Running crop pollination to large areas

Hello, everyone,
I am planning to run the pollination model for a very large region in Brazil (Atlantic Forest biome; that cover about 1,1 million km2). Given the size of the study area, I cannot assume that the same bee species occur throughout the entire region.
I could not find an approach to account for the distribution of species directly in the model.
One option I’m considering is running the model separately for different regions and then combining the outputs. I could run it separately for each ecoregion of the biome (that might be about 7 or 8 ecoregions). First, I define the species list for each ecoregion, then I run the model separately for each one. Does this seem like a good approach? Suggestions?
Thank you,
Fernando

Hi Fernando -

While I’m not an expert on this model, I do have a couple of ideas. First, if you’re running it over a very large area, depending on the resolution of your input data, it might be computationally challenging to run the whole area at the same time, so dividing it into ecoregions as you suggest could be useful from that perspective.

If you’d rather run the whole area at once (and/or computational resources are not a limiting factor), you could try sub-dividing the land use/land cover map, and the related nesting_[SUBSTRATE] categories to allow for the different species that live in each region. For example, instead of having a single “Forest” land use class, you could use the ecoregion map (or some other sort of delineator) to create sub-classes “Forest - ecoregion 1”, “Forest - ecoregion 2” etc. This could potentially make a lot of land use classes, but then you can assign different substrates (such as “cavity_region1”) to each, which can be used to map to different bee species in the Guild table. This is a bit more convoluted, so I’d be interested to hear if anyone has an idea for a more straightforward method.

~ Stacie