Modelling Best Practices Frequently Asked Questions

These FAQs will tackle the thorny issues that plague model builders and draw on the combined expertise of the working group

Image by Iñaki del Olmo

Goals

These ‘modelling best practices’ will be used by the community when building models to improve outcomes for SAR:

  • Improved efficiency, quality
  • Appropriate use of products
  • Better decision making

Modelling Best Practices: Frequently Asked Questions

How much species data is enough to build a robust model?

How do we determine appropriate spatial extents and cell size (model resolution)?

What sources (conventional and unconventional) of data can we look to in Atlantic Canada?

How do we harmonize datasets across jurisdictions?

When is it appropriate to use Topographic Position Index (TPI) as a habitat covariate? How should the data be processed for use in models (e.g., raw data or categories)?

How do we calibrate the predicted probability of occurrence/abundance to suitable and/or critical habitat?

How can we verify or validate the model we’ve built?

What process should we use to obtain the most parsimonious set of predictor variables?

How to efficiently offer support/resources to train students who might start studying SDM approaches but need a lot of guidance to get going?

How many seasons of field data should you collect for model building or validation?