Ecological Resources

The data used in the ecological resources component of the Plan were derived from the Natural Lands Trust's (NLT's) "Smart Conservation" model. This model was developed using a process that involved a broad range of scientists and practicing conservationists organized into workgroups, guiding criteria development according to taxonomic groups: 1) plants, 2) mammals, 3) birds, 4) herps (reptiles and amphibians), and 5) aquatics (water quality and aquatic organisms).

The Smart Conservation model was initially developed for an 11-county area of southeastern Pennsylvania. To adapt it for use in this project, the data were clipped and reclassed for the five counties in southeastern Pennsylvania.

Layers

To prioritize land in the region for its ecological resource value, a modeling technique similar to that for agriculture was used. NLT's first version of the Smart Conservation model used 15 data layers to prioritize the ecological resource value of the region's lands. The updated model utilizes 21 data layers, which include the following:

• Vertebrate Habitat subcomponent
  • Mammals
  • Fish
  • Herps
  • Birds
  • Important bird areas (IBAs)
  • Important mammal areas (IMAs)
• Aquatic Resources subcomponent
  • National Wetlands Inventory (NWI)
  • Hydric soils o Floodplains
  • Forested water quality
  • Riparian buffer quality
  • Water quality (2002) 303[d] list
  • Headwaters protection
  • Impervious cover, 2000
  • Impervious cover change, 1985 to 2000
• Terrestrial Resources subcomponent
  • Interior forest habitat
  • Natural vegetation habitat blocks
  • Contiguous grassland habitat blocks
  • Contiguous barrens- or scrub/shrub-type habitat blocks
  • Steep slopes
• County Natural Areas Inventory and PA Natural Heritage Program subcomponent

Detailed Layer Descriptions

Wildlife/Fish Subcomponent

- Mammals, Fish, Birds, Herps:
These data layers originated from the Penn State University (PSU)/Environmental Resources Research Institute (ERRI)/Pennsylvania Gap project. In 2000, PSU/ERRI released their habitat modeling layers to predict where vertebrate species are most likely to be found in Pennsylvania according to land cover, species range, and other habitat determinants, such as elevation, topography, or other such physical, map-able determinants (such as stream corridors for Louisiana water thrush).

NLT took these statewide species layers and, working with the Expert Taxa Advisory Groups, which were convened to inform the SmartConservation™ project, removed any non-native species and species that are not endemic to the Pennsylvania Piedmont ecoregion. Once the species lists had been compiled and finalized, each species was ranked by the Expert Advisory Groups according to conservation value (CV). Conservation value was derived by considering various aspects of a species' role in the ecosystem, such as whether it was a keystone species or whether it was sensitive to disturbance or fragmentation. General population trends were also considered, (to the extent they were known), while rarity (primarily with relation to population trends) was also taken into account. CV values ranged from 0 to 10, with primarily 0, 2, 5, and 10 being used to represent "no," "little," "medium," and "high," respectively. Upon finalization of the CV assignments, NLT added each species, using its CV weight, according to the taxa group to which it belonged. The results were normalized to a 1-10 scoring scale using a 10% quantile classification system.

- Important Bird Areas (IBAs):
IBAs have been defined as core and buffer polygons across areas of SEPA by the Pennsylvania Audubon Society. Where these areas exist, the researchers boosted the value of these cells to supplement the data for bird habitats. Core areas were assigned an additional score of 3, and buffers, a score of 1 (or essentially 33% of the full 10-point "mammals, fish, birds, and herps" score).

- Important Mammal Areas (IMAs):
IMAs have been defined a polygons across areas of SEPA by the IMA Committee. These areas supplement the other GAP and IBA data in this subgroup. Core areas were assigned a score of 3 (or essentially 33% of the full 10-point "mammals, fish, birds, and herps score).

Aquatic Resources Subcomponent

- National Wetland Inventory (NWI):
Wetlands were assigned scores by type as follows:

• 1 point: Substrate-only type wetlands (e.g., no vegetation, just rock, sand, or mud types)
• 2 points: Open water and aquatic bed wetlands
• 4 points: Unconsolidated emergent or forested wetlands
• 10 points: Emergent, scrub-shrub, and forested wetland types and combinations

- Hydric Soils:
Data on soil survey digital mapping for counties throughout the Piedmont were compiled from the Soil Survey Geographic Database (SSURGO). Hydric soils were mosaiced into a single coverage, clipped for the NLT expanded Piedmont ecoregion, and these cells were given a score of 3.

- Flood Plains:
Data were gathered from PASDA, mosaiced into a single coverage for the Piedmont area, clipped for the NLT expanded Piedmont ecoregion, and given a score of 5.⁷

- Forested Water Quality:
The percent of first- and second-order streams was expressed per Pennsylvania Small Watershed and the results reclassed according to a 10-quantile distribution. Forested landcover was selected from the regional land cover data set and expressed as a percent of forested landcover for each Pennsylvania Small Watershed, also on a 10% (10-quantile) classification system. The two data sets were then added together and divided by 2 and then normalized once again on a 10% quantile basis. A Pennsylvania Small Watershed given a score of 1 represents a watershed that is in the lowest 10% for a combination of forested land cover and percent length of first- and second-order streams, while a score of 10 for a PA Small Watershed indicates it is in the top 10% of watersheds for forested land cover and percent length of first- and second-order streams.

- Riparian Buffer Quality:
Riparian buffers of approximately 100 feet were created in the maps on either side of all streams or water bodies in the region. The regional landcover was then ranked for quality in support of aquatic habitat conditions by the Aquatics Expert Advisory Committee, such that the 15 landcover classes were assigned one of four habitat quality weightings as follows:

• 0 points: Commercial, urban, suburban, quarries, bare transitional
• 2 points: Row crops, recreational grass
• 5 points: Hay/pasture
• 10 points: All forest, water, and wetland types and bare rock (natural)

A Focal Variety algorithm was run on the clipped riparian buffer landcover quality weightings to indicate where aggregations versus fragmentation of land cover types existed. A 0, 2, 5, or 10 score was assigned where there were 4, 3, 2, or 1 landcover types within the focal variety zone of analysis (which used a 3-cell-by-3-cell analysis area).

The Focal Variety results were then multiplied by the weighted aquatics land cover habitat results and divided by 10.

The streams and water bodies results layer was then split from the original coverage into separate data layers as follows:

• All first- and second-order streams: 10 (or 0.625)
• All third- to fifth-order streams: 5 (or 0.3125)
• All streams of sixth order and higher and isolated water bodies: 1 (or 0.0625)

These three separate data sets were then mosaiced back together again using the weights noted above (as suggested and approved by the Aquatics Expert Advisory Committee).

The resulting data layers represent riparian buffer quality in very small linear spatial arrangements. Because such small areas will essentially get "buried" when compiled with broader spatial surfaces, it was felt it would be more meaningful to represent the final results by Pennsylvania Small Watershed. Therefore, the final step in the analysis was to convert the linear riparian quality values into averages per Pennsylvania Small Watershed, classified by 10% quantiles.

- Water Quality (2002) 303[d] List:
Stream segments from the 303[d] GIS data set were clipped to the PA small watershed boundaries and an average score obtained based on the quality ranking system provided above. The resultant map was then recalibrated to show results on a 10-quantile basis.

Since this data set was still incomplete across the entire ecoregion as of 2005, NLT used an interim 10-point ranking system that averages water quality results per PA small watershed throughout the Extended Piedmont ecoregion, as follows:

Attaining = 10 points
Unattaining = 0 points
Unassigned = 5 points

- Headwaters Protection:
The Aquatics Expert Advisory Committee desired to highlight the critical importance of headwater features, such as seeps, springs, and ephemeral streams, as well as the importance of first- and second-order streams in maintaining water quality in general. It was also noted by the group that headwater areas are more defensible from upstream pollution threats. As such, they tasked NLT with formulating a way to generate a measure that indicated "location in watershed," such that lands lower in a watershed were less valuable than lands higher in a watershed. NLT eventually decided that the best way to represent these values was through use of a flow accumulation grid. This grid was created from the Piedmont regional Digital Elevation Model (DEM), and the classification scheme used the following classes:

Values were expressed as averages per Pennsylvania Small Watershed, with final results displayed as 10% quantiles.

- Impervious Cover, 2000:
The Aquatics Advisory Committee helped NLT assign values to this data set, which became available from PSU via PASDA in early winter 2003. Impervious cover averages were generated per Pennsylvania Small Watershed. An "impact" of impervious cover ranking system was used to classify the results, centered around critical threshold impact values provided by Woods Hole Research Station (WHRS) and the Center for Sustainable Watersheds (CSW) of 6%, 10% and 20% respectively, where WHRS has research that implies water quality is largely unimpacted below 6% impervious cover watershed wide; and the CSW proposes that water quality is less impacted where impervious cover is 10% or less watershed wide and greatly impacted where impervious cover is 20% or greater watershed wide. Using these cornerstones for the ranking system provides us with the following value system:

- Impervious Cover Change, 1985 to 2000:
Using the impervious cover data from PSU from 1985 and 2000, and averaging it per Pennsylvania Small Watershed as described above, the 2000 condition was compared to the 1985 condition and the difference mapped in a new data set. Resulting values were classified using a 10-quantile classification system. Thus, a 10-score represents the watersheds which show the greatest amount of increase in impervious cover, while a 1-score represents the smallest percentage increase in impervious cover across a small watershed. In this manner, a high priority is placed on preserving lands in those watershed that are rapidly changing from a rural to an urbanized or suburbanized state.

Terrestrial Resources Subcomponent

- Steep Slopes:
Historically, steep slopes have deterred development to such an extent that they are somewhat of a predictor of intact forest conditions. If the slope is steep enough, there is a good chance that extensive timbering and thus high grading and soil compaction have been avoided in these areas. In addition, these slopes should be protected to reduce the threat of erosion. The location of slopes was calculated for the region using the DEM and assigned scores as follows:

• 0%–15%: 0
• 15%–25%: 2
• 25%+: 4

- Interior Forest Habitat:
NLT obtained an Interior Forest Habitat GIS layer from PSU/ERRI, which selected forest types from the landcover data set and applied a 300-foot buffer to clip away external "edge." The remaining forests were considered Interior Forest Habitat and were ranked according to size (in acres), as suggested by the Birds Expert Advisory Committee. The rankings have been calibrated specifically for conditions across the Pennsylvania Piedmont region:

- Natural Vegetation Habitat Blocks:
All natural vegetation and land cover classes were split out from the regional landcover data layer. The regional road data layer was compiled to show regional landscape fragmentation and size of remaining landscape blocks. Block size values were assigned based on input primarily from the Mammals Expert Advisory Committee, with regional adjustments based on conditions across the Pennsylvania Piedmont, as follows:

- Contiguous Grassland Habitat Blocks:
All hay/pasture land cover types were clipped from the regional land cover data set (92-94) for the Expanded Piedmont ecoregion. The layer was then intersected with the regional landscape blocks, as used in the Interior Forest and Natural Vegetation descriptions. Each contiguous Hay/Pasture polygon within a landscape block was then ranked according to size, using a scoring system as follows:

Contiguous Barrens or Scrub/Shrub Habitat Blocks:
All Bare/Transitional land cover types were clipped from the regional land cover data set for the Expanded Piedmont ecoregion. The layer was then intersected with the regional landscape blocks, as used in the Interior Forest and Natural Vegetation descriptions. Each contiguous Bare/Transitional polygon within a landscape block was then ranked according to size, using the following scoring system:

County Natural Areas Inventory & Pennsylvania Natural Heritage Program Subcomponent/Rarity Assessment

This subcomponent evaluates potential habitat areas for rare, threatened and endangered species. Data from the County Natural Areas Inventory and the Pennsylvania Natural Heritage Program were used to assign value to habitat areas.

NLT used The Nature Conservancy (TNC)'s explicit procedures for ranking and maintaining data on rare species and natural communities as the basis for this subcomponent. NLT used three measures to arrive at a composite rarity ranking: 1) the rarest element; 2) the number of rare elements; and 3) a weighted average element-occurrence score. Once each habitat was ranked, cells were reclassified using the quantile method.

Data used for this subcomponent is as of the most recent update of the CNAI publication within each county (up until March 2002); with the exception of rare plants which were updated by PA DCNR through March 2002 for each county in the Expanded Piedmont ecoregion.

Subcomponent Compilation and Final Ecological Resource Results
To generate the final ecological resource value scores, each layer was added with the others in its subcomponent. The following table summarizes the scoring values for each of the subcomponents and their respective layers.

The cumulative scores from each of the four subcomponents were normalized back to a 1 to 10 (10%) quantile classification system. While reclassing the data back to a 10% quantile system has the benefit of allowing easy data compilation and comparison as part of a relative ranking system, it also has the disadvantage of changing the proportional weight of each subcomponent from its original value to a uniform 25% for each subcomponent (since there are 4 subcomponents). In order to recalibrate the scoring to achieve the original subcomponent values, an adjustment factor was required. The adjustment factors used to accomplish this goal are shown in the table below.

Following the combination of the four subcomponents, the results of the ecological resource assessment for the Expanded Piedmont region were clipped to the five-county SEPA region and again recalibrated to show the results in 10% quantiles. Developed lands, protected lands, and large areas of open water were then removed from the analysis. The resulting ecological resources map displays priorities 6 through 10 in progressively deeper shades (see Map 3: Ecological Open Space Priorities).

Before the removal of developed and protected lands, the results indicate that high-priority ecological resource lands (levels 8, 9, and 10) in the region total 402,700 acres. Removing developed land leaves 332,600 acres. Of these high-priority lands, 76,600 are already protected, leaving 256,000 acres of unprotected and undeveloped high-priority ecological resource lands.

High-priority ecological lands are not evenly distributed: The greatest concentration of high-value ecological land is located in a band stretching from the Unami Hills in Montgomery County across the northern tier of Bucks County to the Delaware River. In Bucks County, this area encompasses the Quakertown Swamp; Rock Hills; the Tohickon, Cooks Creek and Tinicum watersheds; and the Pallisades. The second largest high-value area is located within the Big Woods region of Chester County. Remaining high-value lands are scattered throughout the rest of the 5-county region, but are found primarily in northern and eastern Bucks County, southern Chester County, and western Delaware County. River and stream corridors, as well as wetlands, highlands, steep slopes, and forested areas generally have high values throughout the region.

As with agricultural lands, much of the high-priority land for ecological resources is contiguous to other high-value land and derives part of its value from this connectedness. High-priority agricultural lands are generally not high-priority for ecological purposes. However, as will be revealed in the combined analysis, a subset of lands are a high-priority for both uses.

Note that the ecological resource value of most lands within Philadelphia is low, due to development. Several stream corridors have high values where they enter Philadelphia, but develop lower values as they pass further into the city.

⁷ The PASDA data used for this layer is based on the Federal Emergency Management Agency's (FEMA's) Flood Insurance Rate Maps. These maps were digitized by Pennsylvania DEP and assembled into a statewide coverage in 1996. FEMA's floodplain maps are currently in the process of being updated. New floodplain maps reflect increases in stormwater runoff from development that has occurred since the FEMA maps were first prepared in the 1960s, 70s and 80s, as well as more detailed elevation data. For example, a Temple University study to remap the floodplains of the Pennypack Watershed (September, 2006) shows a 24 percent increase in floodplain area over existing FEMA maps.