I am estimating sediment generation in a basin in Mexico. I am running two scenarios one with some conservation infrastructure (reforestation patches, gabion dams, living fences, etc) and one without it to estimate how much these activities are helping prevent sediment generation and erosion. I find that when I run the scenarios with the conservation activities the USLE, sediment export, and sediment deposition are higher while sediment retention is lower. This goes completely against the logic behind this work, is it possible that im doing something wrong?
Thanks for sharing your inputs. I just requested access to that GDrive folder.
SDR results depend on how different land cover classes are parameterized in the biophysical table, namely the values for cover-management (C) and support practice factors (P). Lower values will lead to less soil loss (USLE) and sediment export, while higher values will cause the model to predict greater USLE and export. To attain the results you may expect, the conservation infrastructure features you mentioned must be represented in the LULC input with corresponding C and P values reflected in the biophysical table. See this section of the User Guide for more information.
An important note about dams and the SDR model, is that once sediment reaches a stream or river, the model assumes it ends up at the catchment outlet, thus no in-stream processes are modeled. To deal with this, we recommend delineating watersheds/subwatersheds by inputting dams points as outlet features. See the Appendix for how to do so. The final paragraph of this section also discusses how to handle dams.
If you continue to need assistance, please let us know here and be sure to grant me access to the GDrive folder you shared. Please also be sure that the log file (.txt) is provided.
Thanks for your answer. As far as I understand, higher cover-management (C) values will lead to lower soil loss and sediment exports. On the other hand, support practice factors (P) with higher values will lead to greater USLE and export. Is this correct? That is what my LULC classes currently represent.
No, you have C backwards. Higher C values cause greater soil loss. Lower values will lead to less loss. See the usle equation in the User Guide. This sounds like the cause of your issue.
Unfortunately, I cannot interpret the data layers you shared because they are no metadata and no LULC raster, but it seems no longer necessary. Adjust your C values so that land types like forest have low values (near 0; retain sediment well) and types such as barren land have high values (near 1; easily eroded).