NDR model: watershed definition

HI all,
I have to define the watershed for a spring in a karst context. Should I also consider the subterranean component?
kind regards
Fabio

Comments

  • PerrinePerrine Moderator, NatCap Staff
    Hi Fabio, 

    The answer is probably yes, although I'm not sure I fully understood the question! 
    If you are interested in changes in water quality at this spring, you need to know where the water comes from. This can be challenging in karstic landscapes.

    Importantly, think about whether InVEST is the best model for your question (the representation of subsurface flow is extremely simple, and relies on some empirical coefficient that you need to derive from the literature).

    Cheers

  • Hi Perrine

    thank you for your answer. From literature I know where the water comes from. I tried to run the model but it doesn't work, the log says that datasets don't intersect.

    Most of the water in my catchment infiltrates, so I wonder whether conceptually I could use InVEST anyway.

    Cheers

    Fabio


  • PerrinePerrine Moderator, NatCap Staff
    Hi Fabio, 

    The log error may indicate an issue with the input projection: make sure all spatial files have the same projection.
    Like I said, the InVEST nutrient model uses a very simple representation of nutrient removal (using empirical retention efficiency parameters, for both surface and subsurface flows). This may be enough to illustrate a point conceptually, but has limitations if you want to use absolute values of model outputs for your analyses.

    Cheers

  • Hi Perrine,
    yes, I get errors about the projections being different. I tried reprojecting with the project tool in ArcGIS but still doesn't work. Could you suggest any other tool?
    Thank you for the disclaimer, I'll remember it when evaluating results.
    Many thanks
    Fabio
  • Actually I managed to solve that and to run the model!
    Only thing is that I get meaningless values (between  -3.40282 e+038 and +3.40282 e+038...
    Regards
    Fabio
  • RichRich Administrator, NatCap Staff
    Hi Fabio, are you inspecting the results in ArcGIS?  Sometimes Arc doesn't naturally rescale the visualization of the rasters.  Can you manually set the style to clamp to max/min, or click around on some pixels and set some reasonable values?
  • Hi Rich, I don't know how to do that, but here are the outputs attached if you can take a look.
    Cheers
    Fabio
  • swolnyswolny Member, NatCap Staff
    If you're working in Arc, you can do that by right-clicking on p_export.tif, choosing Symbology, then a Stretch type of something like Minimum-Maximum.

    However, when I did that, I still got bad values. So I think something is still going wrong with something in your input.

    ~ Stacie
  • Yes, probably I have to look at my data. Also, this is a karst area, so I should wait for the groundwater module...
    Thanks
    F.
  • RichRich Administrator, NatCap Staff
    Hi Fabio, it looks like your outputs are all nodata.  Can you dropbox me your inputs to richsharp@stanford.edu and I can take a look to see what's going on.
  • Hi Rich,
    thanks for that. I uploaded the files, I hope I didn't make a mess.
    Let me know if any of them doesn't work.
    Cheers
    Fabio
  • RichRich Administrator, NatCap Staff
    Hi Fabio, it looks like your biophysical table is defined for landcover codes in the hundreds (like 211, 412, and 312).  But your landcover map only has codes 1, 2, 3, and 4.  It's likely this is the source of your missing output.  Can you either reclass your landcover map or update your biophysical table codes?
  • Hi Rich,
    thanks for that, that should be the problem.
    regards
    F.
  • I have now corrected the biophysical table. Since it's a karst area, should I put a value in "proportion_subsurface_n"?
    Cheers
    F.
  • I now tried with the corrected table and finally I get meaningful results!
    See pictures attached. Any comments welcome.
    Regards
    Fabio
Sign In or Register to comment.