George & Nyeema (leading)

From george 2010 Oct 4
  • Have sample running, but we don't quite understand how to interpret some of the output - esp the habitat quality - because there's not enough narrative about the case study to make sense of it.
  • George has most of the data ready to do a trial run using the Upper Neuse Land Cover and create a "fake" set of tables for forest species.
  • Nyeema is going to research ideas for what kinds of guilds we might use (seems like forest and grassland habitat), which species in the region fit in those guilds, and how we might rank the relative threats.
  • George is going to contact Ricketts to see how they created to information for the sample data - where did these numbers come from?

From george 2010 Oct 11 09:30
  • Still unable to get the model to run with my own data. I have found a number of issues, and have sent information to Nasser Olwero at WWF to see if he can resolve.
  • The current land cover is processing correctly, now, but I'm having trouble with the threat layers at the moment.
  • Here are some
    • Model does not seem to run if the workspace is not C:\InVEST\Biodiversity ... so I put all of my input files there
    • In the threats and sensitivity files, and the matching data layers, it seems that mixed case causes problems. I've renamed all of my threat layers to contain lower case only. That helped.
    • Current problem is some kind of map algebra complaint on the threat layers - Nasser is working on that
  • This is all extremely frustrating, having spent hours trying to make things work, but it's kind of the way such things always seem to work. This is exactly why we needed to get started with "made-up" data even as we are working on finding real data.

From george 2010 Oct 11 11:40
  • OK, I moved all of my data to C:\InVEST\Biodiversity\Input and ran it again - it worked!! (I renamed the sample directories to InputX, IntermediateX, and OutputX, then loaded my own input directory from a zip file (copied below).)
  • Here are the data I used for my input files:
  • I ran with a current land cover (the 2001 stuff) and no baseline or future. I also erased the pixel size in the run dialog window.
  • I got valid quality and degradation output files.

From george 2010 Oct 11 15:30
  • Talked with Nirmal Bhagabati at WWF about where the LU/LC sample threats and sensitivity came from. Essentially, they came from expert opinion among people who know about tigers (the sample is about tiger habitat). We'd need to find similar expert opinion about such things in the Triangle.
  • I am developing the strong opinion that the best thing we could do wrt biodiversity is to
    • do a sensitivity analysis on threat and sensitivity in 0.1 increments from 0 to 1 - all combinations
    • possibly add to that sensitivity to the maximum distance (for edge effect)
    • compare results to other efforts at identifying habitat for the Triangle and to NHP data on forest critters (which we have from the earlier work done by Hess etal around simple plans: Hess etal 2006. Comparing potential effectiveness of conservation planning approaches in central North Carolina, USA. Biological Conservation 128(3): 358-368.

From george 2010 Oct 14 14:15