Hello,

I am trying to decide if gathering true absence data from various herbarium records is worth the time and effort in terms of the benefit they could provide the model.

I'm creating an SDM over a large area ~ 44,000 km^2. In order to prevent sparseness of background data and to follow the recommendations for the specific modelling techniques I'm using (ensemble of small models with ANN, CTA, and GLM algorithms), I would like to use ~ 10,000 pseudo-absences.

If I intend to use a mixture of true absences with high coordinate certainty and pseudo-absences where pseudo-absences will likely far outnumber the number of true absences that can be generated, will the signal of including true absences likely be significant?

Thanks!