Mapping landscape connectivity for large spatial extents
Mapping landscape connectivity across large spatial extents is an important component of ecological reserve network designs and species recovery plans. It can, however, be limited by computational power. One way to overcome this problem is to split the study area into smaller tiles, map landscape connectivity within each of those tiles, and then merge tiles back together to form composite connectivity maps.
We tested the effects of landscape structure on the accuracy of composite landscape connectivity maps created from tiles and tested two methods to increase this accuracy.
We correlated replicate, composite current density maps with untiled maps. We tested whether our findings depended on the composition of the landscape by testing maps with corridors, barriers, different mixtures of high- and low-cost habitat, and road networks.
We found that composite current density maps underestimated large-scale connectivity and overestimated the contribution of small habitat patches to overall connectivity. These biases became more pronounced as the tiles became relatively smaller. Landscapes with corridors or barriers were particularly sensitive. We increased the accuracy of tiled maps by increasing pixel size or by averaging several maps created using a “moving window” approach.
There is a trade-off between tile size and pixel size when modelling connectivity across large spatial extents. We suggest using the largest tile size possible when tiling is necessary, in conjunction with increased pixel size and a moving window method to increase accuracy of the composite current density maps.
KeywordsCircuitscape Connectivity Moving window Tile size Tiling Resolution
We thank the Wildlife Research and Monitoring Section of the Ontario Ministry of Natural Resources and Forestry for financial and logistic support and R Marrotte for advice on making our script more efficient. We are grateful to the anonymous reviewers whose thoughtful comments have improved our manuscript.
- Anantharaman R, Hall KR, Shah VB, Edelman A (2019) Circuitscape in Julia: high performance connectivity modelling to support conservation decisions. http://arxiv.org/abs/1906.03542
- Anderson MG, Clark M, Olivero Sheldon A (2012) Resilient sites for terrestrial conservation in the northeast and mid-Atlantic region. The Nature Conservancy, Boston, p 168Google Scholar
- Beyer HL (2012) Geospatial modelling environment. Version 0.7.2.0. http://www.spatialecology.com/gme
- Bivand R, Keitt T, Rowlingson B (2017) rgdal: bindings for the geospatial data abstraction library. R package version 1.2-8. https://CRAN.R-project.org/package=rgdal
- Bivand R, Rundel C (2017) rgeos: interface to geometry engine - Open Source (GEOS). R package version 0.3-23. https://CRAN.R-project.org/package=rgeos
- Crooks KR, Sanjayan M (eds) (2006) Connectivity conservation (Vol. 14). Cambridge University Press, CambridgeGoogle Scholar
- Doyle PG, Snell JL (1984) Random walks and electric networks. Mathematical Association of America, WashingtonGoogle Scholar
- Environment Canada (2012) Recovery strategy for the woodland caribou (Rangifer tarandus caribou), boreal population, in Canada. Species at Risk Act Recovery Strategy Series. Environment Canada, Ottawa, p 138Google Scholar
- Hijmans RJ (2016) raster: Geographic data analysis and modeling. R package version 2.5-8. https://CRAN.R-project.org/package=raster
- IUCN (2018) Connectivity conservation. https://www.iucn.org/theme/protected-areas/wcpa/what-wedo/connectivity-conservation
- Leutner B, Horning H, Schwalb-Willmann J (2018) RStoolbox: tools for remote sensing data analysis. R package version 0.2.3. https://CRAN.R-project.org/package=RStoolbox
- McRae BH, Shah V, Mohapatra T (2013) Circuitscape 4.0 user guide. The Nature Conservancy. http://www.circuitscape.org
- McRae BH, Shah VB, Edelman A (2016) Circuitscape: modeling landscape connectivity to promote conservation and human health. The Nature Conservancy, 14Google Scholar
- MNRF (1998) Introduction to the Ontario Land Cover Data Base: First edition. Land classification data [computer file]. Ontario Ministry of Natural Resources and Forestry, PeterboroughGoogle Scholar
- MNRF (2007) Southern Ontario Land Resource Information System (SOLRIS). Land classification data [computer file]. Ontario Ministry of Natural Resources and Forestry, PeterboroughGoogle Scholar
- MNRF (2010) Ontario Road Network. Land classification data [computer file]. Land Information Ontario Database. Ontario Ministry of Natural Resources and Forestry, PeterboroughGoogle Scholar
- MNRF (2011) Ontario Hydro Network. Land classification data [computer file]. Land Information Ontario Database. Ontario Ministry of Natural Resources and Forestry, PeterboroughGoogle Scholar
- MNRF (2013) Built-up area. Land classification data [computer file]. Land Information Ontario Database. Ontario Ministry of Natural Resources and Forestry, PeterboroughGoogle Scholar
- Nova Scotia Lynx Recovery Team (2006) Provincial recovery plan for the Canada lynx (Lynx canadensis). Nova Scotia Department of Natural Resources, Nova Scotia, p 32Google Scholar
- Pebesma EJ, Bivand R (2005) Classes and methods for spatial data in R. R News 5:9–13Google Scholar
- R Core Team (2018) R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, AustriaGoogle Scholar