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

RIOS pre-processor runs, but produces empty raster outputs

kfreemankfreeman Member
edited November 2014 in Pre/Post-Processing
When I run the RIOS pre-processor, it will hang at the point it is generating the riparian continuity index and never ends. It generates a stream network and output layers for upslope source, downslope retention, and riparian indexes for the models Erosion, Phosphorus Retention, and Nitrogen Retention. The stream network shape file looks reasonable to me. However, every raster output: the upslope source layers, the downstream retention layers and the riparian indexes that are created are entirely (the full raster extent) NoData. Strangely, when first viewing these output layers they appear to have values, but when examined cell by cell, there is no data. So for instance, a riparian index created will look exactly like the stream network, but the cells that look like the stream data are NoData values, as are the cells around them.

I am using a cell size of 10m, for a large area, 14469 columns by 12317 rows and running the Erosion, Phosphorus, Nitrogen and Flood Mitigation models.

Does anyone have any suggestions for what might be causing this result?

Comments

  • RichRich Administrator, NatCap Staff
    Hi, I pinged a couple of our RIOS scientists about this issue and they're still working on it! Sorry for the delay!
  • swolnyswolny Member, NatCap Staff
    Sorry, I thought I responded last week, but the reply doesn't appear here.

    So the script is hanging doing riparian continuity for Flood Mitigation? Have you tried running the script only with Flood, not with the other objectives checked? I wonder if it's hitting a memory issue or something with that large DEM. Which version of Arc are you using?

    How odd that the results look fine but have NoData values. If you look in the raster Properties/Source, what are the min and max values? If you look at the Symbology, what are the values there? If there's not a Stretch set, what happens if you set one? I'm trying to figure out if there really are NoDatas, or if it's somehow just not showing it to you correctly. If there really are NoDatas everywhere, that would certainly be caused if there's an input that doesn't overlap the rest, or has all NoData values. Otherwise, you could look through the Intermediate folder and see if there are any rasters in there that are also NoData, that could help tell where it's failing.

    If these don't help, perhaps you can make your Flood datasets available and I can try running it.

    ~ Stacie
  • Hello,

    Thanks for your reply. I am using ArcGIS 10.2.2. What I did is clip the area (and every input data layer) to a single HUC10 as FGDB grids and checked only Erosion model, to see if it was size that was causing the problem. Even with the smaller area, I get the same NoData results. However, it does not hang - the preprocessor ends normally.

    For erosion downslope index Min/Max values show in properties as blank, in Symbology they show as the large NoData value. Applying a color to NoData shows it is entirely NoData.

    For erosion riparian index Min/Max values show in properties as blank, in Symbology they show as the large NoData value. It displays a stream network.

    For erosion upslope source, Min/Max values show in properties as 0, in Symbology they show as 0 also. If I apply colors in Symbology to NoData and Background, the study area displays as 0 and the rest is NoData.

    Many of the intermediate files are also blank or 0. These files are starting with: cind, er_cwgt, er_flowlen, er_ind, ret_.

    Does this help point to something?


  • When I ran the flood model alone the flood_slope_index.tif did have values between 0.33 and 1. The upslope, donwslope and riparian outputs were either NoData or 0.

    I have attached the layers I used to run the flood model,

    Thanks,
    Kathy
  • swolnyswolny Member, NatCap Staff
    Thanks, Kathy. I'll try to run these and let you know what I find.

    ~ Stacie
  • swolnyswolny Member, NatCap Staff
    Hi Kathy -

    I'm just trying to run these layers, and it looks like the general class/activities table is missing - can you please send it? I could make one up, but would rather use yours, in case it's what's causing the trouble.

    Another thing I'm wondering: Are the inputs located in the geodatabase when running the preprocessor? I didn't explicitly support .gdbs in the script, and wonder what would happen if you save the inputs as rasters outside of the geodatabase and use those as inputs instead. Give that a try and see what happens.

    Thanks...

    ~ Stacie

  • Hi Stacie,

    I'm sorry I forgot the tables. They are attached. I ran the preprocessor with tif files originally and had the same result. When I clipped the study area to a single HUC, I thought I would try FGDB files to see if that would solve the problem.

    Kathy
  • swolnyswolny Member, NatCap Staff
    It doesn't look like the tables actually got attached. Thanks for the info about running them with .tif files.

    ~ Stacie
  • OK, I attach them again. It looks like the zip file uploaded.

    Kathy
  • swolnyswolny Member, NatCap Staff
    Thanks, Kathy - the activity table is the problem. In this table, the column 'lucode' must have values that correspond to the 'Value' column in your land cover map. But the values in the attached activity table do not correspond to the land cover class values. I changed the lucode values in the activity table and the preprocessing results look good. Give this a try and let me know how it goes.

    ~ Stacie
  • thanks kfreeman i was looking for this thanks
Sign In or Register to comment.