| GAP home | USGS home |

(Status as of April 2005)

Southwest Regional GAP (SWReGAP)

An update is under way for the five-state region encompassingArizona, Colorado, Nevada, New Mexico, and Utah. Statecoordination for the project is facilitated through the SWReGAPweb site <http://leopold.nmsu.edu/fwscoop/swregap/default.htm>.

Anticipated completion date: October 2005

Contacts:

Julie Prior-Magee, SWReGAP Coordinator
USGS/BRD, Las Cruces, New Mexico
jpmagee@nmsu.edu
505-646-1084

Pat Comer and Keith Schulz
NatureServe, Boulder, Colorado
pat_comer@natureserve.org, 303-541-0352
kschulz@natureserve.org, 303-541-0356

Arizona: Kathryn A. Thomas, PI
USGS/BRD Southwest Biological Science Center
Colorado Plateau Research Station, Flagstaff
Kathryn_A_Thomas@usgs.gov, 928-556-7466 x235

Colorado: Donald L. Schrupp, PI
Colorado Division of Wildlife
Habitat Resources Section, Denver
hqwris@lamar.colostate.edu,
303-291-7277

Nevada: David F. Bradford, Co-PI
U.S. EPA—Office of Research and Development, Las Vegas
bradford.david@epa.gov
702-798-2681

William G. Kepner, Co-PI
U.S. EPA—Office of Research and Development, Las Vegas
kepner.william@epa.gov
702-798-2193

New Mexico: Ken Boykin, Co-PI
NM Coop. Fish and Wildlife Research Unit, Las Cruces
kboykin@nmsu.edu
505-646-6303

Utah: John Lowry, Co-PI
RS/GIS Laboratory, Utah State University, Logan
jlowry@gis.usu.edu, 435-797-0653

Land cover: The RS/GIS Lab at Utah State University is theregional land cover mapping lab for the five-state Southwestregion. The regional project focused on four major objectives in2004: (1) completing land cover modeling activities; (2) modelvalidation; (3) producing a regional mosaic of the mappingzones; and (4) developing a data delivery system for theprovisional land cover product.

The majority of natural and seminatural land cover classes weremodeled using a decision-tree (DT) classifier. Advantages ofDT include the ability to use both continuous and categoricalpredictor data sets with different measurement scales, goodcomputational efficiency, and an intuitive hierarchicalrepresentation of discrimination rules. Decision-tree modelswere validated by generating initial models using 80 percentof available samples, while withholding 20 percent of samples.Withheld samples were randomly selected and stratified bycover class. Withheld sample polygons were intersected throughthe land cover map to create an error matrix, presenting users,producers, and overall “accuracies.” Using the 4 km overlapregion between mapping zones, a “cutline” was used to edge-match adjacent mapping areas where land cover discontinuitiesresulted from the modeling process. The resulting five-stateregion mosaic was qualitatively reviewed by the five state teamsand NatureServe. Following review, a limited number of errorswere “flagged” for final editing. The “edits” were determined tobe relatively easy to correct with localized recoding, or a simpleconditional model, and were made to the regional map.

The SWReGAP land cover data set is currently available tothe public with “provisional” status from <http://earth.gis.usu.edu/swregap/> (see Figure 1). Because the data set encompassessuch a large region, the web site allows users to downloadspecific geographic segments of the region, such as individualstates, counties, or ecoregions. Additionally, the web site offersan Internet map server from which users can interactivelyclip a specified rectangle in the region. The clipped data set issubsequently bundled with metadata and made available fordownloading.

Figure 1. SWReGAP provisional land cover data.

Animal habitat modeling: The regional project focused on sixobjectives during 2004: (1) collecting habitat modeling attributes;(2) creating region-wide modeling data sets; (3) creating adatabase to facilitate association compilation, expert review, andmodification, and potential end-user application; (4) internaland expert review; (5) accuracy assessment; and (6) conductinga regional animal-habitat modeling workshop in Las Vegas,Nevada, in March 2004.

The region is working with an MS Access database to facilitatedata collection and to compile taxa specific information formodeling purposes. The intent was to create a data set thatmanages information and was used to construct each taxon’swildlife habitat relationship model. Included within the databaseis a method for defining range limits using the eight-digithydrologic unit code (HUC). The database also incorporatesthe core data layers the region had identified to be minimallyaddressed in each wildlife habitat relationship model. These coredata layers are land cover, elevation (minimum and maximum),slope, aspect, soils, hydrology (distance to and association withpermanent water), and patch size. Species were allocated to eachstate based on expertise and species distribution. These stateswere responsible for creating the habitat models for those species.The individual databases were then combined and currently anymodification to the database is done through an online connectionto the master database. All species data collection is completeas of this report, with modifications occurring as internal andexternal reviews are completed.

The region has undergone an internal review process to checkconsistency within the models and to provide the frameworkfor an external review. The internal review is complete as ofFebruary 14, 2005. Expert review is beginning as of February 1,2005, and is scheduled to be completed by April 30, 2005. Theregion will complete the standard gap analysis habitat modelingmeasure of agreement, as well as a measure of agreementwith existing species occurrence records. States are currentlyidentifying qualified species lists for the standard measure ofagreement; these lists will then be provided to the regionallaboratory. The Arizona project, in coordination with the regionallaboratory, is identifying a procedure for using existing data tomeasure the degree of concordance between habitat models andspecies occurrence records. This analysis will be done as modelsare completed, and is scheduled to be finished by July 2005.

Land stewardship mapping: The final regional stewardshipand management status map is expected to be complete by June2005. External review of stewardship mapping products beganin December 2004. Nevada stewardship and management statusmaps have undergone the external review process, resulting inupdates to internal parcel boundaries and refinements to the GAPstatus codes. The draft maps for Colorado are also completeand the external review is scheduled for March 2005. Externalreviews are also scheduled for Arizona, New Mexico, and Utahthis spring. As a conservative estimate, the regional stewardshiplab has collected over 300 management planning documentsfrom various federal, state, and county entities. The process ofreviewing current management plans, interviewing various landstewards, and assigning the GAP status codes is complete. Inaddition, most of the digital boundary information for all fivestates has been collected, and cooperators have been generouswith providing digital parcel data layers. Currently, the regionalstewardship lab is in the process of assembling the GIS databaseusing the geodatabase format to maintain data integrity. In aneffort to keep the stewardship mapping effort consistent acrossthe region, the regional stewardship lab digitizes additionalinternal information when digital information is unavailablefrom local sources. This effort is designed to provide a consistentproduct across the region in both the level of mapping detail andthe assignment of the GAP status codes. In addition, the detailand refinements of the stewardship product will create a betterassessment in the final gap analysis.

Analysis: Analysis for SWReGAP will take place when allmapping tasks are completed. Land cover analysis and animalhabitat modeling analysis will begin in May 2005.

Reporting and data distribution: All products derived fromSWReGAP are scheduled to be complete by approximately October 2005.