Bayesian areal disaggregation regression to predict wildlife distribution and relative density with low-resolution...

Area of Science:

• Ecology
• Conservation Biology
• Spatial Statistics

Background:

• Accurate spatial population data are crucial for conservation and managing human-wildlife conflict.
• Low-resolution areal counts are common but limit detailed species distribution modeling.
• Existing methods struggle to provide fine-scale wildlife distribution insights.

Purpose of the Study:

• To introduce and demonstrate Bayesian areal disaggregation regression for ecological applications.
• To convert low-resolution wildlife count data into high-resolution species distribution models.
• To assess deer population dynamics and range shifts in Ireland.

Main Methods:

• Applied Bayesian areal disaggregation regression to hunting bag return data.
• Utilized high-resolution environmental raster data for disaggregation.
• Validated model predictions against independent deer surveys and alternative distribution models.

Main Results:

• Successfully disaggregated regional deer counts to pixel-level distribution data.
• Documented significant increases in relative population density for red, sika, and fallow deer.
• Observed extensive range expansion for all three deer species across Ireland.
• Validated disaggregated model accuracy through high correlations with independent data.

Conclusions:

• Bayesian areal disaggregation regression accurately captures fine-scale animal distribution patterns.
• This method enables the use of previously disregarded regional count data for wildlife monitoring.
• The approach offers new possibilities for adaptive management and conservation biology in a changing world.