How fast do landscapes change? A workflow to analyze temporal changes in human-dominated landscapes

ContextAnthropogenic activities alter natural habitats, with impacts on species that live in human-modified systems. Often abrupt, anthropogenic influences not only alter the availability and distribution of suitable habitats for species, but also the ability of species to perceive variations within the landscape. Researchers studying the drivers of species distribution and behavior often use "static" land-cover maps as descriptors of habitat, which are most typically characterized at predictably cyclical seasonal scales. Changes that occur over shorter temporal scales are rarely quantified, and there is a lack of understanding of how landscapes change within seasons.ObjectivesWe propose a generic work-flow to identify the temporal scales at which changes in land-cover patterns can be detected within a landscape.MethodsWe use easily calculated landscape metrics such as patch area, inter-patch distance (ENN) and shape complexity (SHAPE), obtained using high-resolution satellite imagery. We conducted pairwise comparisons for each metric and LULC class separately, at temporal scales corresponding to 15, 30, 45 and 60-day intervals, using a case study from central India.ResultsWe observed that changes in landscape structure and in land-cover classes can be detected even at a 15-day time period in human-dominated landscapes. In our case-study, agricultural fallows showed the highest proportion of change-points. The grassland class was the most stable across metrics and time-scales. Among metrics, SHAPE was the most stable and ENN was the most dynamic, indicating that while patch structure remained relatively stable, patch configuration changed more rapidly.ConclusionsWe suggest that when studying animal resource use and movement, particularly in anthropogenically modified systems, matching the temporal resolution of landscape-level data to animal movement data is critical, as broad-scale data may miss key triggers of animal response.