Mapping Potential Bird Habitats for the
Southern Blue Ridge Province

Scott M. Pearson

Biology Dept., Mars Hill College, Mars Hill NC 28754

Email: spearson_at_mhc.edu, Phone: (828) 689-1402

I. Introduction

During 1996-1998, the Southern Appalachian Forest Coalition (SAFC) and consulting ornithologists held meetings to discuss the status of bird populations in southern Blue Ridge Province and to consider strategies for conservation of single and groups of bird species on public and private lands. The conservation of these birds is complicated due to the relatively large number of species and the complexity of habitats (from local to landscape spatial scales) in this region. During these meetings, a scheme for combining bird species into a small number of groups emerged that should simplify questions about protecting and managing these species. This scheme involved categories of habitat use, and habitat definitions were based on elevation zones and coarse habitat types. Therefore, maps of these habitat types became desirable.

Four maps of potential bird habitat were developed using elevation and topographic data obtained from USDA-Forest Service and land-cover data from the Southern Appalachian Assessment (SAA) (Hermann 1996). The purpose of the maps is to show the spatial location, pattern, and abundance of coarse habitat types that are used by bird species in the southern Blue Ridge Province. These maps may be used by land managers and researchers to examine the relative abundance and spatial pattern of habitats in their areas of interest.

The maps were produced using ArcView 3.0 (ESRI 1997). They are grid (raster) maps projected in Albers, the same projection as the SAA map layers. The cell size is 100 m. The spatial extent of the maps covers the southern Blue Ridge Province in western North Carolina. This extent covers all of the Pisgah and Nantahala National Forests and intervening private lands. The extent of the mapped region (hereafter referred to as "study area") was determined by the areas covered by elevation and landform maps. A description of the four habitat maps and the steps involved in producing them are outlined below. Throughout this document, names of thematic maps included in the ArcView Project are shown in bold type. This work was performed in collaboration with SAFC, based in Asheville, NC.

II. Intermediate Steps and Map Products

A. Elevation and Landform data

Steve Simon (US Forest Service, National Forest in NC, Asheville NC) provided a data set entitled "The Nantahala-Pisgah Digital Terrain Model GIS Data Base." It consisted of ArcInfo 7.x coverages which are compatible with ArcView 3.0. These grid coverages consisted of digitial elevation models (DEM), maps of landform index (LFI) (McNab 1993), and polygon coverages of NC Hydrological Units. The DEM and LFI maps were based on level-2 (mostly) and level-1 DEMs obtained from USGS. Simon edited and edge-matched the USGS DEMs to produce an elevation map of the study area, defined by the hydrological units containing lands in these two national forests. Using this elevation map, he then produced a map of LFI. The DEMs show elevation in feet above sea level.

As a basis for the bird habitat maps, two preliminary maps were produced using Simon's DEM and LFI coverages. A map of elevation zones produced showing regions with the following four elevation categories: <3500 ft, 3500-3999 ft, 4000-5499 ft, and >=5500 ft. Likewise, Simon's LFI map was reclassified to highlight landscape positions likely to harbor sheltered, mesic cove forests. The landform index of McNab (1993) produces continous values. Sites having a larger LFI are more sheltered from insolation and have more convex landforms. Therefore, LFI is greatest at the bottom of narrow, deep ravines and least on ridges and peaks. Based on a comparison of Simon's LFI map and USGS 7.5 minute quad maps (hard copy) of the author's study sites in these National Forests, a threshold value of 0.32 was chosen. Sites with LFI values greater than or equal to this value were classified as coves. These reclassifications were performed using ArcView's Spatial Analyst extension.

Simon's data set was projected in NC State Plane. Hugh Irwin (of SAFC) reprojected the reclassified elevation and LFI maps into Albers to be compatible with the SAA data. The elevation zones and coves maps can be found in the "Intermediate Maps" view of the ArcView project.

B. Land Cover

Land-cover maps for this analysis were derived from the SAA data sets. The coverage "Landcov" on SAA CDROM no. 3 having the final 17-category classification scheme was used. Please refer to the SAA data (Hermann 1996) for a formal description of the data sources and techniques used to produce the SAA land-cover map. The spatial and cover-class accuracy of the bird habitat maps is dependent on the quality of the SAA land-cover data. One intermediate map of coarse forest types was produced by reclassifying the SAA data. This map consists of four classes shown in Table 1. There were no areas of SAA class "17 indeterminate- clouds, shadows" in the study area. The SAA data were resampled to grid-cell size of 100 m during this reclassification. This map is also stored in the "Intermediate Maps" view.

Table 1. Reclassification of SAA land-cover categories for a map of coarse forest types.
Coarse Forest Type SAA Land-cover Class (number and description)
Deciduous 1 Northern Hardwood Forests
2 Mixed Mesophytic Hardwood Forests
3 Oak Forests
4 Bottomland Hardwood Forests
Mixed Conifer-Deciduous 8 White Pine / Hemlock / Hardwood Forests
9 Mixed Pine / Hardwood Forests
Conifer 5 White Pine / Hemlock Forests
6 Montane Spruce-Fir Forests
7 Southern Yellow Pine Forests
Non-forest 10 Herbaceous
11 Barren
12 Agriculture - Pasture
13 Agriculture - Cropland
14 Wetlands
15 Developed
16 Water


C. Analysis Mask and Resolution

ArcView's Spatial Analyst extension was used to produce the final maps. During these analyses, the spatial resolution was set to 100-m grid cells. An analysis mask was also imposed. This mask was based on the LFI map from Simon because this map had the most limited spatial extent. The mask is called mask_lfi in the "Intermediate Maps" view.

III. Final Bird Habitat Maps

The following four maps were derived from the intermediate maps described above and/or the SAA land-cover data. For each map, the cover classes are described. The numerical class number (i.e., grid cell value) is given in parentheses.
(Click on the highlighted heading to see an image of the map. Use the browser's BACK key to return to this page.)

A. High Elevation Areas.

These areas are of interest due to a number of endemic species in the southern Blue Ridge Province. Habitats in these areas have been and are subject to anthropogenic impacts such as logging, air pollution, and introduction of alien species. The map, high elevation, consists of three categories: (1) Elevations <4000 ft; (2) Elevations >=4000 ft, non-forest; and (3) Elevations>=4000 ft, forest. The elevation zones and coarse forest types maps were used to produce this map.

B. Low Elevation, Mesic Cove Forests

Cove forests in this region support some of the greatest diversity and abundances of breeding birds. The map, low elevation coves, consists of two categories: (0) Other, and (1) Coves. Cove forests were defined as cells having the following three characteristics: (a) elevation <4000 ft, (b) LFI >0.32, and (c) forested land-cover. The intermediate maps elevation zones, coves, and coarse forest types were used. Aspect was not incorporated in this analysis because LFI values in Simon's data set were already adjusted to incorporate aspect. Northeast-facing slopes received higher LFI values.

C. Hemlock Forests

Hemlock forests at high elevations may be used by boreal species. At lower elevations, birds of cove forests and species of interest, such as Swainson's Warbler, Northern Parula, and Black-throated Green Warbler, use these forest stands. The map, hemlock, consists of three categories: (0) Other cover types, (1) Hemlock forests, elevation <3500 ft, and (2) Hemlock forests, elevation >=3500. The elevation zones map was used. Hemlock forests were not mapped as a separate category in the SAA land-cover data because this species did not have a unique spectral signature and is often mixed with others in forest stands on the landscape. To map hemlock forest, SAA land-cover classes 5 and 8 (see Table 1, conifers) were aggregated. Due to this aggregation and the 100-m cell size employed, the absolute area of this forest type could be over- or underestimated by the hemlock map. This map should, however, provide users with general information about the location and relative abundance of this forest type in different regions of the study area.

D. Northern Hardwood Forest Zone

Northern hardwood forests are used by some boreal species and by bird species, such as Rose-breasted Grosbeak, that are noteworthy for the southeastern US. The map, n. hardwood, consists of five categories. The elevation zone of northern hardwood forests was set at [3500 ft>= elevation <5500 ft]. Category (0) shows areas outside of this elevation zone. The next four categories show different land-cover categories within this elevation zone: (1) Deciduous forest, (2) Mixed forest, (3) Conifer forest, and (4) Non-forest. The intermediate maps elevation zones and coarse forest types maps were used.

D. Public Lands

A map of public lands was also derived from the SAA data set. This map shows categories 1, 2, and 4 from the SAA map, pub_own. These categories include National Park and National Forests. Some state- and county-owned lands may not be included.

IV. Literature Cited

ESRI. 1997. ArcView GIS. version 3.0a. Environmental Systems Research Institute; Redlands, CA.

Hermann, K.A. (ed.) 1996. The Southern Appalachian Assessment GIS Data Base CD ROM Set. The Southern Appalachian Man and the Biosphere Program. Norris, TN.

McNab, W. H. 1993. A topographic index to quantify the effect of mesoscale landform on site productivity. Canadian J. For. Research 23:1100-1107.