very different for those that are overestimated. But if sites on vegetation from which there is only few measurements are not well representative of the whole ecosystem we can still have large biases. We can also notice that if GPP is estimated from the in-situ Rs:GPP ratio of 0.54 and the new global Rs estimate of 93 GtC we find a global GPP value of 172 GtC which is even larger that estimation given from GPP_Rslit!.
One important point is that, in opposite to Rs estimation which is only based on direct measurement or estimation for flux tower, GPP can be estimated by different independent methods. This is unfortunately only briefly discussed in the paper and only shown in figure S5. This is however an important point. In particular looking to table S2 we can see that a large majority of estimates used in the study come from MODIS which gives all relatively similar estimations in the lower range of the estimated GPP. Then the mean GPPlit of 113 PgC given as reference calculated by averaging all the GPP estimates from literature is mainly dominated by MODIS estimations and then relatively low. For instance, if instead of averaging all the values we first average the values coming for each method (I.e NDVI, satellite driven models, FLUXCOM, FLUXNET, SIF, O18, IPCC, global carbon) and then averaging the mean value of each method to give a similar weight to each method independently of the number of available estimation the mean value is 130 GtC which is still lower than estimated from Rs but closer however. More generally, as briefly mentioned, some methods like O18 of SIF for instance gives estimates of 162 GtC and 147 GtC which are in agreement with GPP estimated from Rslit of 149GtC ! But as I mentioned previously, it doesn’t prove that these values are more realistic than estimation based on remote sensing NDVI which are around 112GtC but at least it shows that the estimations or GPP and Rs are not necessarily fully inconsistent are claimed in the title. Likewise the fact to say that GPP is probably underestimation is not necessarily true for all the method considered. So it would be important to have a deeper discussion on the different methods of estimation of GPP and not just look to the average of these estimations.
Concerning the discussion of Rs:GPP simulated by ESMs there is an important limit which is related to the fact that only two versions of the same ESM model (CESM2) is considered in the study Considering the large spread of modeled respirations in the different ESMs, it is very difficult to give a general conclusion on simulated Rs:GPP ratios in ESMs based only on a single model. Obviously I understand that it is related to the fact that only CESM2 report the partition of respiration within the different biomass compartments in the CMIP6 database. However there is some ways to see how CESM2 behave (in term of ratio between respiration and GPP) compared to others models. For instance, one can see how ratio of total Ra to GPP and Rh to GPP in CESM2 compare to others models. Probably, most of the models provide also the partition of biomass between the different compartments. Even if respiration coefficient can be different for the different compartments, weighting Ra with the relative biomass of each compartment could be a first guess of root respiration as well. So to be able to discuss what is the ratio of Rs to GPP from the models point of view it is very important that CESM2 could be compared to others ESMs to see if it is, for the different ratio of respiration to GPP, in the middle of models range or at one of the end members. In particular from what I understand in the text Ra_root:GPP is 21% and Rs:GPP 27-29% that would means that Rh:GPP is only 6-8 % which seems unrealistic.
Add a panel show RH:GPP ratio (similar as RS:GPP)
I am not familiar with the datasets of partitioning ratios, but these measurements are notoriously scarce and obtain a very high variable importance in the bootstrapping. I therefore would have expected a more comprehensive discussion of their representativeness and potential influence on the results. For example, to what extent are climatic gradients within vegetation types represented in this database and should they be?
With temporal representativeness and GPP group considered:
Table 2 move to SI
SRDB updating