Dynamic occupancy models for analyzing species' range dynamics across large geographic scales

Area of Science:

• Ecology
• Macroecology
• Biodiversity Science

Background:

• Large-scale biodiversity data are crucial for predicting species' responses to global change and macroecological studies.
• Analyzing biodiversity data is challenging due to spatial sampling heterogeneity, observation processes, and unexplained spatial effects.
• Understanding dynamics at large spatial scales requires advanced analytical methods.

Purpose of the Study:

• To develop and apply dynamic occupancy models for analyzing large-scale atlas data.
• To estimate local colonization and persistence probabilities alongside occupancy.
• To account for spatial autocorrelation and heterogeneous sampling in biodiversity data analysis.

Main Methods:

• Developed dynamic occupancy models incorporating conditional autoregressive and autologistic models for spatial autocorrelation.
• Applied models to Southern African Bird Atlas Project (SABAP1 and SABAP2) detection/nondetection data for the hadeda ibis.
• Utilized a flexible hierarchical approach for analyzing grid-based atlas data.

Main Results:

• The models accurately reproduced the hadeda ibis range expansion into drier regions of South Africa.
• Colonization of new areas was significantly influenced by the density of occupied neighboring grid cells.
• Detection probability varied significantly across space due to factors like sampling effort, observer, and seasonality.

Conclusions:

• Dynamic occupancy models offer a flexible and robust approach for analyzing grid-based biodiversity atlas data.
• The models successfully account for complex factors like heterogeneous sampling and spatial correlation.
• This approach enables the examination of dynamic changes in species ranges at large spatial scales.