SDR Model_results

Hello,

i tried to run the SDR model, but im not sure if the outputs maps are correct. There are large areas of NoData in the outputs maps <<avoided_export>> and <<sed_deposition>> that i can’t explained, because the Nodata areas are larger than the areas with streams.

i run the model with flow accumulation 100 and 8000 but there is still large NoDATA areas. i run the model with flow accumulation= 8000, in order to comapre with my streams that i exporded from DEM. The real streams are kind of same with the Stream.tif output maps, but the streams.tif has biger stream width.

I would like to ask you how can i deal with this problem.
For the Borselli K Parameter, MAXIMUM sdr value, Borcelli IC0 parameter and Maximum L value i putted the default value.
here is a file with my data and results file (SDR_WITH FLOW ACC_100 and SDR_WITH FLOW ACC_8000)

https://drive.google.com/drive/folders/1-zJ41ew3O8SFIDQsFGftX11qJO_3aDqb?usp=sharing

Thank you in advance,
best regards

Hello @Tsiatsiari and welcome to the forums!

These large areas of nodata that you’re seeing are not expected, and I am looking into it. Thank you for including your data!

James

Hello,

Thank you very much for your immediate response. if you have any question about the data feel free to ask me.

I am looking forward for your reply.

best regards
Tsiatsiari.

Thanks for your data @Tsiatsiari !

So the issue here is that there is imperfect overlap between some of your layers. Specifically, there are places where your DEM has valid pixels, but your erosivity and erodibility are nodata. This mismatch is causing the issues you’re observing.

I believe this issue has been fixed in the development build linked below, which will mutually mask out all input layers where any pixels in the aligned stack of rasters are nodata. Could you try it out and confirm that it fixes the issue?

https://storage.googleapis.com/natcap-dev-build-artifacts/invest/phargogh/3.13.0.post649%2Bg6743ff2a2/workbench/invest_3.13.0.post649%2Bg6743ff2a2_workbench_win32_x64.exe

Thanks!
James

Hello,

Thank you very much for your help.

I run the Sediment delivery ratio model using the link that you sent me (development build), but the model did not run. You can find the error message linked below.

I also replaced the R- factor raster with better resolution data. There isnt no data area in erosivity and erodibility raster (linked below). but the problem is still there.

I would like to ask you how can i solve this problem.

Thank you in advance.

InVEST_error.txt (4.5 KB)

https://drive.google.com/drive/folders/1-zJ41ew3O8SFIDQsFGftX11qJO_3aDqb?usp=sharing

Whoops, sorry about that @Tsiatsiari and I’m sorry about the delay! The issue you mentioned has been fixed in the newer development build that I am linking below. Could you confirm that this addresses the issue you observed?

https://storage.googleapis.com/releases.naturalcapitalproject.org/invest/3.14.0.post121%2Bg2a02f756a/workbench/invest_3.14.0.post121%2Bg2a02f756a_workbench_win32_x64.exe

Thank you,
James

Hello,

Thank you very much for the correct link and help. i am really appreciated. I run the model using the development build link, but unfortunately there is still large area of no data in avoided_export.tif and sed.deposition.tif .

This time there is also no data area in the other outputs files (sed. export.tif, avoided_erosion.tif urls.tif). but no data area is only accros the river and the high value of the rasters is just inf, not number.

i also run again the model using the official version inVEST 3.14.0 and the high value of rasters is number, not inf.

You mentioned that there is no data area in erosivity and erotibility data, i checked again but there is value everywhere.

when you run the model using the development bulild link, is there large area of no data?

do you think can i do something else to solve this issue?

I am sharing the results from the development build link (run_2) and from offical version of Invest 3.14.0 (run)

https://drive.google.com/drive/folders/1b3LtOmWRl0pJpM7DPV78G1X-mTx6Pgp9?usp=sharing
Thank you in advance.

best regards

Hello @Tsiatsiari , and I’m sorry for the delay! Between projects and a holiday week, I’m still catching up on some things.

So, the outputs that you shared (thank you for those!), which are in Run_2.zip, look as expected. The nodata areas I’m seeing look like streams to me … the Multiple Flow Direction routing algorithm that InVEST is using tends to create wide streams where the landscape is flat. And I’m pleased to see that the first bug you had posted about appears to be fixed, where the only nodata regions we are seeing are on the outside of the raster, or where there are streams. You can also take a look at stream.tif to confirm what the model is detecting as a stream.

Let us know if you have any other questions!
James

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.