Brad A. Holt, Stephen P. Warren

Implementation of Ecosystem Management requires the linking of multi-resource inventory data with detailed spatial data to support ecosystem analysis at a variety of spatial scales. The planning landscape is first quantified through the use of an ecological land classification system to determine a measure of ecological complexity and temporal stage of forest development. A classification tool called the Ecosystem Diversity Matrix is described along with the data management processes to facilitate use of the classification system within the planning landscape of the Idaho Southern Batholith. The framework of the matrix is used to describe historical disturbance regimes, existing landscape conditions required to support biodiversity, and to determine desired future conditions.

Today, there are many different agencies, organizations and private corporations proposing various strategies for implementing and addressing the social, economic and ecological objectives of ecosystem management. Unlike traditional resource management which has focused on individual resource values, ecosystem management focuses on managing multiple resource values across ecological communities and plans for the sustainability of all of its resource values. In other words, it integrates and considers a variety of resources together as opposed to planning for each resource separately. It also extends beyond traditional ownership boundaries and considers the ecological integrity of the landscape.

This fundamental change to the larger context of ecological communities and the environment requires data that describes ecological processes and interactions among components at a variety of spatial scales, from a sub-forest stand scale to the landscape. Currently, there are many efforts to develop resource management tools and collect resource inventories which meet ecosystem management objectives at the landscape scale. This paper describes some of the resource tools that have been developed for integrating forest and ecological inventories into resource and GIS databases and then illustrates how these tools can be used to implement a ecosystem management process in west-central Idaho.

In 1994, Boise Cascade Corporation initiated an ecosystem management project to develop an ecosystem strategy for the 5.8 million acre Idaho Southern Batholith landscape. The Idaho Ecosystem Management Project has implemented an existing ecological land classification system known as the Ecosystem Diversity Matrix (Haufler 1994) to quantify ecological diversity in the Idaho Southern Batholith landscape. As a ecological land classification system, the (EDM) provides the foundation for ecosystem management planning and represents the primary tool for quantifying existing conditions, describing historical conditions, and developing targets for desired future conditions. The EDM also provides an avenue for Boise Cascade to quantify its contribution to ecosystem diversity both today and into the future by incorporating its knowledge of existing forest conditions with potential plant community successional pathways to predict potential outcomes over time.

The purpose of the EDM is to describe the planning landscape in an ecological context. To accomplish this task the matrix is comprised of two axes. The horizontal axis describes ecological complexity through habitat type classes which characterizes combinations of overstory and understory vegetation associations. A habitat type class system is an ecological classification system which describes the biotic potential of the land as expressed through combinations of environmental interactions which determine the vegetation found on any given site (Daubenmire 1968). A habitat type classification system describing potential or climax vegetation has been developed for central Idaho by Steele et al.(1981) and is the basis for describing ecological complexity in horizontal axis of the EDM. The habitat type classes (i.e. columns on the horizontal axis of the EDM) are groups of individual habitat types that have been combined due to similar productivity potential, influence of historical disturbance patterns and ecological functions. This classification of habitat type classes allows an understanding of successional trajectories, provides the predictability of future seral variants of plant communities, an describes the productivity of a given site for different tree species.

In contrast, the vertical axis describes vegetation growth stages or vertical forest structure by characterizing the physical attributes of overstory and understory in a forest stand. More importantly, it describes and quantifies the sequential development of vertical forest structure systematically over time and space from a shrub seedling vegetation growth stage to an old growth or old forest growth stage. The quantification of such forest structure and analysis of the variation found within the forest across the landscape is a measure of overall ecosystem health and is a potential indicator of wildlife habitat suitability.

The intersection and combination of habitat type classes and vegetation growth stages (e.g. an individual cell in the EDM) results in new information classes defined as ecological land units or ELU's (Table 1.). Each ELU describes the existing vegetation for both overstory and understory characteristics, and predicts the ecological processes associated with the forest site such as successional pathways, site productivity, forest health, and habitat suitability. When ELU's are mapped across the entire landscape, the EDM becomes the framework for quantitatively characterizing region ecosystem diversity, it provides a basis for assessment of wildlife habitat quality, it quantifies the contributions of ecosystem diversity by all land owners, and it forces land managers and planners to recognize the dynamic nature inherent in ecosystems.

Table 1. Simplified Ecosystem Diversity Matrix Populated with Acres by Ecological Land Unit.

Once the landscape is classified, ecological land units (ELU's) become the base unit for field inventory sampling. Inventory field staff collect a variety of resource data to describe forest stand attributes associated with each ELU on the landscape to support data requirements for timber and wildlife resource planning models and tactical management decisions. Field data including measures of overstory and understory vegetation, down-woody material, snags, stumps, and horizontal cover are entered directly into hand-held data recorders. Each data recorder is equipped with a Global Positioning System receiver which collects the real-world coordinate of each plot location. The plot data is tagged with the spatial position and down-loaded into a resource data base to store and maintain information on ELU's overtime.

Implementing the ecosystem diversity matrix in GIS requires two essential layers of mapped resource information. First, habitat type classes need to be identified. Typically this has been achieved by using habitat type classification system dichotomous keys and field mapping techniques to identify individual habitat type polygons. Once mapped digitally into GIS, habitat types are then reclassified into habitat type classes and a new GIS habitat type class input layer is created. Secondly, a separate layer of existing vegetative characteristics is needed to supply information on the current vegetation growth stage present. Currently the vegetation growth stage GIS layer is generated from an existing forest stand layer. For each forest stand polygon, a set of forest structure decision rules are applied to the forest inventory information associated with that forest stand and a vegetation growth stage classification is made. When each of these layers have been processed, the habitat type class layer (i.e. the columns in the EDM) is overlaid with vegetation growth stage layer (i.e. the rows in the EDM) within the GIS and the resulting union of map polygons creates the ecological land unit layer (Figure 1). With the creation of a new ELU map layer, the number of acres by ELU can be calculated and used to populate the individual cells of the EDM.

Figure 1. Overlay of Habitat Type Classes with Vegetation Growth Stages to Create Ecological Land Units.

With an Ecosystem Diversity Matrix classification system and GIS ecological land unit (ELU) layer in place, the challenge is to develop the capability to electronically link GIS information to the matrix and visually display the quantity of acres or other significant ecological variables for the entire landscape both presently and into the future. To accomplish this, an Arc Macro Language (AML) program was written to graphically display quantitative values associated with the ecological land unit coverage, in the form of a 3-D bar chart.

The AML program first generates a GIS ArcInfo matrix coverage specifying the number of rows and columns that correspond to the ecosystem diversity matrix (i.e. 37 vegetation growth stages by 12 habitat type classes). INFO database items are added to the generated coverage for the matrix row, column, and relate attributes to create the link to the mapped ELU polygon coverage. Row and column attributes are populated through a series of "DO" loops in ArcEdit and the resulting concatenated Arc Attribute Table (AAT) with row, column, and relate items are ready to be related to ELU coverage. Once this processing is complete, the generated matrix coverage is then transformed to produce the 3-D look for the whole matrix. Two copies of the matrix are transformed to the same locations; one to serve as the matrix base and the other for mapping the vertical bars for the 3-D effect. Next, a frequency file is constructed from the ecological land unit polygon coverage for mapping acres of each ELU onto the transformed matrix coverage. The ELU item value is the first level frequency item and acres are calculated as a second level frequency item.

Once the relate between the frequency INFO file and the transformed matrix coverage are established, the 3-D graphical display of acres for each ELU can be mapped. The mapping of the vertical bars on the matrix coverage occurs within ArcPlot. In ArcPlot, the AML selects all of the arcs of the transformed matrix coverage and steps through the selected set one by one to select the vertice locations for the base of each bar on the chart. It then relates to the ELU frequency file to calculate and store a new set of (x,y) coordinates which are vertically above each of the vertices for each ELU cell to form the bar graphic effect. The height of each vertical bar (i.e. the distance above the four vertices of each ELU cell) is determined by the relative number of acres associated with that ELU cell in the ELU map coverage. Once the relative heights have been established for each ELU cell, the AML then proceeds into a shading routine and maps stored (x,y) coordinates for each ELU. In other words, the routine shades a coordinate polygon in the form of a vertical bar with the height of the bar associated with the number of acres for each ecological land unit in the ELU coverage. Figure 2. illustrates an example of the graphical display of acres for each ecological land unit in the matrix for the entire landscape.

Figure 2. 3-D Graphical Display of Ecosystem Diversity Matrix Linked to Mapped GIS Ecological Land Unit Polygons.

With the mapped ecological land unit GIS layer linked to the 3-D graphical display of the Ecosystem Diversity Matrix (EDM), the matrix can then be used to implement specific steps in an ecosystem management process. In the Idaho Ecosystem Management Project, the staff and project partners are using the EDM as a framework for describing historical disturbance regimes, existing landscape conditions, ecological land unit thresholds overtime and desired future conditions required to support biodiversity. The following discussion describes each of the primary components that the project and cooperators are pursuing in the ecosystem management process (Haufler et al. 1996) and how the EDM is used as a foundational element to support the development of each process component.

Understanding past disturbance regimes and how they operated across a landscape provides essential reference information for ecosystem management. First, it provides information on the ecology of vegetation growth stages described in the Ecosystem Diversity Matrix and describes a rationale for their distribution across the landscape. Secondly, it provides information on the natural history of the area and conditions to which native species were adapted. In the Idaho project, a 1915 timber cruise was used to reconstruct historical stand conditions for low elevation habitat type classes. The results revealed that under past historical disturbance regimes, the forest stands on lower elevation habitats were influenced in structure and species composition by frequent low to moderate intensity understory fires. With this and other historical information, an Ecosystem Diversity Matrix (i.e. Historical EDM) describing the range of historical conditions can be developed for the planning landscape.

Another component of the ecosystem management process is the quantification of existing landscape conditions. The EDM provides a means of quantifying ecological land unit contributions by each landowner in the landscape planning area. Once the habitat type class and vegetation growth stages for the entire landscape have been overlaid to create a ELU layer and the acres of ELU's have been quantified, further analysis of the existing landscape can be conducted. For example, a newly constructed EDM with ELU's provides a framework for understanding, documenting and quantifying both biodiversity and the available habitat components for dependent species across the landscape. In addition, the matrix, in combination with mapped ELU polygons in a GIS, also provides information on the potential spatial distribution of native species habitat in the planning landscape.

Another primary component of the ecosystem management process is the concept of adequate ecological representation. Adequate ecological representation is defined as the sufficient size and distribution of inherent ecosystems to maintain viable populations of all native species dependent on these ecosystems (Haufler 1994). Determination of adequate ecological representation is based on understanding historical disturbance regimes and characteristics of the ecological land units that occurred under these regimes.

Coarse Filter Examination of the Landscape

The development of adequate ecological representation is dependent on a coarse filter examination of the landscape. In this approach, coarse filter refers to the use of an appropriate classification of the landscape, the EDM, to quantify acreage thresholds of ecological land units. In essence, the EDM is the classification tool for performing a coarse filter assessment of the number of ecological land units which are essential for maintaining ecosystem diversity and function across the landscape. The 3-D graphic display along with quantitative acreage measures is an important tool for both developing and evaluating the appropriate acreage thresholds for maintaining ecosystem integrity and viable populations of native species.

Fine Filter Examination of the Landscape

Once the coarse filter examination is complete and an EDM describing adequate ecological representation is established, a fine filter examination must occur. A fine filter examination of the landscape refers to assessing land management decisions based on the needs of individual species or guilds (Roloff 1994). In other words, the adequate ecological representation EDM must be further evaluated with single species or guild assessments to assure that there is sufficient habitat within the established ELU acreage thresholds to maintain viable populations of any native species of concern. The fine filter or species assessment approach examines both the quantity of ELU's as well as the quality or spatial distribution of ELU habitat across the landscape.

In the Idaho Ecosystem Management Project, the fine filter/species assessment process is supported through habitat suitability index models for selected species. Vegetation variables related to species needs are identified and quantified in the habitat model for each ecological land unit in the matrix. The habitat models are programmed into a GIS to incorporate the spatial arrangement of species' requirements. These models are typically developed in ArcInfo Grid and use focal functions or "moving windows" to assess the quality of habitat within a species home range or territory. The resulting GIS output layers represent habitat suitability surfaces that are derived from and reflect the spatial distribution of habitat quality for the existing landscape of ecological land units. These habitat map surfaces of habitat potential (see Figure 3.) are then assessed for the selected wildlife species to determine whether the existing habitat quantity, quality, and spatial distribution is adequate.

Figure 3. Habitat Suitability Index Map Surface for Flammulated Owls Derived from ELU GIS Layer in the Idaho Southern Batholith.

Together, the mapped ecological land units describing adequate ecological representation, the coarse filter 3-D bar chart illustrating the quantity of ecological land units, and the species assessment GIS surfaces can be used to benchmark future ecological objectives. If the coarse-filter evaluation indicates that these baseline conditions are provided over time, it is assumed that habitat conditions for native species are being met. The species habitat assessments are used as checks to assure the proper functioning of the coarse filter thresholds. This coarse and fine filter approach should be made prior to and throughout the forest planning period to ensure that ecological objectives are maintained in future forest plans.

The last component of the ecosystem management process is the determination of desired future conditions. With ecological objectives defined, successful ecosystem management then integrates social and economic objectives for the landscape. Economic considerations include needs of natural resource based economies while social considerations include diverse demands for natural resource based recreation and aesthetics, cultural and archeological values, and other concerns unique to the landscape. Identification and quantification of these values are necessary to determine the desired conditions for the landscape. The culmination of the ecosystem management process is determination of desired future conditions. The EDM is important for understanding the contributions of each landowner at any point in the planning horizon. The adequate ecological representation matrix is used to set the needed threshold requirements for ecological objectives. The existing conditions matrix depicts landowner contributions toward these thresholds and where future contributions may be needed. The Ecosystem Diversity Matrices in concert with economic and social objectives are then used to define desired future conditions. Once desired future conditions have been identified, the management activities needed to produce them overtime need to be identified, implemented, monitored, evaluated, and adjusted to maintain adequate ecological representation while providing the optimum mix of social and economic benefits.

Daubenmire, R. F. 1968. Plant communities: a textbook of plant synecology. Harper and Row, N.Y. 300 pp.

Haufler, J. B., Mehl, C. A., and G. J. Roloff. 1996. Using a coarse-filter approach with species assessment for ecosystem management. Wildlife Society Bulletin 24(2):200-208.

Haufler, J. B. 1994. An ecological framework for forest planning for forest health. J. Sustain. Forestry 2:307-316.

Roloff, G. J. 1994. Using an ecological land classification system and wildlife habitat models in forest planning. PhD. Dissertation. Michigan State University. East Lansing, Michigan. 203 pp.

Steele, R., R. D. Pfister, R. A. Ryker, and J. A. Kittams. 1981. Forest habitat types of central Idaho. USDA Forest Service. Gen. Tech. Rep. INT-114. 138pp.

Brad A. Holt
Manager, Idaho Ecosystem Management Project
Boise Cascade Corporation
P.O. Box 50
Boise, Idaho 83728
(208) 793-2586
(208) 793-2712
email idahoeco@aol.com

Stephen P. Warren
GIS Programmer/Analyst, Idaho Ecosystem Management Project
Boise Cascade Corporation
P.O. Box 50
Boise, Idaho 83728
(208) 793-2480
(208) 793-2712
email idahoeco@aol.com