Multiscale resistant kernel surfaces derived from inferred gene flow: An application with vernal pool breeding...

Area of Science:

• Landscape genetics
• Spatial ecology
• Conservation biology

Background:

• Species-environment relationship models benefit from multi-scale spatial data assessment.
• Landscape genetics studies have not previously optimized resistance surfaces across multiple spatial scales.

Purpose of the Study:

• To model multiscale landscape resistance surfaces for estimating gene flow resistance in two salamander species.
• To evaluate the performance of multiscale resistance surface models compared to single-scale models.

Main Methods:

• Developed multiscale/layer landscape resistance surfaces for spotted (Ambystoma maculatum) and marbled (A. opacum) salamanders.
• Optimized resistance surfaces using forest land cover, normalized vegetation index, traffic rate, and topographic curvature at various Gaussian kernel bandwidths.
• Fit species-specific resistant kernels and scored vernal pools based on local habitat and landscape connectivity.

Main Results:

• Multiscale resistance surface models significantly outperformed models using original spatial scales.
• Optimized surfaces identified specific predictor variables and bandwidths for each species (e.g., forest/NDVI for A. maculatum; traffic/curvature for A. opacum).
• Vernal pool connectivity was effectively modeled by optimized resistance surfaces, with higher-scoring pools found in forested areas with high density and low resistance.

Conclusions:

• A novel multiscale analytical approach successfully optimizes landscape resistance surfaces for modeling gene flow.
• This method provides valuable insights for amphibian conservation, particularly for vernal pool species.
• Findings underscore the importance of scale-dependent predictors in landscape genetics and conservation planning.