Gap Analysis in Riverine Environments
Scott P. Sowa
Missouri Resource Assessment Partnership, Columbia, Missouri
Background
As Jennings (1997) mentioned in Bulletin No. 6, the National Gap Analysis Program is in the initial stages of developing the aquatic component of Gap Analysis. This effort, to date, has included drafting general technical documents to guide the development of pilot projects as well as establishing two pilot projects. These include a project for the upper Allegheny River basin in western New York, initiated in 1995, and a statewide project for Missouri, initiated in 1997. The purpose of this article is to outline the basic approach developed by the Missouri Resource Assessment Partnership (MoRAP).
The project addresses several objectives; however, the three primary ones are: 1) develop an objective method for identifying gaps in biodiversity conservation in riverine environments, and set priorities for filling those gaps, 2) identify problems of and methods for effectively integrating the terrestrial and aquatic components of Gap Analysis, and 3) document information needs, successes, failures, obstacles, time, and costs, which will assist other states with similar efforts.
We established some priorities for our project to make it more reasonable in scope and to help us maintain a more structured approach. First, we are strictly focusing on riverine environments. Missouri is essentially a "stream state" with the majority of our aquatic biodiversity concerns situated in riverine environments. Second, although Gap Analysis will continue to include all taxa, this project focuses on fish, mussels, crayfish, and snails. These four taxonomic groups were selected primarily because of the availability and quality of existing sampling data.
Approach
There are five major steps to the approach we developed. The first step involves delineating and mapping a 1:100,000 digital data layer of valley segment types for Missouri. These valley segment types can be viewed as the lotic counterparts of wetland or lake type classifications (Figure 1). To accomplish this step, we are using the Aquatic Community Classification System developed by The Nature Conservancy (Lammert et al. 1996). This hierarchical classification system focuses on ecological regions and hydrogeomorphic variables to delineate distinct valley segment types. Using this digital data layer, we will then generate the critical base line inventory statistics required for conducting accurate assessments and developing meaningful conservation priorities. These statistics include how many valley segment types there are within each ecological section (Bailey 1980), how many miles there are of each type, and where they are (Figure 2).
 Figure 1. Hypothetical example of two very different valley segment types delineated by following The Nature Conservancy’s Aquatic Community Classification System. |
 Figure 2. An example of how base line (i.e., total number of miles) and assessment (e.g., percent in public land) statistics for each valley segment type will be generated for each Ecological Section in the state. These statistics represent the first step toward setting meaningful conservation priorities. |
The second step incorporates digital instream and watershed management and land use information into a coarse-level analysis process (Figure 2). Essentially, the question being addressed in this step is how well are we currently conserving each of Missouri’s valley segment types. Answering this question requires relativistic comparisons using percentage statistics: for example, calculating the percentage of each valley segment type currently in the public trust, the percentage of the total stream miles each valley segment type represents within each region, and the percentage of each valley segment type that can be classified as high-quality. There are numerous other calculations like these that will be incorporated into this step of the process. The key point is that this step needs to be flexible enough to meet a wide variety of user needs yet remain focused on establishing first-cut biodiversity conservation priorities for each region.
As can be seen in steps one and two, we are not using biological information to develop our initial conservation priorities. Our initial reasoning is that all types of the various riverine environments must be considered as having conservation value, regardless of their biological communities. Therefore, we first focus on the rarity and the conservation status of riverine types themselves. Consequently, riverine environments inherently low in diversity are weighted equally with those inherently high in diversity at the outset. This avoids the problem of developing conservation priorities which simply focus conservation efforts on those inherently diverse environments while ignoring very unique environments that may have relatively simple communities. I believe that the number of rare species tends to be strongly associated with community diversity, at least in aquatic environments; however, this relationship is probably scale-dependent. Even conservation assessments that focus on rare, threatened, or endangered species will tend to establish conservation priorities favoring inherently diverse environments rather than unique environments. This may not be true for all regions; for instance, the very simple communities of the Arid Southwest harbor a large number of rare aquatic species. However, this association between rarity and diversity definitely holds true for the majority of the nation as well as those regions covering Missouri (Figure 3).
 Figure 3. Example of the association between diversity and rarity at broad spatial scales for two taxa in Missouri. With such an association, establishing statewide conservation priorities based on either diversity or rarity data would lead to conservation priorities focused in the Ozark Plateau Ecological Section even though the other three Ecological Sections harbor extremely unique communities which in many instances are currently facing greater threats to their persistence. This emphasizes the importance of classifying resources under consideration into meaningful ecological units and establishing initial conservation priorities based on the rarity or conservation status of these units regardless of their biological communities. This figure also points out the importance of taking a hierarchical approach to establishing conservation priorities under which priorities are established within a regional context. |
Once conservation gaps are identified for the valley segment types and initial priorities established, we move on to step three in which distributional data for all known species of fish, mussels, crayfish, and snails are used to predict the community potential of each individual valley segment in the state. To accomplish this task, we need three different pieces of information in digital format (Figure 4). First, we need to know the statewide distribution of each species. More specifically, we need to know all of the watersheds (14-digit Hydrologic Units; USDA 1992) in the state in which a given species exists. Second, we need to know all of the general habitat requirements or affinities of each species so we can predict their distribution throughout the watersheds in which they are known to occur. (Few species are found throughout entire watersheds; most reside in specific segments of watersheds such as headwater segments or only warm-water segments.) Finally, we need the valley segment data layer, which provides the habitat type template for predicting local distributions. This is analogous to the land cover layer for predicting the distribution of terrestrial species. The end product of this exercise will be a 1:100,000 digital data layer of valley segment types for Missouri, with each segment attributed with the fish, mussel, crayfish, and snail species likely to occur in that segment under pristine conditions (Figure 5).
 Figure 4. Diagram showing the three digital information sources required to predict the biological potential of a given valley segment. Since existing land uses are not factored into this modeling effort, the resulting predicted distributions represent the potential aquatic community of a given segment under pristine conditions. |
 Figure 5. An example for the Little Piney Creek watershed showing the difference between the species known to occur in a given Hydrologic Unit and the subset of species predicted to occur within a selected valley segment. |
In step four we use the distributional data developed in step three to revise our initial conservation priorities (Figure 6). Specifically, we use this information to identify specific locations of "high-priority" valley segment types which: a) are relatively high in species richness, b) serve as centers of endemism, or c) harbor species of special concern. For simplicity’s sake, we label these valley segments as "segments of biological significance." Like the second step, step four must also remain flexible to meet a wide variety of user needs. This fourth step is necessary and important because the same valley segment type will be found in many different locations and, due to zoogeographic factors, these different locations will often have different biological assemblages. This fourth step thus serves an important role in further refining management options since even slight differences in species assemblages among locations may have a significant effect on decisions related to biodiversity conservation.
 Figure 6. Example of how biological statistics are used to revise initial conservation ranks for high-priority valley segment types. |
The final step in the overall process involves further refining our conservation priorities by identifying specific valley segments which are both biologically significant and high-quality examples of a particular valley segment type (Figure 7). The resulting final maps will then show the locations at which our biodiversity conservation efforts should be focused and where we can assume we will most likely succeed. To assess the relative quality of each valley segment, we will develop "quality-ranking" models. To accomplish this we are working with resource professionals from around the state to identify the major "stressors" and management activities within each ecological section that either positively or negatively influence our riverine environments. Once identified, we will compile digital data for those major stressors and management activities. We will then develop a protocol, to account for these major stressors and management activities, which examines both the local conditions surrounding a given segment (e.g., is it channelized or lacking a natural riparian corridor?) as well as the condition of its surrounding watershed (e.g., percent forested, road density, potential pollution sources, acres of CRP, etc.).
 Figure 7. Example of how specific locations are identified for selected valley segment types at which conservation efforts should be focused and are most likely to succeed. |
When this five-step process is completed, thousands of miles of stream will have been examined, resulting in a workable number of high-priority stream segments on which to focus conservation efforts. The three most important aspects of the aquatic component of Gap Analysis are: 1) it provides an objective (i.e., data-driven) approach for assessing biodiversity conservation needs, 2) it provides the common framework necessary to make truly relative conservation assessments across states, regions, watersheds, etc., and 3) it has built-in flexibility to account for a wide variety of user needs. Information from the Missouri project will assist state and federal resource agencies in making decisions pertaining to new land acquisitions, new management plans, or for identifying focus areas for land owner incentive programs and research.
For more information, see http://www.cerc.cr.usgs.gov/morap/ or contact Scott Sowa at:
MoRAP
4200 New Haven Road
Columbia, MO 65201
Phone: 573-875-5399 ext. 1715
E-mail: scott_sowa@usgs.gov
Literature Cited
Bailey, R.G. 1980. Description of the ecoregions of the United States. USDA Forest Service, Washington, D.C. Miscellaneous Publication No. 1391.
Jennings, M.D. 1997. In pursuit of the aquatic component of Gap Analysis. Gap Analysis Bulletin 6:35.
Lammert, M., J. Higgins, D. Grossman, and M. Bryer. 1996. A classification framework for freshwater communities. Proceedings of The Nature Conservancy’s Aquatic Community Classification Workshop; New Haven, Missouri; April 9-11, 1996. The Nature Conservancy, Arlington, Virginia. 16 pp.
USDA. 1992. Mapping and digitizing watershed and subwatershed hydrologic unit boundaries. Natural Resources Conservation Service, Missouri State Office, Columbia, Missouri. Prepared under National Instruction No. 170-304, issued July 1992.