A DIGITAL ATLAS OF THE VASCULAR PLANTS OF UTAH: DEVELOPING A GIS DATABASE FROM HERBARIA RECORDS.

R. Douglas Ramsey
Thadeus Tilton
Bonnie B. Banner
Allan Falconer

Department of Geography and Earth Resources, Utah State University, Logan, Utah 84322-5240. voice (801) 797-3783, fax (801) 797-4048, Email - doug@nr.usu.edu

ABSTRACT

Collection locations of over 400,000 vascular plant specimens of 2,438 species were digitized from the Atlas of the Vascular Plants of Utah (Albee, Shultz, and Goodrich, 1988). Source maps consisted of 1:6,000,000 approximate scale Utah shaded relief maps with points representing collection locations by species. Location points were transposed onto these maps from the herbaria records of three major universities and the Forest Service. These source maps were hand digitized into an ArcInfo database to reproduce the Atlas in digital form.

INTRODUCTION

In 1988 Albee et al. published the "Atlas of the Vascular Plants of Utah". This work is a compilation of herbaria collection locations of 2,438 species of vascular plants in Utah. This database represents collections from herbaria located at three major research universities, the Bureau of Land Management, the U.S. Park Service, and the U.S. Forest Service. The Atlas required seven years to compile with each specimen voucher visually checked and mapped by the authors and cooperators in the various herbariums.

This database, in the original hard-copy map form, represents an important body of work depicting the distribution of vegetation species throughout the state. The Atlas is used by ecologists, evolutionary botanists, morphologists, physiologists, reproductive biologists, etc. to understand the distribution of vegetation along latitudinal, elevational, and ecoregional boundaries. As a GIS database the Atlas can be used in conjunction with other small-scale (large area) ecological datasets to provide biogeographical information, potential range distribution, and sampling adequacy.

While the Atlas is an important piece of work and voluminous in it's coverage of vegetation distribution. Like all geographic datasets, it makes certain asumptions and has limitations that preclude certain types of analysis. This paper will detail the process of converting the Atlas to digital form and also provide an understanding of the limitations of using small scale databases.

THE ATLAS OF THE VASCULAR PLANTS OF UTAH

The Atlas is a compilation of collection locations of voucher specimens generated over a 30 year period by a host of individuals in a variety of private, state, and federal agencies. The Authors critically examined approximately 400,000 specimens and mapped the sample location of over 73,000 vouchers. Where the same species was collected in approximately the same area, only one location was mapped. Further, where the location of a specific voucher was in question it was not mapped.

The authors used 1:6,150,000 shaded relief maps of Utah to locate vouchers. This represents a limitation of the atlas that will be examined in the following text. Figure 1 is a representative map of the distribution of Sego Lily (Calochortus nuttallii T. & G.), the state flower of Utah. Species distribution maps are organized alphabetically by species, genus, and family respectively. Accompanying each map is the latin name, authority, common name, brief habitat description and elevational ranges.

The purpose of publishing the atlas was not only to document those areas that have been sampled for individual species but more importantly identify sampling gaps and direct future activities to those areas with little or no sampling.

METHODOLOGY

The atlas was digitized over the span of 1 year by three student technicians. To reduce digitizing variation between technicians, a menu interface was created using ArcInfotm menu and AML tools. Reference tics were positioned at each corner of the state and each distribution map was referenced to the same state GIS layer. All technicians worked together to maintain consistency with the database and were each instructed to digitize the center of the mapped sampling point. Each species distribution map was individualy digitized and stored in it's own coverage. Attribute information consisted of genus, species, low elevation, and high elevation for each species. All coverages are named according to standard genus/species acronyms and stored in individual family workspaces.

To spatially evaluate sampling, and mapping bias, all points for each species were combined into family-wide databases and then into a collection-wide coverage. Family-wide databases contain all digitized points for each species within that family. The collection-wide coverage contains all points digitized for the entire atlas. Each point maintains it's original attribution in all coverages.

RESULTS

Figure 2 shows the distribution of sampling points for the collection-wide coverage. A cursory examination of the distribution of points in both figures shows that for the individual species (Figure 2) there is no obvious bias in the distribution and mapping of individual points. While some species are naturaly restricted in geographic range, and there may be a field sampling bias, there is no apparent mapping bias when one examines individual distribution maps. However, when the collection-wide map is displayed there is a decided mapping bias. By simultaneously displaying all 73,000 points one can easily discern political boundaries of individual counties in Utah. There was an apparent subcontious bias on the part of all three authors (Albee, Shultz, and Goodrich) to avoid placing sample locations on top of the county boundaries. The original shaded relief map used to located voucher specimens also contained county boundaries, water bodies and major water courses to aid in locating sample points. Personal conversations with the original authors indicate that there was no discussion relative to the cartographic positioning of the points. While samples collected along county boundaries are placed within the correct documented county, the lack of samples directly over a boundary shows a consistent bias.

Another limitation to the atlas is the size of the original points used to show sample locations. The dot size is approximately 10.7 km in diameter on the ground, effectively setting the minimum resolution of approximately 91 km2. This, coupled with the unknown error of positioning the points makes the atlas a general representation of sample location which was it's intended purpose.

Digital maps are being made available for each species over the world- wide-web. The universal resource location address is http://www.nr.usu.edu. The database is located in the Department of Geography and Earth Resources home page under Utah Geography.

DISCUSSION

The Atlas of the Vascular Plants of Utah is one of only two publications of it's type in the nation. The material presented within it is inteded to provide taxonomists with a guide to where samples of individual species have not been done, thus guiding future collection efforts. However, it fails to show areas of the state that have had little or no sampling as a whole. It is difficult to evaluate overall sample density by examining 2,438 individual distribution maps. By placing the atlas into a GIS we are able to identify those areas lacking in overal sample density.

Distribution maps in the atlas are intended to show where samples have been collected, not the potential distribution of individual species. Therefore, the lack of samples of a particular species in one part of the state with appropriate habitat does not indicate that the plant does not grow there, simply that it was never collected there. However, since these data were compiled from over 400,000 voucher specimens collected by a host of private, state, and federal agencies over the span of 25-30 years makes the atlas one of the most complete sources of habitat information for individual species available. Biogeographical analysis of species distribution can be carried out with a certain level of confidence. Such analysis however, should be limited to the scale of the information and not extrapolated to finer levels of resolution.

The digital form of the atlas can be used to evaluate species distribution and limits within and between available ecoregions delineations. Species richness between and within ecoregions and subregions can be carried out to help evaluate biodiversity.

BIBLIOGRAPHY

Albee, B. J., L. M. Shultz, and S. Goodrich. 1988. Atlas of the Vascular Plants of Utah. The Utah Museum of Natural History, Salt Lake City. Occasional Publication No. 7. 670pp.