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

Water Yield Model: 2.x much faster than 3.0

Mike_AMike_A Member
edited August 2013 in Terrestrial Models
I was wondering if there has been any progress made with the standalone 3.0 hydrological models? I just ran the Water Yield model (latest and greatest 2.5.5) for all of Mozambique with a 90m DEM. I discovered that the ArcMap toolbox version performed the task much faster: 2.x = 5 min vs 3.0 = 30 min

Any idea what might be causing these performance issues? Do I need to worry about the two versions giving me different results?

Thanks,

Mike


I'm not sure how to share my data logs within the forum. I tried copy/paste but it exceeded the character limit.

Comments

  • kglowinskikglowinski Member, Moderator
    Regarding sharing the logs, could you upload them as files? There should be something under the comment box that says "Attach a file."
  • Ok, I think I have them uploaded now. I didn't see that button before. Let me know if you need anything else.

    Thanks again.
  • RichRich Administrator, NatCap Staff
    Hi Mike, sorry to hear it's slow. Generally the standalone models are orders of magnitude faster than the ArcGIS ones so I'm surprised to hear this behavior. I'll ask the engineer that works on this model what he thinks.
  • DougDoug Administrator, NatCap Staff
    Thanks Mike, we are currently looking into this issue and have hopefully narrowed down where it is occurring. We are hoping that by the next release this will be resolved.

    As far as the difference in results between the two models, they both should give roughly similar outputs. However, there are differences in the actual files that are output. The 3.0 model now allows for optional Sub-Watershed files and will only do sub-watershed analyses for the Water Yield Model (not Scarcity or Valuation). We have also done away with a lot of the raster based outputs and gone to shapefile outputs that hold all the results.

    Thanks for your post and helping us improve our models!
  • DougDoug Administrator, NatCap Staff
    Hi Mike, we have been looking into the performance issue you mentioned. It seems that there is a difference in run time performance between our 32 bit and 64 bit versions when it comes to this model. I'll save you the details, but it seems like the 32 bit version runs fine and I would recommend using this version. The 64 bit version will at the moment run considerably slower. Thank you for your patience and let us know if anything else comes up. In the future we will attempt to address the 64 bit issue.
Sign In or Register to comment.