Matthias Leu, Steven E. Hanser, and Steven T. Knick. 2008. The human footprint in the west: a large-scale analysis of anthropogenic impacts. Ecological Applications 18:1119–1139.

Appendix A. Methods used to acquire and develop spatial data sets of anthropogenic features.

We prepared 16 spatial data sets delineating the extent of anthropogenic features across the western United States. These data sets were the foundation of the input models and the human footprint model. First, we describe spatial data acquisition and manipulations (see also metadata of each spatial data set) according to: (1) point features (e.g., campgrounds, landfills), (2) line features (e.g., roads, irrigation canals), and (3) polygon features (e.g., agricultural land and urban areas). Second, we outline the development of four spatial data sets used to evaluate the output of the human footprint model. Third, we show data implemented to calculate physical effect area of anthropogenic features. All spatial data sets were based on a 90-m resolution, converted from shapefile to grid, obtained between November 2002 and December 2003, processed using ESRI ArcView 3.2 or ArcMap 9.0. (ESRI 1998), and are housed, except the campground spatial data set, on the SAGEMAP PROJECT website (http://sagemap.wr.usgs.gov) for public access.

Point features

We developed the campground spatial data set by digitizing locations of campgrounds referenced in DeLorme gazetteers (DeLorme 2001) for each state. We used road and hydrology data sets as reference layers when delineating the locations of campgrounds and merged the state-level campground layers to produce a western United States map of campgrounds. We obtained fire ignition locations between 1986 and 2001 from the National Fire Occurrence Database and selected human-caused fires using the “Cause2” attribute. Data for landfills were obtained from multiple sources. For California we used the solid Waste Information System Database obtained from California Integrated Waste Management Board website (http://www.ciwmb.ca.gov/SWIS/). All landfill and transfer stations were selected with the following string: "OPERATIONA" = 'Active' AND "ACTIVITY" = 'Solid Waste Landfill' OR "ACTIVITY" = 'Large Volume Transfer/Proc Facilit' OR "ACTIVITY" = 'Limited Volume Transfer Operation' OR "ACTIVITY" = 'Limited Volume Transfer Operation' OR "ACTIVITY" = 'Medium Volume Transfer/Proc Fac' OR "ACTIVITY" = 'Small Volume Transfer Station' OR "ACTIVITY" = 'Solid Waste Disposal Site'. The Wyoming landfill data were obtained from the Wyoming Geographic Information Science Center Clearinghouse’s Digital Wyoming Atlas (http://www.wygisc.uwyo.edu/clearinghouse/All.html). The Colorado landfill data, in shapefile format, were obtained from Colorado Hazardous Materials and Waste Management Division website (http://www.cdphe.state.co.us/hm/hmgis.htm). For Idaho, Montana, New Mexico, and Utah, the landfill data were downloaded from the USGS website (http://water.usgs.gov/lookup/getspatial?landfill). For Nevada, we received landfill spatial data sets from the Nevada Division of Environmental Protection (Brenda Harpring, Environmental Scientist, 333 W. Nye Lane Carson City , Nevada 89706). For Oregon we received spatial landfill data sets, spatially referenced by township/range/section (T/R/S), from the Department of Environmental Control (Cameron Oster, Oregon DEQ). The corresponding T/R/S was selected in ArcGIS to delineate landfill locations as the center point of the section. For Washington, we downloaded the data from (http://www.ecy.wa.gov/services/gis/data/data.htm) and selected landfills from the Facilities Location Coverage. For Arizona we obtained a map from the Arizona Department of Environmental Control (http://www.adeq.state.az.us/environ/waste/solid/images/map.gif) depicting active landfill locations. Locations were digitized in ArcGIS using 2002 TIGER/Line highway as a reference layer at 1:250000 scale. Spatial data sets on oil and gas wells were downloaded from the USGS National Oil and Gas Assessment website (http://energy.cr.usgs.gov/oilgas/noga/) and included the location of active and inactive wells up to 1995. Spatial data sets on rest stops were developed using information provided on websites from State Department of Transportation (DOT) for Oregon (http://www.tripcheck.com/Pages/RAentry.asp), Washington (http://www.wsdot.wa.gov/mapsdata/geodatacatalog/Maps/noscale/DOT_Cartog/restarea.htm), Nevada (http://www.usroadconditions.com/nvrest.html), Idaho (http://www2.state.id.us/itd/ida-road/restareas.pdf), Utah (http://www.dot.state.ut.us/mnt/RestAreas/), Arizona (http://tpd.az.gov/gis/maps/pdf/az_rest_areas.pdf), Montana (http://www.mdt.state.mt.us/departments/transportation_planning/misc/pdf/99restarea.pdf; ftp://ftp.mdt.state.mt.us/map/hwymap_side1.pdf), and Wyoming (http://dot.state.wy.us/web/e_docs/restarea.pdf). TIGER/Line roads and U.S. Census Gazetteer Cities were used to digitize the location information obtained from the State DOT information. For California we obtained coordinates of rest areas from (http://www.dot.ca.gov/hq/maint/ra/Statewide.htm) and imported the data as an event theme and converted to shapefile. For Colorado we obtained data from Colorado DOT (http://www.dot.state.co.us/App_DTD_DataAccess/GeoData/index.cfm?fuseaction=GeoDataMain&MenuType=GeoData). We re-digitized these rest area locations based on the Colorado DOT data to increase accuracy of the location along the roadway.

Line Features

Density of linear features, such as roads, power lines, telephone lines, and railroad tracks have been shown to impact synanthropic predator abundance patterns (Knight et al. 1995). Following Knight et al. (1995), we built a linear features spatial data set using linear features such as roads (i.e., interstate highways, federal and state highways, secondary roads), railroads, power lines, and irrigation canals. However, we lacked spatial data on telephone and feeder power lines. We merged all spatial data shapefiles and calculated density of linear features using a moving window analysis with 1-km search radius around each 90-m cell (Arc/Info: "LINEDENSITY"). Because Knight et al. (1995) found that Common Raven (Corvus corax) abundance indices increased when four to five linear features ran parallel to the sampling transect, we selected density values greater than 3.5 km/km² in the final output. This density threshold translated into four to five linear features/ km². For interstate highways, federal and state highways, and secondary roads, we used U.S. Census 2000 TIGER/Line data (http://www.esri.com/data/download/census2000_tigerline/index.html) to delineate roads throughout the study area. First, we merged all county-based data sets and then created three spatial data sets, interstate highways (CFCC2 = 'A1'), state and federal highways (CFCC2 = 'A2' OR CFCC2 = 'A3'), and secondary roads (CFCC2 like 'A%' AND (not (CFCC2 = 'A1' OR CFCC2 = 'A2' OR CFCC2 = 'A3'). To develop an irrigation canal spatial data set, we obtained the US Census TIGER/Line hydrology data for all counties in the study area (http://www.esri.com/data/download/census2000_tigerline/index.html). We then merged all county data sets, selected features with the attribute CFCC2 = 'H2', and exported this query to a new data set. We obtained power line corridors from the Interior Columbia Basin Ecosystem Management Project (ICEBMP, http://www.icbemp.gov/). The spatial data for railroads was produced by clipping the 100K Railroads line network from the Bureau of Transportation Statistics, National Transportation Atlas Databases (http://www.bts.gov/gis/) to the study area extent.

Polygon features

For Agricultural land, we used the "Sagestitch" map (Comer et al. 2002), downloaded from the SAGEMAP website (http://sagemap.wr.usgs.gov/), to delineate the current distribution of agricultural land in the western United States (excluding New Mexico and Arizona). We reclassified Sagestitch to contain only agriculture lands using the query: Value = 101. For eastern Washington, we replaced Sagestitch with a contemporary data set, Shrubsteppe Mapping of Eastern Washington using Landsat Satellite Thematic Mapper data, categories 7 and 11. Furthermore, because the Sagestitch map did not cover the entire study area, we appended missing agricultural lands using state GAP coverages for Arizona, New Mexico, California, Oregon, and western Washington. Agricultural land was selected out of the corresponding GAP shape files or grids using the following queries: Arizona: Grid-Code = 82, California: CNDDB1 >= '20000' and < '30000' and SP1A >= '20000' and < '30000', New Mexico: covertype_code = 9110 OR 9120, Oregon: ORGAP = 125, Washington: WAGAP >= 300 and WAGAP < 400. We developed a grid of forest habitat from the “Sagestitch” map (http://sagemap.wr.usgs.gov), and GAP landcover data (New Mexico, Oregon, Arizona, Washington, and California) by using a crosswalk among forest types (deciduous forest, mixed deciduous/conifer forest, conifer forest) and the input data layers (Table 1).

TABLE A1. Cross walk employed for reclassification for input data sets values to forest spatial data set values.

We developed the human populated area spatial data set following a modification of Schumacher et al. (2000), with populated areas defined as > 1 person/ha (i.e., rural/exurban but including suburban and urban as defined by Marzluff et al. 2001) by incorporating census and land ownership data. The census data were downloaded from the U.S. Census 2000 TIGER/Line data (http://www.esri.com/data/download/census2000_tigerline/index.html). We merged all census block polygon shapefiles for the western states. Because populated areas on federal land are negligible, we created a layer of private and BIA land (Bureau of Indian Affairs) using the westUS_own (http://sagemap.wr.usgs.gov) ownership layer and the query: Pub_pvt = Pvt. We used the raster calculator in ArcGIS to determine area of census blocks that were private (Priv_cb_fid = priv_owner * [cb_fid-1]). We then calculated area (ha) for each polygon and joined the 2000 census block data table to the private landownership shapefile. In a new field, we calculated population density (individuals/ha) corrected for public land in census blocks and then selected areas of population > 1 individual/ha to create the populated areas spatial data set.

Evaluation of the human footprint extent

The Nature Conservancy Ecological Provinces (The Nature Conservancy 2001) were downloaded from (http://sagemap.wr.usgs.gov/ftp/regional/TNC/tnc_us_eco2001.zip). We delineated land stewards using the westNA_own shapefile (http://sagemap.wr.usgs.gov) and converted it to a grid layer using the CA_OWN attribute as grid values. Roadless areas were downloaded from United States Department of Agriculture (USDA), United States Forest Service (USFS) Roadless Area Conservation website (http://roadless.fs.fed.us/documents/feis/data/gis/coverages/) using the categories defined in the metadata: "1B = Inventoried Roadless Areas where road construction and reconstruction is prohibited; 1B-1 = Inventoried Roadless Areas that are recommended for wilderness designation in the forest plan and where road construction and reconstruction is prohibited; 1C = Inventoried Roadless Areas where road construction and reconstruction is not prohibited". To delineate protection status of lands, we obtained stewardship maps for each state from the National Gap Stewardship Status GAP analysis program (http://www.gap.uidaho.edu). In the analyses, we used all management codes except code 0 (water). Verbatim descriptions of status codes 1–4, taken from the metadata, are listed in Table A2.

TABLE A2. Definition of National Gap stewardship map management codes (downloaded from the GAP analysis program site).

Development of the physical effect area

TABLE A3. Methods employed to calculate the physical effect area of anthropogenic features.

† For point features we reported total area covered by anthropogenic feature, for line features we reported width of anthropogenic features.

‡ Data were collected in Ada County, Idaho, and in Churchill, Mineral, Pershing, and Washoe County, Nevada. For anthropogenic features other than interstate and state/federal highways, means are based on independent data points; interstate and state/federal highways were measured twice in cases where there was a reduction/increase in traffic lanes.

LITERATURE CITED

Comer, P, J. Kagan, M. Heiner, C. Tobalske. 2002. Current Distribution of Sagebrush and Associated Vegetation in the Western United States (excluding NM and AZ). Interagency Sagebrush Working Group. (http://sagemap.wr.usgs.gov).

DeLorme. 2001. Atlas and Gazetteer. DeLorme, Yarmouth, Maine, USA.

ESRI. 1998. Arc/Info version 7.1. Environmental Systems Research Institute, Inc., Redlands, California, USA.

Ingelfinger, F. M., and S. Anderson. 2004. Passerine response to roads associated with natural gas extraction in sagebrush steppe habitat. Western North American Naturalist 64:385–395.

Knight, R. L., H. A. L. Knight, and R. J. Camp. 1995. Common ravens and number and type of linear rights-of-way. Biological Conservation 74:65–67.

Marzluff, J. M., R. Bowman, and R. Donnelly. 2001. A historical perspective on urban bird research: trends, terms, and approaches. Pages 1–17 in J. M. Marzluff, R. Bowman, and R. Donnelly, editors. Avian ecology and conservation in an urbanizing world. Kluwer Academic Publishers, Boston, Massachusetts, USA.

Neatherlin, E. A., and J. M. Marzluff. 2004. Responses of American crow populations to campgrounds in remote native forest landscapes. Journal of Wildlife Management 68:708–718.

Schumacher, J. V., R. L. Redmond, M. M. Hart, and M. E. Jensen. 2000. Mapping patterns of human use and potential resource conflicts on public lands. Environmental Monitoring and Assessment 64:127–137.

The Nature Conservancy. 2001. The Nature Conservancy's Ecoregions of the United States. (http://sagemap.wr.usgs.gov).