For the input guild table. In my case I use the parameters from nested species distribution values. Relative values representing habitat suitability index (each species 0-1) for the bee species. So similar to the sample guild table (values don’t sum to one). SDM are at larger scale including environmental (bioclimataic and land-use) variables, so I intend to use the second InVEST step with the local habitat filters.
The three different possibilities to prepare an input table, do they have an effect on the output and interpretation of the pollinator supply index and pollinator abundance index (PS and PA)? I think not, because all the pixel values are related to the total (100% = 1 or >1…).
As far as I can see PS and PA also do not sum up to 1 (over all pixels and also not over all species at a pixel). So what is 100% in PS and PA, all pixels together that also not need to sum up to 1? I am still looking for the exact wording of PS and PA in the context of the input guild table. Someting like this so that non-experts and non-modelleres can understand: For species 1, pixel 1 with (PA =0.1) has 10% of all species 1 from the total country and pixel 2 with (PA=0.2) has 20% of all species 1 from the total country.
Maybe it will help to find an answer to my question if I bring sa(s) from equation 5
Crop Pollination (Pollinator Abundance) — InVEST® documentation (storage.googleapis.com)
As discussed before, the relative abundance of the input table does not need to sum up to 1.
My question now: Does InVEST automatically transforms input table that are not summing up to 1 to relatiave values summing up to 1 for sa(s)? I am missing this process in the documentation part. this may clarify the meaning of the following indicators.
The definition of the sum of sa(s) is clearly not per pixel, but the entire species pool (=1). So this must be the values of the input table.
Correct, the PS (pollinator supply) score does not need to sum to 1. The pollinator supply summarizes the combined impacts of the nesting quality at a specific location and the quality of the foraging landscape surrounding the nest. A score of 1 would mean that both the nest and all of the surrounding landscape is ideal or perfect. It is a relative index and should be positively associated with abundance and diversity. The Pollinator abundance index then summarizes the landscape of supply - providing an index related to the abundance and diversity of bees one might see foraging from the nests. The relationship between PA and observed abundance is not linear so a score of 0.1 does not necessarily mean that one should expect 10% of the abundance of diversity - see pasted image from earlier work relating the PA score to observed bees. We related abundance to a log function.
In other work, a linear function has worked. See Cusser et al. 2023 - Public and private economic benefits of adopting conservation tillage for cotton pollination. Agriculture, Ecosystems & Environment 342, 108251.
The input tables of Nesting and Floral resources can have an impact on the score. For example, we performed a sensitivity analysis of the estimates for nesting and floral suitability and found that it depended on the cover types. See table 2 here: Modelling pollination services across agricultural landscapes | Annals of Botany | Oxford Academic
The sum of the species abundances in the input table does not need to sum to 1 and I believe yes, the InVEST tool will transform them so they become proportions. They indicate the relative abundance of the species in the landscape your analyzing. This helps provide an overall pollinator habitat score that is weighted by the relative abundance of the bees you are representing in the assessment. @jdouglass can you confirm this bit about the abundances being transformed?
Yes, I confirm that this is exactly what is happening behind the scenes. Species abundance and also foraging activity transformed in this way. The values in the table do not need to sum to 1, and InVEST will transform them so that they become proportions in the range 0-1.