Hi, currently I’m trying to conduct sensitivity analysis on NDR model, hoping that I can gain an insight on the extent of each parameter that poses an impact on the result. Nonetheless, after trying multiple time, all of the result is completely empty.
What have you tried so far?
For conducting sensitivity analysis, I created pseudo input data, including DEM, Precipitaion, 5 pseudo scenarios for LULC, Watershed and Biophysical Table. Then, I set up Threshold Flow Accumulation to 10,000 and 2 for Borselli K Parameter. Any advice would be extremely helpful, Thanks!
I took a quick look at your raster inputs, and they have 30 columns and 30 rows, for a total of 900 pixels. If you’re using a Threshold Flow Accumulation value of 10,000, that means that 10,000 pixels must drain into any given pixel for it to be considered part of a stream. But there aren’t even 10,000 pixels in the whole raster. So no streams will be created, which means that the NDR part of the model won’t work properly, since it’s partly based on the downslope retention between each pixel and the stream. If you look at the files stream.tif and what_drains_to_stream.tif in the intermediate Workspace folder, they’ll probably illustrate this.
I’m a little concerned that you’re using such a small area of interest - is this what you’re actually using for your study?
Hi @swolny , first of all thank you for the reply, it helps me a lot!
Regarding the size of the area of interest, the pseudo DEM is solely for sensitivity analysis. My actual study area is the catchment of a reservior located in New Taipei city, TW, called Feitsui Reservoir. Thus, my input raster for the actual study area has 8853columns and 7132 rows.