How measures of agricultural landscape patterns are affected by crop rotation dynamics

Abstract

Landscape patterns have been measured as a fundamental part of landscape ecology, especially with increasing computational power and availability of landcover data. Among the challenges associated with landscape measurement is identifying the appropriate scale of application, both temporal and spatial. Measurement of landscape patterns in highly-dynamic landscapes is especially challenged by temporal scale selection. This short communication focuses on how landscape pattern measurements vary in response to normal variability in a highly-dynamic landscape For a small agricultural landscape with known crop rotation dynamism sequences, a simulation model provided 30 years of crop rotations. These 30 different landscapes were measured with landscape pattern measurements and the variability of the measurements was assessed. Almost half of the measurement values were outliers, and 12 of the 30 years had at least one outlier landscape pattern measurement value, even though the basis of landscape change was a normative agricultural practice. For a consistent but dynamic landscape, some landscape measurements or simulation years had high variability, demonstrating that inferences from highly-dynamic landscapes need to be based on fine-resolution time steps. The application of landscape pattern measurements continues to be a common method in landscape ecology, and temporal dimensions are key to detecting change and what might be causing or responding to that change. The temporal resolution of change implies that highly-dynamic landscapes require more-frequent sampling, and that comparisons for agricultural landscapes over long time periods requires more than a single temporal endpoint.