Menu Close

Questions about running LANDFIRE models spatially

Syncrosim Forums ST-Sim & State-and-Transition Simulation Models Questions about running LANDFIRE models spatially

Viewing 1 post (of 1 total)
  • Author
  • #4178

    Here is a recent set of questions I received by e-mail (my answers in bold):

    My goal is pretty simple, and one we have touched on before: how many acres of each disturbance type would a BpS experience annually? Getting this in a non-spatial way is remarkably easy-thanks. I would like to take the next step and use mapped initial conditions and have run into some questions:
    1. I was able to successfully load in my “Vegetation Type” tif file, which is a raster of the BpSs of LANDFIRE Map Zone 61. See top screen capture below. The Raster File Attributes are correct. I have 3 “technical” questions:
    a. According to the ST-Sim spatial inputs video there should be an ID column in the Strata properties, but I do not see one. Is this a problem? See second screen shot. I am guessing it is. I did not see a way to “tell” St-Sim which attribute in my raster links up with the “Name” in St-Sim.
    To see the Stratum ID column under “Project | Definitions | Vegetation Type” right click on the table and select “ID”. The column should then be visible.

    b. Is there a way to see if ST-Sim has the correct number of acres per strata (BpS)? I clicked around and did not find that.
    After a run you can generate a chart of “State Class Area” and “disaggregate by” vegetation type. You will then see the area of each vegetation type charted over time. In your case, these should be constant.

    c. My thought was to keep the LANDFIRE rule of equal proportions of state classes as initial conditions. I assume this is OK.
    That depends on how long you are running your model for and what questions you are asking. If you are running over a long period to equilibrium conditions, this should be ok.

    And- a broader question-is this the best way to get annual acres per disturbance per BpS? St-Sim locked up. I would like to do this for the country and worry about the computing load of such a run.
    Depending on how many cells are in your landscape you may need to split your runs into several jobs. This kind of task is best done on a server, using Amazon or some other cloud computing platform

    I like being able to average the results form multiple Monte Carlos.
    You can plot either acres or proportion of area disturbed. These can show results for individual Monte Carlo simulations or be averaged

Viewing 1 post (of 1 total)
  • You must be logged in to reply to this topic.