ecoConnect: Regional Ecosystem-based Connectivity

Landscape connectivity is vital to the long-term survival of plant and animal populations. At varying scales, connectivity is necessary for feeding, migration, mating, genetic diversity, and dispersal to new areas. As climate change alters habitat, species will be forced to alter their ranges, dispersing great distances over generations. Where connectivity is broken by unsuitable habitat, development, and roads, populations may face extinction. ecoConnect, our model of regional connectivity, informs conservationists reconnecting the landscape in the Northeast.

We start the analysis by selecting points with high ecological integrity in a 2 km grid. These points are connected to other points within 5 km based on development, road traffic, and ecological similarity to each starting point. These connections are made with random low cost paths, mapping a large number of suboptimal routes, yielding paths that represent local connectivity. We count the number of times each path is used in connecting each pair of points across the entire Northeast, and use this regional analysis to weight local connectivity, resulting in models of regional connectivity that are:

Truly regional: rather than chaining local or medium-scale connectivity, this approach assesses paths hundreds of miles long, from Virginia to Maine.

Multi-scale: connectivity is assessed at multiple scales, so results are meaningful for a town, a state, or the entire Northeast.

Ecosystem-aware: rather than simply connecting natural systems, connectivity is estimated for forests or wetlands or even finer groups such as ridgetop forests or floodplain forests, and paths are based on ecological similarity to the target systems.

Independent of pre-defined conservation targets: rather than connecting existing or aspirational conservation cores, all of the landscape is connected, so conservation targets may be brought in later in the planning process or omitted entirely.

ecoConnect can be used to assess the level of connectivity among existing conservation land, and to target additional land to conserve connectivity, including “unlovely” land important for connectivity that might not otherwise be protected. It can also be used to assess viable locations for wildlife road passages, and it highlights bridges that already provide connectivity under high-traffic roads.

Development of regional ecosystem-based connectivity was funded by the USGS Northeast Climate Adaptation Science Center and USDA’s Renewable Resources Extension Act (RREA).

---

Documentation

A description of ecoConnect: https://landeco.umass.edu/web/lcc/dsl/ecoconnect/dsl_documentation_ecoConnect.pdf

Technical report: [Not yet available]

---

Web viewer

You can view ecoConnect for the Northeast using our simple web viewer: https://umassdsl.shinyapps.io/ecoConnect

We have received funding from USGS Northeast Climate Adaptation Science Center to develop a more complete web viewer for ecoConnect in the Northeast, including the Index of Ecological Integrity and a site scoring tool. We expect to post the initial version in Spring 2024, replacing our current stopgap version.

Download GIS data

GeoTIFFs of regional connectivity for the northeastern United States are available for four ecosystems:

1. Forests and forested wetlands (573 MB): https://landeco.umass.edu/web/lcc/dsl/ecoconnect/ecoconnect_forests.zip
2. Nonforested wetlands (420 MB): https://landeco.umass.edu/web/lcc/dsl/ecoconnect/ecoconnect_wetlands.zip
3. Ridgetop systems (224 MB): https://landeco.umass.edu/web/lcc/dsl/ecoconnect/ecoconnect_ridgetops.zip
4. Large river floodplain forests (66 MB): https://landeco.umass.edu/web/lcc/dsl/ecoconnect/ecoconnect_floodplains.zip

Served GIS data

Results are also available as a Web Map Service on our GeoServer:
https://umassdsl.webgis1.com/geoserver/wms

1. Forests and forested wetlands ecoConnect:Forest_fowet
2. Nonforested wetlands: ecoConnect:Nonfo_wet
3. Ridgetop systems: ecoConnect:Ridgetop
4. Large river floodplain forests: ecoConnect:LR_floodplain_forest