Assessing the interaction between landscape structure and biodiversity – method development and application

Severe declines in UK butterfly species over recent years have been attributed to changes in agricultural practices, increased urbanisation, and possible effects of global warming, which have together changed the structure of our landscapes (1). Of concern, the relationship between landscape structure and butterflies is not fully understood. Much of the empirical data reveals contradictory responses between butterfly species to landscape structure, and many studies have only focused on single habitats and often used inappropriate spatial scales (2). Understanding the interaction between landscape structure and butterflies is essential, to halt further population declines.

The aim of my research is to assess whether landscape pattern metrics can be used as surrogate measures of butterfly distribution and diversity, and ultimately be used to predict the impact of future changes in landscape structure, facilitating sustainable land management. Initially I identified landscape pattern metrics that discriminated between different landscape structural components at a range of spatial scales. A moving window analysis, applied to the Land Cover Map 2000 and Warwickshire Phase 1 Habitat map, was used to compute selected landscape metrics at a scale relevant to which butterflies perceive their landscape. Spatial assessments of butterflies have been generated from wildlife survey data obtained for Warwickshire and summarised at a scale of 1km², a standard resolution for wide scale biodiversity monitoring.

Butterfly abundance and distribution data, standardised to account for uncertainty and temporal gaps, has been related to landscape diversity, composition and selected landscape pattern metrics, summarised at the same spatial scale. Logistic regression analysis based on landscape components were applied to identify potential predictor models of butterfly distribution for all species and species grouped by their ecological attributes, as classified by Shreeve et al., (3). These groups were used to capture the behavioural response of species movement to functional connectivity. Developed models will be compared between the two landscape data sources to identify the level of precision required for developing relationships with biodiversity. A key outcome of the model development will be the identification of data requirements and assumptions for the application of these models to other species and landscapes.