Manipulating rasters between timesteps?

[Taylor Mutch]

Hi Leonardo and Colin,

I was playing with the following idea in my head and I wanted to see if this is appropriate or not within the ST-Sim world:

Using spatial inputs and outputs from Landfire (converted to the appropriate state class values for each veg type), would it be appropriate in the modeling sense to modify the input raster between each timestep (i.e. run ST-Sim for one timestep, alter the resulting output raster based on exterior information such as climate info), possibly change some transition probabilities to simulate increased or decreased probability of a given state transitioning along a given pathway, and then run again on the new raster, repeating this process for N timesteps?

This would have the effect of potentially using other spatial models in tandem with the ST-Sim model (hydrology models, habitat suitability, intactness).

What are your thoughts on this (high level) process?

Cheers,
Taylor

[leonardo-frid]

Hi Taylor,

This is possible using our existing, but not yet documented plugin architecture for ST-Sim and SyncroSim. There are currently two new prototype examples of people doing this. TNC Nevada is using a dynamic link to a sage grouse habitat suitability model written in R which is used to modify ecological restoration probabilities in space each timestep. USGS is linking ST-Sim to the FARSITE fire growth model to define where fires happen on the landscape for each timestep. Both these examples allow for feedback between the state of the landscape at a particular timestep and the dynamics of the external model which in return affects transition probabilities in space. The main interface for this is the use of transition spatial multipliers. Because the functionality is brand new, we are still in the process of developing the documentation and some simple examples on how to link external models. This should all be ready quite soon.

Regards,
Leonardo

[Author]

Posts