The landscape of the Litchfield Hills is 75% trees. Connecticut as a whole is about 60% forested, and actually loses more forest cover now than it replaces through natural succession. We still have forested uplands in Northwest Connecticut that are of sufficient size to sustain a broad array of animals that depend on contiguous, intact forest habitats for their survival.
The Litchfield Hills Greenprint has developed a novel way of defining these areas and is using it to help its conservation partners set regional priorities for conserving large forest habitats. We did so because we were not satisfied with existing data sources for this resource of regional significance.
Both The National Audubon Society and the Nature Conservancy recognize exceptionally large blocks of relatively intact forest habitat in Northwest Connecticut. In TNC's case, its Lower New England / Northern Piedmont Ecoregion Conservation Plan (2000) emphasized predominantly forested areas bounded by roads encompassing at least 15,000 acres, which TNC considered the minimum area needed to withstand the impacts of major natural disturbance events. TNC believes conserving these forest blocks can serve as a "coarse filter" for most of the region's terrestrial biodiversity.
It is a very coarse filter, however. There are very few of these places and TNC had to ignore a number of smaller roads and fragmenting features to define the areas. It is hard to explain why some roads were considered fragmenting while others are overlooked. I remember one fragmentation metric that did not consider roads where two people could stand and toss a Frisbee for five minutes without having to step out of the way of an oncoming car. I give them points for creativity, but it is still an imprecise means of determining traffic volume and more than a tad subjective.
Merely buffering a digital road layer and selecting larger patches of interior forest as the Greenprint had initially done also proved inadequate, since Connecticut's Geographic Information Systems (GIS) road layer is a real mess and lumps together private driveways, trails, and roads digitized from old topo maps. This makes it exceedingly difficult to select as "fragmenting" a set of roads of a specific width, surface, and traffic volume using digital data.
We needed to be able to account for the impacts of habitat fragmentation as we defined core forest habitat. We decided not to consider roads when defining boundaries, but instead to use the 2002 remote land cover data: the most current available for Connecticut. These data layers recognize 12 land cover types, 11 of which occur across this landscape. The results are shown on the map above in pale green.
Here is how we did it:
- We selected altered land cover types - grass and turf, agricultural fields, barren land, utility lines, and developed areas (buildings and associated road infrastructure).
- We buffered them by 300 ft. We chose this distance to account for some of the significant edge impacts on core forest habitat, which include the spread of invasive exotic plant species from disturbed areas and invasion by brood parasites that threaten interior forest nesting birds.
- We then selected the remaining terrestrial land cover types (coniferous and deciduous forest, forested and non-forested wetlands) and made a single shape file.
- We clipped the buffered areas from the habitat shape file, and eliminated patches of contiguous habitat that were <190 acres in size. What remained represented our region's predominantly forested core habitat.
- We added the lake cover layer , which we did not want to buffer or to count toward the computation of overall terrestrial habitat. Nothing more resembles egg on one's face than standing up before a commission or land trust and pointing to a purposed forest habitat that is 90% below the surface of a lake.
This is what it looks like close up. Permanently protected open space shows up in dark green, while the habitat is pale green. You can see that small roads that had closed canopies were not detected by the satellite, but houses in clearings and large open fields were identified and buffered. The remote data looks at 100' pixels and classifies them by their dominant land cover type.
These forest habitats cross jurisdictional boundaries and occur across many ownerships. The number of forest landowners in southern New England doubled in the last decade, but the overall area of forest declined. The inescapable conclusion is that forest parcels have been divided and are managed - if managed at all - in lots of ever diminishing size and often without regard for the larger forest system in which they occur.
This kind of resource mapping offers landowners, land managers, land trusts, municipalities and regional planners a new way of understanding the distribution and protected status of our larger forest habitats. If you overlay these data with actual parcels of land, you start to see opportunities to locate development so that it creates less habitat fragmentation, or identify large forest parcels to try and conserve. We believe that we in the Litchfield Hills need to conserve at least 20,000 additional acres of this forest habitat where it expands cores and connects corridors, and do so in the next dozen years.
These maps do not tell us whether there is a willing landowner, or how the forest is managed. That sort of data comes from those most closely connected to the resource and the communities in which it occurs. The maps help focus attention on the resource and structure that discussion.
Comments