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

ERROR in Nutrient Delivery Model

wenwen Member
edited October 2016 in Terrestrial Models
Hi everyone,
I am running the NDR model.
I can successfully run the model without any error message,
but when I open the output "n_export" and the "p_export" raster,
it shows negative values and no data areas. This cannot be correct. The shp file also gives "n_export_tot" and "p_export_tot" negative values.

Can you help me with the problem?

The link is the input data and the output files:

Thank you !

Wen

1394 x 246 - 49K
856 x 575 - 75K
Post edited by wen on

Comments

  • RichRich Administrator, NatCap Staff
    Hi Wen, it looks like your primary issue might be hydrological pits in your DEM.  You can get a hint that something like this is happening if you look at the flow accumulation raster and stream raster in the intermediate outputs.  

    The problem that's occurring here is the value for the effective retention is based distances to stream pixels, but in many parts of your DEM there are microwatersheds that are hydrologically isolated.  That's not to say they're actually there, it's just an error in the DEM.  Take a look at your "effective_retention" raster in intermediate outputs to see this.

    Anyway, in turn that makes a highly negative NDR_0 which drives the whole solution negative.

    Can you try filling the hydrological pits in your DEM with ArcHydro or SAGA and re-running with the new DEM?
  • wenwen Member
    Hi Rich,
    Thank you so much!
    After I filled the DEM and increased the threshold flow accumulation values, the flow accumulation raster and stream raster seems very well.
    Though the stream raster is good, the output "x_export" and the "NDR_x" raster still shows highly negative values.
    How can I solve this problem and keep going on?

    This is the output data :

    thank you!

    Wen

    1436 x 873 - 94K
    997 x 581 - 119K
  • RichRich Administrator, NatCap Staff
    Hi Wen, it still looks like there are hydrologically isolated watersheds in your DEM.  I've highlighted some of them here by coloring "d_dn" (distance to stream) over the stream network in cyan.  Those bluish values are indicating very high distance to stream values which really means there's no stream to route to.  You can also see this if you zoom in on the flow accumulation raster in this area.

    Also note by increasing the stream threshold, you shorten the pixels that are classified as streams.  Normally that's not a problem if the DEM drains, but if it doesn't, that means there will be swaths of pixels that have high d_dn values like you have above.

    So the right solution here would be to continue to fix your DEM, but a hacky solution might be to lower your stream threshold until "d_dn" resolves to a nice layer.  (Perrine can chime in about how that's a terrible idea) :)
  • wenwen Member
    edited October 2016
    Hi Rich,
    Thank you for your comment,
    I have tried a lot these days
    It seems that all the sinks in the DEM were already removed, but the result still looks alike.
    I just dont know why some pixels still shows very high distance to stream values.
     How can I continue to fix the DEM after all the pits had already been removed?
    Is there have any solution to get the reasonable value of the output data?
    Thank you for helping me for the problem.

    Wen

    Post edited by wen on
  • RichRich Administrator, NatCap Staff
    Hi Wen, I'm still getting the same error on my end too.  It's getting complicated enough that I don't have an easy answer.  I'll work more on this and hope to get back soon.
  • wenwen Member
    edited October 2016
    Hi Rich,
    Thank you for your attention to this matter.
    Please let me know if there is any possible way to solve it.(maybe I can directly input the stream data to skip the dem part?)
    Thanks again! 
    Post edited by wen on
  • RichRich Administrator, NatCap Staff
    Hi @wen, easy fix!  Your biophysical table has effective retention values in them that are greater than 1.  For example, I saw an 5 and an 80.  Maybe you meant for those to be .05 and .8?  At any rate, when I lowered those below 1 I was able to get nutrient export values that looked normal.

    Sorry for the LONG debug time on this!

  • wenwen Member
    Hi Rich,
    I am really grateful for your time and effort,
    I finally got reasonable nutrient export values.
    I am sorry to keep bothering you, you have been a lifesaver! :D
    Thank you so much!

    Wen
Sign In or Register to comment.