Good evening, I am being questioned about my results generated by Invest, according to my examining board it is not consistent that the value Good evening, I am being questioned about my results generated by Invest, according to my examining board it is not consistent that the value of retained material is higher than USLE, as well as the amount of soil exported.
Could you help me validate this result? Follows spreadsheet.material is higher than USLE, as well as the amount of soil exported.
Could you help me validate this result? Follows spreadsheet.BACIA COM 5000.csv (2.1 KB)
Hi @Mariane -
As the User Guide notes the sed_retent output from the model “represents the avoided soil loss by the current land use compared to bare soil, weighted by the SDR factor. This index underestimates retention since it does not account for the retention from upstream sediment flowing through the given pixel. Therefore, this index should not be interpreted quantitatively.”
So we don’t compare the sed_retent result with USLE or sed_export, we only report sed_retent as an index from low to high, and use this index to value the current landscape’s ability to keep erosion from occuring. Future versions of the SDR model will have the ability to quantify retention from upslope, but the current versions (as of InVEST 3.7) do not have this feature yet.
For USLE and sed_export, these are quantitative values, and I’m wondering if you have done any calibration/validation of these results (usually sed_export) against observed data, so you have more confidence in these absolute values?
~ Stacie
Good afternoon Stacie,
The C-P values used in my research were obtained in the literature. Unfortunately no work was found that has done a modeling in the watershed in question as it was done by me, now that I have more knowledge of the tool I aim to validate the generated data. Our doubt at this point would be that this data would refer to retained soil. Thank you so much for the return, I am open to suggestions for improvement.
Are there any stream gauges in the watershed that record sediment levels in the water? That is the data we usually use for validating the model results. If there is none, then we recommend only using the relative values (low->high) and focusing on landscape patterns when reporting the results.
If you need to accurately quantify the amount of sediment that is retained, then I suggest looking into different models that provide that ability.
~ Stacie
gratitude for the advice 