Developers and friends.
When I used the InVEST model to do the assessment of spatial and temporal changes in water purification (NDR), I found that the results varied very little from one period to another, about 0.2%, and this result is obviously problematic because the changes in land use from one period to another are significant, and there is a lot of forested and cultivated land that has been converted to urban land. At first I thought it was the effect of rainfall, so when I entered the same rainfall raster data and then evaluated it, the change was still very small, I would like to ask you what is the reason for this?
In addition, the validation of the results is a difficult problem. Although my friends said before that we can use the relative values of the assessment results for the study, my supervisor still asked me to do the validation, I would like to ask if there is any good method for the validation of the assessment results.
I hope you can help me with the above two questions. Thank you!
Hi @littlepencil -
When you say that the results varied very little, is this related to the per-pixel results, or the total nutrient change across the whole basin?
If there is a lot of land use change, I would also expect a significant change in modeling results, and would need to look at the data for both scenarios to perhaps understand why. Do make sure that you’ve reviewed how the nutrient loading and efficiencies are different between the different land cover types, since it may be, for example, that there’s much less nutrient loading from urban land than by the agriculture that was there before.
I know of one paper that might help you think about a sensitivity analysis and validation with NDR, which was co-authored by NatCap hydrologist Perrine Hamel:
Redhead et al 2018, National scale evaluation of the InVEST nutrient retention model in the United Kingdom
~ Stacie