2 min read

Moving forward in fall

2019/10/22 
Warning in .load.config(override.config): Your configuration file is
missing the following entries: tables_type. Defaults will be used.
Warning in .load.config(override.config): Your configuration contains the
following unused entries: data_tables. These will be ignored.
Warning in .check.version(config): Your configuration is compatible with version 0.8.2 of the ProjectTemplate package.
  Please run ProjectTemplate::migrate.project() to migrate to the installed version 0.9.0.

So, after a hiatus on summer (for no particular reason, I just took a break after the Davis talk and never got it picked up), I’m returning to this project. Where are we?

  • The results for Ler are essentially complete. The results as I presented them at Davis broadly make sense. The one thing I need to deal with is to truncate the populations (recall that I had occasional replicates that spread ridiculously far). However, I subsequently noticed in Williams & Levine (2018; p. 878) that only 8 pots were installed beyond the furthest seed. So I can implement that in the code, and rerun the simulations. Hopefully that won’t change the results much!
  • For RILs, it looks as though I still have some work to do to complete the dispersal parameterization. I’m also not sure if the model will run; I recall lots of things breaking right at the end. Indeed, I’m not even sure where the code that ran the model is; the functions are in lib/model.r It looks like the code is in src/make_Ler_sims.R and src/make_RIL_sims.R. The latter look very hacked up; when I try to run it I get an error that RIL_params is not found, so I’ll have to dig into ProjectTemplate to see why the munge script isn’t getting run.
  • There is a great deal left to write in both the parameter estimation document and the TRACE document.

So here are some to-dos:

  • Ler
    1. Find a place in the code to add the 8-pot truncation
    2. Re-run the simulations and analyze the results
    3. Complete the parameter estimation description for Ler (a big job!)
  • RIL
    1. Check to see where things are with the kernel estimates
    2. Write up the RIL parameterization (this will help get me back to an understanding of what we’ve done)
    3. Try to get the existing simulation code running (this looks like it’s just for the evolving populations)
    4. Make simulation code for non-evolving populations
    5. Generate some results for mean and var of spread
  • General principles
    1. Look back at some of the other patterns in the data and think about trying to reproduce them
    2. Review the TRACE elements and decide whether to proceed with this documentation.