This forum is shutting down! Please post new discussions at community.naturalcapitalproject.org

Seasonal Water Yield calculate curve number Error 0

Hello,

I am
working with the Seasonal Water Yield Model which I want to use as nutrient
runoff proxy input for the NDR model.

Every time
I run the model I receive an error I don’t understand.

I was
reading some discussions in the forum but nothing I found solved my problem

Can you
help me with this error or give me a hint?

 

Thanks

Eleia

Comments

  • Hello
    again
    ,

    I solved
    the problem. I used the wrong soil group raster with 0 values instead of
    noData. Now the model passes the calculate curve number part but I get the next
    error. Maybe the holes in my soil map are the reason? 

    I’m sorry for this basic
    questions but maybe someone give me an advice what I’m basically doing wrong?

    I would much appreciate it

    Thanks 

  • RichRich Administrator, NatCap Staff
    I'm not sure what might be going on there. Could you Dropbox your data stack to me to see if I can reproduce and fix? My Dropbox email is richsharp@stanford.edu
  • RichRich Administrator, NatCap Staff
    Hi @Eleia, I just ran your data, and yes your soil map is full of nodata holes. The model naturally masks nodata across the datastack. So it looks like the DEM isn't routing because it's getting chunked out by the nodata holes. But I suppose the real problem is the nodata holes in first place. 

    Any chance you could make a soil map layer that is filled all the way out across the DEM?
  • RichRich Administrator, NatCap Staff
    Hi @Eleia, I tried to run the model with your data this morning, but I don't see the biophysical table in the dropbox you sent me. Could you add that, let me know, and I'll take a look?
  • Hi Rich,
    thank you for the hints. Unfortunately, in my test area in Austria the soil map is available only in this form. Perhaps there is a somewhat costly way to create a map covering the whole area. Maybe I'll try that.
    In my Dropbox the biophysical table is visible. I uploaded it again.

    Thank you very much
  • RichRich Administrator, NatCap Staff
    Oh sorry, I posted that last comment to the wrong thread! Yes I can see your table, but the nodata holes are your issue.
  • Hello Rich

    to test the model I filled the NoData holes with some correct but random values (I had done thisalready  in the past without success). In addition, I cliped all data to a smaller study area. So every dataset covers the whole AOI

    Now the model is running! That is the good news. The bad thing is that I need different or new data now because I don't have soil values for Forest sites.

    Many thanks for the help. I have to go back to the beginning and rethink it.

    Thank you and best regards
    Eleia
  • RichRich Administrator, NatCap Staff
    Hi @Eleia, it looks like the issue was related to how I was handling nodata values in the raster stack. In your case the ET0 rasters did not completely overlap with the DEM thus creating some nodata holes there. This was similar to a bunch of issues other users have been having and I think I've fixed it with this build of InVEST. At the very least I was able to run through your datastack completely on my end. Can you try it out and let me know if that fixes your issue? These fixes will roll into our release of InVEST 3.4.0 so feel free to upgrade then.

  • Hi Rich,
    thank you for the nightly build.
    With this version the model is running. Even with the soil dataset with NoData holes the model runs completely.

    Now a general question. What happens in those places where no soil data is available? How does the model calculate the CN values? Should the model be applied with these large areas with missing soil data.

    Attached the log file from successful the test run on my machine.

    Thank you
  • RichRich Administrator, NatCap Staff
    Hi @eleia,

    Here's the new rule: any biophysical layer pixel that is nodata that also overlaps a valid DEM pixel is treated as "0". That means the CN is 0 everywhere you have nodata. I imagine this isn't want you want at all, so might be a good idea to fill in something there that makes sense in your context.

Sign In or Register to comment.