Pollination model edge effect

[sibylles]

Dear community

Could it be possible, that we underestimate the pollination potential at the border of an InVEST map? e.g. the border of a country, of a sample of squares?

Here a nice visualisation from Stacie Wolny

My reflection:
We do not know what is outsite the squares. As far as I understand the pollination model there are some squares at the border (of a country or sample of squares) that will have pollinator supply from just only 2 or 3 neighbouring squares or landscapes. flight distance are sometimes > 1000 m, so we may miss some pollinator supply at the border.
Especially for smaller countries edge effect may be relevant.
Furthermore I will work with 180 1km squares. In this case there are no direct neightbouring squares, but sometimes for some bees with large flight distances there may be a supply from another square. How does InVEST processes “no data” in the neighbourhood if actually a supply of bees is needed?

@swolny
@jdouglass
@elonsdorf

Kind regards
Sibylle

[elonsdorf]

Hi Sibylle, InVEST ignores data outside these borders. Whenever we use it for our work, we make sure to buffer the extent of our analysis by 2x the averaging foraging value of the pollinator with the largest forage range to make sure we capture the external sources of pollination, precisely because there could be edge effects or spillover. Here is a paper that might be relevant for dealing with edge effects and spillover where we used the pollinator model. https://besjournals.onlinelibrary.wiley.com/doi/full/10.1002/pan3.10138

Hope this helps!
Eric