ISIMIP3b: CWatM total actual water withdrawal (all sectors) not matching the sum of sectors. #50

We've received the following report about water withdrawal components in the data submitted by the CWatM modelling group. Please take of this issue when analyzing the data:

Dear ISIMIP developers,

While exploring the actual water withdrawal components in CWatM ISIMIP3b, I've made the following observations:
1- The only available sectoral withdrawal components are 'actual domestic water withdrawal' (adomww), 'actual livestock water withdrawal' (aliveww), and 'actual industrial water withdrawal' (aindww). Using the summation of these sectoral variables and comparing it with Total actual water withdrawal (all sectors) (atotww), the results do not match at all. Computing the min, max, and mean ratio percentages [atotww/(adomww+aliveww+aindww)] resulted in the following, 148%, 913%, and 275% (avg.), respectively.

2- Reading through the CWatM model manual and publication, it explicitly states that irrigation is one of the computed sectoral withdrawal components which are not currently available within ISIMIP3b output variables, this could be the reason? Since in the article, it stated that irrigation is mainly abstracted from renewable groundwater resources, I've tried a second comparison. In the current comparison, I've used the 'Total Actual Water Withdrawal (all sectors) from groundwater resources' as an addition to the sectoral components that builds the 'atotww' --> [atotww/(adomww+aliveww+aindww+atotwwgw)] and resulting in 129% (min), 548% (max), and 212% (avg.). Although it is highly unlikely that atotwwgw should be treated as a sectoral component but I've made the assumption for the sake of exploration that it composes the irrigation part.

For reference, I've used multiple outputs that correspond to the same climate forcing input datasets and reference period and the discrepancy is consistent. As an example, the following 5 datasets:

'cwatm_gfdl-esm4_w5e5_historical_histsoc_default_adomww_global_monthly_1850_2014.nc'

'cwatm_gfdl-esm4_w5e5_historical_histsoc_default_aindww_global_monthly_1850_2014.nc'

'cwatm_gfdl-esm4_w5e5_historical_histsoc_default_aliveww_global_monthly_1850_2014.nc'

'cwatm_gfdl-esm4_w5e5_historical_histsoc_default_atotww_global_monthly_1850_2014.nc'

'cwatm_gfdl-esm4_w5e5_historical_histsoc_default_atotwwgw_global_monthly_1850_2014.nc'

Any explanation for the made observation?

Best regards,
Ahmad Awad

It turned out that irrigation withdrawal is missing in the formula to match the model's characteristics for the total water withdrawal. We hope to receive both, irrigation withdrawal and irrigation consumption data in the near future.

Thank you for the feedback. Since you have mentioned the consumption part, then I would like to highlight that consumption variables for domestic, industrial and livestock are not on the ISIMIP3b portal for CWatM (only withdrawals)