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

SDR watershed values

Hello,

I have been trying to run the model to generate values for the watershed area and keep on encountering a problem where values are not written to all the different subwatersheds, instead they are just written to three. The values are for sediment retention and export as well as total USLE. Attached is the log and the shapefile generated. 

Thank you for your time,
Thomas

Comments

  • swolnyswolny Member, NatCap Staff
    Hi Thomas -

    Thanks for sending that information. What do the sed_export and usle output rasters look like? Do they have appropriate values, or is something wrong at the pixel level? And what do the output streams (stream.tif) look like? Is the model generating streams in the watersheds without values in the shapefile? Sometimes, if streams are not generated (usually because the area is very flat) we'll see results like that.

    Also, since you're using InVEST 3.4.4, I suggest trying the new version 3.6.0 if possible. Updates have been made to the stream-generating in SDR that might help, if streams are not being generated. (But, if the topography is difficult, or the DEM is bad, there's not much that the model can do...)

    ~ Stacie
  • Hello Stacie,

    Thanks for asking. I believe it is a pixel level error, the streams look fine. Under one of the rasters it says -inf. Linked are the results I am getting. 

    Thank you for your time,
    Thomas Price

  • swolnyswolny Member, NatCap Staff
    Can you send me your model inputs? It's not obvious to me what's going wrong from looking at the outputs. Thanks.

    Did you get a chance to try InVEST 3.6.0 to see if you still have this problem?

    ~ Stacie
  • Hi Stacie,

    I downloaded and ran the new version of the model 3.6. I had the same results occur. Linked are the inputs and outputs from the model. 

    Thank you for your time,
    Thomas Price

  • swolnyswolny Member, NatCap Staff
    Hey Thomas -

    Using InVEST 3.6, I tried something that got me past the -inf in the output rasters. Your input raster cell sizes are not round numbers, but have a lot of decimal places, like 32.43758368. I resampled them to round numbers (like 32, or 30 for the DEM) and the -inf went away. And, now I get USLE values aggregated for all watersheds. So that's something to do - go back to your original datasources, I assume that you had to reproject them, so when you do, give them a cell size like 30x30. 

    It's also worth noting that your erosivity raster doesn't cover the edges of your study area, so you'll be missing output data there. For coarse data like erosivity, we recommend clipping the raster to an area larger than your study area so you have complete coverage.

    But... I still don't get aggregated output for sed_export or sed_retention. You do indeed have an ok stream network, and the sed_export.tif and sed_retention.tif rasters look ok (if unusually concentrated in particular patches) so I don't know what's going wrong with the aggregation.  

    @Rich, can you take a look at this? I'm stumped. Here are Thomas' resampled inputs. Thanks!

    ~ Stacie

  • Hi Stacie,

    Thanks for the recommendation on the input erosivity raster and to resample the rest of the inputs to round numbers. I will try messing around with the model more and see if I come across some work around to get the results divided by watershed. 

    Thank you for your time,
    Thomas Price
Sign In or Register to comment.