Assessing the comparability of landscape connectivity maps: An experimental approach

Background/Question/Methods

Maximizing landscape connectivity is a key objective of conservation planning, but the many approaches used to describe landscape connectivity suggest that generalities may be obscured by methodological differences among studies.  We compiled spatial data from six connectivity projects for 18 species that overlapped in a 132,122 km²portion of the Pacific Northwest, and used those data as the focus of our investigation.  Using the resistance data layer generated for each species by the original researchers, we ran landscape connectivity analyses using two algorithms, one based on circuit theory and the other based on least-cost path modeling. 

Results/Conclusions

Comparisons among all possible pairs of species revealed that correlations between resistance layers were generally weak (mean r = 0.164) indicating that resistance maps were highly variable among species.  Pair-wise comparisons of current maps also indicated low similarity among species (mean r = 0.287).  Mean overlap in 1 km-wide least-cost paths was 13%.  Similarity in current maps between pairs of species was highly associated with pair-wise similarity in the resistance maps on which analyses were based (R² = 0.55, slope = 0.979).  In contrast, the relationship between resistance and overlap in buffered least cost paths was weaker (R² = 0.23, slope = 0.164).  Our results suggest that least cost paths modeled for many different species across several studies tended to be spatially coincident despite substantial differences among input resistance layers.  Pinpointing conservation efforts in a relatively small portion of the landscape may help protect low resistance corridors used by many different species.