Data and code for "Spillovers and legacies of land management on temperate woodland biodiversity"

Datasets and R code to accompany the publication "Spillovers and legacies of land management on temperate woodland biodiversity".

Abstract: 

Species distributions are a product of both current spatial configuration of habitats and legacies of historic land use. Here, we explore current and historic drivers of species distributions, considering combined effects of spatial spillovers and temporal legacies, both within and between habitat types. We fit Bayesian hierarchical occupancy models to data on 373 species from four taxa (ground beetles, birds, vascular plants, small terrestrial mammals) from a chronosequence of 134 woodlands (10 to >250 years old) in temperate agricultural landscapes in the United Kingdom. Both spillovers and legacies affect species richness and community composition and, critically, these effects interact. Real-world combinations of spillovers and legacies result in different biodiversity responses compared to the individual factors in isolation. Woodland patches in landscapes with more old woodland and lower amounts of historical woodland loss tend to host more bird and plant, but fewer beetle species. Failing to account for these drivers, particularly legacy effects, gives a distorted view of habitat suitability. In consequence, the same management actions may result in unexpectedly different outcomes depending on the spatial and historical context within the landscape. A better understanding of spillovers and legacy effects on species distributions is required to design biodiversity-friendly, cost-effective land management.

Dataset is in .xlsx format with 8 separate sheets:

'Predictors' = predictors used in the study. Four site-scale predictors: Age (in years, derived from maps), Area (in hectares), Shape (ratio of site perimeter divided by a circle with the same area), Structure (standard deviation of tree DBH collected in field surveys). Six landscape-scale predictors (i.e., of the surrounding 3-km radius landscape): Proportion currently covered by woodland, Proportion of the current woodland that is at least 100 years old, Proportion that was previously woodland but is now another habitat type, Proportion that is covered by 'Trees outside woodlands', i.e. hedgerows and isolated trees, Proportion of agriculture that is arable, Livestock density (as Livestock Units)

'Coordinates' = site coordinates.

'Beetles', 'Birds', 'Mammals', 'Plants' = Presence / absence data for individual taxa.

'Unused predictors' = site-level data for road density and urban area, and proportion of arable agriculture and density of livestock. These were not used in the study (see paper for explanation).

'Landscape composition' = proportion of woodland and agriculture in a series of radii around each study site.