Assessing Landscape Fragmentation Dynamics with Fourier Transforms

Abstract Quantifying landscape fragmentation is crucial for understanding its impact on biodiversity. Within landscape ecology, there is an ongoing effort to create metrics that systematically characterize landscape patterns and predict their effects on various species. In this study, we utilized Fourier transforms to measure landscape fragmentation in urban and heterogeneous landscapes over periods of 24 and 12 years, respectively. Analyzing Landsat images from 1986, 1999, and 2010, we focused on the first principal component (PC1) to evaluate changes in fragmentation over time. Our analysis revealed a significant increase in fragmentation in urban areas from 1986 to 2010, with a particularly noticeable rise between 1999 and 2010. Specifically, the average Fourier spectrum decreased slightly from 16.0 to 15.9 throughout the study period, indicating increased fragmentation at both urban and landscape levels. The sensitivity analysis further showed that removing 6,000 cells from the original image caused the mean Fourier spectrum to drop from 16.0 to 15.5, demonstrating the method’s effectiveness in detecting subtle land cover changes. Our findings demonstrated that Fourier analysis effectively detects subtle landscape changes, particularly in larger study areas and over longer time intervals. However, it lacks detailed information on specific types or quantities of altered land cover. To address this limitation, integrating Fourier analysis with land cover classification datasets could enhance understanding of landscape dynamics and support more informed conservation and management decisions.