I have some questions about the calculation of annual AET in the Seasonal Water Yield module.
I design the scenario 1 to test whether the precipitation at month m would contribute to the monthly AET of month m+1 ~ 12 in this module. In brief, I set all 12 months’ input data as the input data of month m. And then I set the data related to precipitation (raster data of monthly precipitation, table data of number of monthly rain events) of Nan-m months to “0” value. I set α to 1/12.
The output results (Table 1) showed that the value of annual AET was the same whether the value of m was set to 9 or 12, which seems to indicate that the precipitation at month m would NOT contribute to the monthly AET of month m+1 ~ 12 in this module design.
Besides, I added another scenario. In brief, I set all 12 months’ input data as the input data of month m. And then I set the data related to precipitation (raster data of monthly precipitation, table data of number of monthly rain events) of month 1 ~m-1 to “0” value. We set α to 1/12.
The output results (Table 2) showed that as the value of m increased from 9 to 12, the increment of annual AET was very small, which really confused us. In addition, when we set α to Pm-1/P, the increment of annual AET became a little bit bigger (Table 3).
I am really confused of such small increment of annual AET in the second scenario and looking forward to your guidance.
Note: I have attached the sumrised output data (Table 1, Table 2, and Table 3) in TEST_20230329.csv.
Regarding table 1, it is possible that annual AET might not change as a result of precipitation. As you can see from the user’s guide, the equation for monthly AET is the minimum of:
the monthly potential evapotranspiration PET_{i,m} (no precipitation involved)
the available water (which involves precipitation).
My guess is that option 1, above, is what is happening, where the monthly potential evapotranspiration is generally smaller than the available water, but you could confirm that by comparing the equations with your various inputs and the model’s outputs.
Regarding table 2, I still suspect that this may be explained by your PET_i. AET_i is calculated for each individual pixel, so if your PET_i pixels are consistently smaller than the available water (which again, involves precipitation), then the aggregate values you have in your table will not change much even if you have noticeable variation in your precipitation.
All of this is spelled out with much greater precision in the user’s guide chapter (now available in Chinese and Spanish, if you prefer either of those!), so I would absolutely recommend taking a close look there if you have any questions about how these inputs relate to the outputs and how they might change.
Hope this helps, and let us know if you have any further questions!
James