Notes and Announcements
Next Annual GAP Meeting
Make your plans now for the 1998 Annual GAP Investigators Meeting, to take place on July 20-24, 1998. The meeting will be hosted by the California GAP Project, headed by Dave Stoms and Frank Davis at the Department of Geography, University of California-Santa Barbara. The sessions will be held on the picturesque UCSB campus, beautifully located among palms and eucalyptus trees on a plateau overlooking the Pacific ocean, about 10 miles from downtown Santa Barbara. Accommodations in dormitories as well as hotel rooms will be are available.
The meeting is open to GAP investigators, their staff, project collaborators, and others interested in GAP methods and results. Please consult the following Web site for more information on the Santa Barbara area: http://www.biogeog.ucsb.edu/projects/gap/gap98. Information on the meeting will be posted to that site as it becomes available. A call for papers will be sent out in January.
A Progress Report on WY-GAP Extension
With the official completion of the Wyoming Gap Analysis Project (WY-GAP) this past February, we are now developing a coordinated approach to provide access to these valuable data via the Internet and through a coordinated extension program. This effort is known as the Wyoming Bioinformation Node or WBN. The establishment of the WBN is funded by the USGS and is part of a distributed federation of biological data and information sources, also known as the National Biological Information Infrastructure (NBII).
The WBN (http://www.sdvc.uwyo.edu/wbn) is also part of a larger spatial data clearinghouse effort being developed by the University of Wyomings Spatial Data and Visualization Center (SDVC). Through the natural resources data clearinghouse of the SDVC, WY-GAP data has found a home. All of the spatial databases being served through the SDVC and WBN include metadata documentation in accordance with the FGDC Content Standard for Digital Geospatial Metadata for implementation under the Z39.50 service protocol which utilizes I-Site and I-Search "browse and search" software.
Another important component of the WBN is the "bioinformation extension program" which is designed to promote the use and integration of the WY-GAP databases into natural resource planning, management, and education. A portable ArcView demonstration of the WY-GAP databases is in development to showcase the utility and integration of these data with other natural resource databases. Also included in the extension program is a county-level planning support project. This pilot project will assist Teton County, Wyoming, in creating a biological decision support system (DSS) in a desktop computing environment to help promote biological considerations in traditional land use planning and management decision-making processes. The system will incorporate biological and other natural resource databases with a simple graphical user interface and custom designed tools that will not require prior knowledge of GIS or biology on the part of the planning staff. We anticipate having the biological DSS operational early next spring.
GAP Data Distribution on CD-ROM
Two innovative CD-ROM products from Arkansas and California are being added to those developed in Utah as model data dissemination and education products. Arkansas is developing the prototype for state final report CDs that will feature hypertext plus linked graphic images of all the GAP data and interpretive analysis maps. California is working with ESRI to develop a CD with GIS functionality built in. Both products are scheduled for release in late 1997. In the meantime, interactive and educational CDs that are very popular continue to come out of Utah (see http://www.usu.edu/~cliff/UGA.html, the Utah Geographic Alliances "Utah GAP Education Project" CD for a great example).
On the national front, the EROS Data Center has agreed to provide Internet access to the GAP data. The New Mexico GAP data will be used to develop and test the interface this winter. The National GAP office plans to produce a standard set of CDs for each completed project that will contain HTML and Adobe Acrobat versions of the report and graphic versions of all coverages (if provided by the project) on one set of CDs, and the GIS and other ancillary data on another set (perhaps with GIS query capability built in). State projects may produce any other CDs they feel appropriate and can develop with their cooperators.
We plan to provide a number of CDs for each state to deliver to cooperators. Further requests for CDs will be met by in-state distribution if a center is available, or a "press/print on demand" center linked through the GAP home page. Because the CDs will not contain any proprietary software, we anticipate being able to sell them for the cost of the media and pressing, plus a small handling fee, by the distribution centers. It is critical that states follow the data delivery protocols to facilitate easy and quick production of the CDs because we will not have a budget for data reformatting. In the coming year, we plan to issue more detailed protocols for directory structures to aid this process. If you have any questions or comments, contact Patrick Crist (firstname.lastname@example.org).
Arkansas Vegetation Map Wins 2nd Place in Contest
The Arkansas Statewide Vegetation Map received a second place in the prestigious 1996 International Intergraph Users Conference. The conference brings together some 4,000 users and developers of Intergraph systems from around the world. According to Dataquest, Intergraph is the worldwide leading producer of GIS software and hardware. Each year the conference organizers sponsor a cartographic map production contest. Winning this contest is particularly significant since Intergraph map production software and hardware is used by the great majority of commercial map production companies, most national mapping agencies (particularly in Europe), and many universities offering cartographic degrees.
The cartographic part of the Arkansas Statewide Vegetation Map was done by Stephan Pollard, a graduate student in the University of Arkansas Geography Department and an employee of the Center for Advanced Spatial Technologies at the University (home of AR-GAP). The map won second place in the "Best Overall Page Layout and Design" category as judged by an international panel.
The map displays the GAP vegetation which has been derived from classified TM imagery and extensive ground-truth data at a scale of 1:600,000. Thirty-seven vegetation and land cover classes are represented. In addition to the map data, Pollard also developed an innovative legend structure which visually displays the structure of the vegetation classification as well as extensive marginalia improving the usefulness and readability of the map.
W. Fredrick Limp
Use of SPOT and CIR by Louisiana GAP
Typically, federal and state efforts to maintain biodiversity have relied on protecting species once they become threatened or endangered. While it is important to protect these valuable species, a better approach is to examine entire regions for native animal and plant species distributions in conjunction with current management practices of conservation lands. Gap analysis provides this regional assessment of animals and plants giving land managers information that extends beyond their jurisdictions to facilitate better management decisions. While gap analysis provides a regional view, it should not be seen as a replacement for intensive site-specific inventories and assessments.
The Louisiana GAP project is using Landsat TM imagery along with auxiliary data sets (SPOT, CIR, field survey data) to classify the 23 vegetation formations in Louisiana. Ten TM scenes were needed to provide complete coverage of the state. There are many circumstances to consider when classifying imagery, especially when dealing with such a large amount of data. We decided on using an unsupervised classification scheme using the ERDAS 7.5 software. The individual TM scenes were subsetted into 900 x 900 25-m pixel blocks, because we prefer the method of having images match our computer screens at a 1:1 ratio to facilitate class identification. Otherwise, when working with a full TM image, the off-screen portions of the image containing a vegetation class being manipulated are not visible.
In 1995, the National Biological Services Southern Science Center acquired 1:65,000 color infrared (CIR) aerial photography for the entire state. CIR aerial photography is being used to augment visual interpretation of the TM data. Contact prints were made to aid survey team members in the field and in the office. The CIR aerial photography has been scanned and stored on CD-ROM, and these CIR images will be made available on the Web.
Field survey teams collected ground-truthing data in several areas of the state to help the image analyst classify the clustered data sets. The locations were picked to provide best coverage of the different vegetation types within the state. Dozens of points were collected throughout the survey areas, most of which were near roads due to access restrictions for private lands.
Approximately the lower third of Louisiana is covered with wetlands, and the National Wetlands Inventory (NWI) provides high-quality, detailed information for these complex wetlands. The data were delineated using 1988 NASA 1:62,500 CIR aerial photography which is also used in classifying the TM imagery. The NWI data sets were segmented to reflect the Louisiana GAP vegetation classification. Aerial photography and other imagery, such as merged TM/SPOT data, will be used along with ground point data for accuracy assessment.
Steve Hartley, Pat ONeil,
and Jimmy Johnston
GAP and Local Government Planning
In the last year, the recognition of the critical role local government must play in biodiversity conservation has spread rapidly throughout the scientific, planning, and conservation communities. At the 1997 Annual GAP Meeting in Reston, USGS Chief Biologist Denny Fenn echoed this need through a quote from In Our Own Hands (Jensen et al. 1993): "County and local governments have the greatest effect on natural resource management in the U.S., yet they have the least access to good science. They also have the least access to biological information that covers large areas which often stretch beyond their jurisdictions."
The role of local government stems from the fact that approximately 80 percent of the land area of the coterminous U.S. is privately owned, and local governments have nearly exclusive jurisdiction over those lands. That jurisdiction influences habitat conservation or destruction most directly through land use planning and zoning. More importantly than area alone, however, privately owned land contains the lower elevation, highly productive, and most threatened ecosystems (repeatedly being confirmed by Gap Analysis projects, e.g., Merrill et al.1996, Davis et al. 1995, Thompson et al. 1997, Edwards et al. 1994).
In response, the application of GAP information to county land use planning is being explored in four county pilot projects. A biological decision support system is being developed for Teton County, WY, by Patrick Crist and Tom Kohley that will allow planners to easily consider impacts on biological resources when reviewing development proposals. In Santa Cruz County, CA, Chris Cogan is using GAP data to aid the county in determining those elements requiring conservation planning in the context of the ecoregion. Daryl Durham is leading a three-county effort in Tennessee to integrate the results of GAP in assessing development projects and comprehensive planning. In Washington state, a two-county planning assistance project is working with the counties to incorporate biodiversity into their open space planning programs. These projects each take a somewhat different approach to spatial and temporal scales and planner needs, whether it is data and tools, planning assistance, or education; but they are all contributing to the "tool box" of methods to make biodiversity a routine consideration of local government planners.
National GAP is solicited frequently by planning-related publications for tools, methods, and case studies that apply our data to local government needs. We plan to form a national repository and working group for such work and welcome all input and interest. If you would like more information or would like to be on the working group, contact Patrick Crist.
Davis, F.W., P.A. Stine, D.M. Stoms, M.I. Borchert, and A.D. Hollander. 1995. Gap Analysis of the actual vegetation of California: 1. The Southwestern region. Madroņo 42(1):40-78.
Edwards, T.C., Jr., C.G. Homer, S.D. Bassett, A. Falconer, R.D. Ramsey, and D.W. Wight. 1995. Utah Gap Analysis: An environmental information system. Final project report 95-1, Utah Cooperative Fish and Wildlife Research Unit, Utah State University, Logan.
Jensen, D.B, M.S. Torn, and J. Harte. 1993. In our own hands. University of California Press, Berkeley, California.
Merrill, E.H., T.W. Kohley, M.E. Herdendorf, W.A. Reiners, K.L. Driese, R.W. Mars, and S.H. Anderson. 1996. The Wyoming Gap Analysis Project final report. Wyoming Cooperative Fish and Wildlife Research Unit, University of Wyoming, Laramie, Wyoming.
Thompson, B.C., P.J. Crist, J.S. Prior-Magee, R.A. Deitner, D.L. Garber, and M.A. Hughes. 1997. Gap analysis of biological diversity in New Mexico using Geographic Information Systems. New Mexico Cooperative Fish and Wildlife Research Unit, Las Cruces, New Mexico.
Status Report on MRLC Activities for 1997
The goal of the MRLC is to provide its partners and other federal users with high-quality remotely sensed data and characterized land cover data for the United States in an on-line database (MRLC National Land Characteristics Database). These data will be accessible to federal, state, and private groups for comprehensive and multiscale natural resource research, planning, and management. The first TM-based national land cover data set, circa 1992, is being generated on a federal region basis and will be completed by December 1999.
Current Status of the MRLC TM Archive:
Descriptive data about the MRLC TM imagery are available through the Global Land Information System (GLIS) which can be accessed through the MRLC home page (http://www.epa.gov/mrlc).
MRLC Data Clearinghouse
A MRLC on-line data clearinghouse is being designed at the EROS Data Center. The objective is to provide users with the ability to either order or download TM-based derivative products from the MRLC partners, including: the regional land cover data sets, GAP state or regional natural vegetation data sets, and C-CAP land cover and change detection data sets. A draft Data Dissemination Plan has been produced, and we expect construction activities to begin in the fall of 1997.
For further information contact Nick Van Driel, EROS Data Center, at (605)-594-6007, e-mail: email@example.com.
Building a Wildlife Habitat Relationships Database with FileMaker Pro
For states where information on habitat relationships has not been compiled for some species, constructing a database is one of the larger tasks required by GAP. However, a user-friendly software package can make it much more manageable. At MT-GAP, we are using FileMaker Pro, a database developed by Claris for both Mac and PC environments.
Disclaimer: Although it may seem like one, this is not an infomercial. Several other packages undoubtedly offer similar advantages. However, FileMaker Pro has a number of features that can simplify the task of building a wildlife habitat relationships database:
More information on FileMaker Pro can be found at the following sites.
Links to CGI software and FileMaker Pro databases on the Web: http://www.claris.com/support/products/filemakerpro/docs/cgi.html
Alliance-Level Classifications for Southeast and Midwest
Two new reports on regional vegetation alliances were produced by The Nature Conservancy (TNC) for the National Gap Analysis Program (GAP) in May 1997. An Alliance Level Classification of the Vegetation of the Southeastern United States by Alan Weakley and others describes each alliance that is found in Alabama, Arkansas, Florida, Georgia, Kentucky, Louisiana, Mississippi, North Carolina, Oklahoma, South Carolina, Tennessee, Texas, and Virginia. An Alliance Level Classification of the Vegetation of the Midwestern United States by Jim Drake and Don Faber-Langendoen includes alliances found in Illinois, Indiana, Iowa, Kansas, Michigan, Minnesota, Missouri, Nebraska, North Dakota, Ohio, South Dakota, and Wisconsin. Digital copies of these reports have been distributed to all GAP projects in these states. For additional copies, contact Elisabeth Brackney at 208/885-3560 or firstname.lastname@example.org.
The national vegetation classification developed by TNC classifies existing vegetation hierarchically, with physiognomic units at the higher levels and floristic units at the lowest levels of the classification system. Descriptions of the lower, floristic levels had not been available until GAP provided support to TNC for this work. The first region to complete alliance descriptions was the East (Sneddon et al. 1994). A revision of Bourgeron and Engelkings (1994) Preliminary Vegetation Classification of the Western United States is currently under way. (See the article on vegetation classification in this Bulletin for more details and literature cited.)
Predicting Plant Species Distributions in Wyoming: A GAP Pilot Project
The initial gap analysis has been completed for Wyoming. That project involved the development of coverages for land cover, land management status, and distributional modeling of terrestrial vertebrates (Merrill et al. 1996). Now, a new initiative will explore potential distributions of sensitive plant species and state and regional endemics.
During the past 20 years, the Rocky Mountain Herbarium has systematically inventoried most of the state (200,000 new collections). This, combined with 100,000 specimens from the preceding 80 years, provides an extensive database on location and ecological parameters for more than 3,000 taxa. Likewise, the Wyoming Natural Diversity Database (WYNDD) contains a wealth of information on sensitive plants. The first phase of the project is to capture relevant data on about 1,000 target taxa (20,000 earlier collections, 30,000 obtained since 1977).
The second phase will be modeling the biogeographic properties of these taxa, based in part on the herbarium and WYNDD databases. This will be done by Walter Fertig (WYNDD employee and Botany Department Ph.D. student) in collaboration with Reiners. The modeling fundamentally will be correlative; the potential locations of a species will be based on similarity with environments of locations in which they are known to occur.
The third phase is to perform "gap analysis" of sensitive plants in terms of protection status of common environments of clusters of species. Using modeled distributions of plant species in aggregate and in groups clustered according to similar site requirements, we will compare these distributions with the land status coverage available from the original Wyoming GAP Project, similarly to the models used for terrestrial vertebrates.
Merrill, E.H., T.W. Kohley, M.E. Herdendorf, W.A. Reiners, K.L. Driese, R.W. Marrs, S.H. Anderson. 1996. The Wyoming Gap Analysis Project Final Report. USGS/BRD Wyoming Cooperative Fish and Wildlife Research Unit, Laramie.
Ronald L. Hartman and William A.
Use of Low-Cost, Commercial GPS for Field Data Collection in Nebraska
This article is not intended as endorsement of a specific product but portrays our attempt to find low-cost global positioning system (GPS) support for our field mapping efforts.
This summer, we had teams of graduate students collecting field data in support of the Nebraska Gap Analysis land cover mapping effort. We used GPS technology to obtain accurate locations in the field. We anticipated having more than one team out at a time, but only had one GPS rover unit. Therefore, we needed a low-cost second unit that could be used as needed for a second field team. For this purpose, we selected Trimbles ScoutMaster Flash, their low-end GPS rover designed primarily with the backpacker and recreational user in mind. The unit is under $500, but has some very nice features, including the ability to store, download and upload points, and the ability to navigate to pre-selected points.
For our purposes, the most important feature was the ability to average readings at each point. In the setup, the user can specify the logging rate and the number of points to be averaged at each site. While the output from the ScoutMaster is not differentially correctable, by averaging numerous records at each site we gained real-time readings that were accurate enough to meet our project objectives. Our accuracy was best when we averaged at least 100 readings per site (we were taking a reading every 5 seconds, so total time for the average was about 10 minutes). Increasing the number of readings to 150 and to 300 did not seem to increase our accuracy, but cutting back from 100 to 50 drastically reduced our accuracy. With at least 100 records our average error was about 26 meters in the easting and 18 meters in the northing.
While our sampling was not extensive, and the error for any one site was as much as 40 meters, we feel confident that the unit is giving us field locations that are within about 1 TM pixel in any direction. The low cost of the unit and the time saved in not needing to download base station files and conduct postprocessing have made this unit a very economical field tool for our project. To date, we have used the unit extensively in our field reconnaissance efforts. As we move into assessing our land cover map accuracy, we will need to reevaluate the level of positional accuracy required and the ScoutMasters ability to provide it. Clearly, it will not replace our primary rover unit, but it has been a valuable supplement to it.
Marlen D. Eve, James W.