The relationship between least-cost and resistance distance

Area of Science:

• Ecology and evolutionary biology
• Spatial modeling
• Landscape genetics

Background:

• Least-cost modeling and circuit theory are widely used to model animal movement and gene flow across landscapes.
• Least-cost distance and resistance distance are the primary outputs of these respective methods.
• The comparative bias and relationship between these two modeling approaches remain underexplored.

Purpose of the Study:

• To investigate the linear relationship between least-cost distance and resistance distance.
• To assess the sensitivity of both metrics to variations in Euclidean distance, spatial autocorrelation, landscape resolution, and data aggregation.
• To understand how methodological choices influence ecological inference.

Main Methods:

• Experimental design to compare least-cost and resistance distance metrics.
• Systematic variation of landscape parameters including Euclidean distance, spatial autocorrelation, pixel number, and data aggregation.
• Statistical analysis to determine linearity and sensitivity of the metrics.

Main Results:

• Least-cost distance and resistance distance were not linearly related without data transformation.
• Resistance distance showed lower sensitivity to the number of landscape pixels and Euclidean distance between nodes.
• Resistance distance was more sensitive to data aggregation compared to least-cost distance.
• Spatial autocorrelation did not significantly affect either metric or their relationship.

Conclusions:

• The choice between least-cost and resistance distance modeling can significantly influence study outcomes.
• Data manipulation, particularly aggregation, introduces differential bias between the two methods.
• Careful consideration of the chosen metric and data processing is crucial for accurate ecological inference of movement and gene flow.