Development and Application of a Decision Support Tool to Analyze Alternatives for Landscapes Composed of Multiple Ownerships

Here's a link to the PDF version of this Thesis

by:

Jeffrey Michael Comnick

A thesis submitted in partial fulfillment of the requirements for the degree of

Master of Science

University of Washington

2002

Program Authorized to Offer Degree:
College of Forest Resource

Table of Contents

List of Figures

Introduction

Literature Review: Spatial and Temporal Complexity in Forest Ecosystems
          The Ecosystem Concept
          Stand Development

Literature Review: Ecosystem Management in a Single Ownership
          Landscape Management
          Rational-Iterative Decision Making Process
          Technical Tools: GIS, LMS, and Toggle
          Technical Tools: Other Computer Programs

Literature Review: Ecosystem Management Across Multiple Ownerships
          Examples of Comanagement
          Legal Barriers to Comanagement
          Key Components of Successful Collaboration

Data and Methods: Study Area Description
          Location
          Site Description
          Applicable State and Federal Laws

Data and Methods: Modification of the Toggle Program

Data and Methods: Modeling the Study Area in Toggle
          Scoping the Study Area
          Identifying Objectives and Defining Measurable Criteria
          Developing Alternatives
          Comparing Objectives and Alternatives

Results: Outputs for Each Objective for BLM and Landscape Alternatives
          BLM Only: No Action
          BLM Only: Alternative E
          Landscape Alternative 1: No Action (BLM and Private)
          Landscape Alternative 2: Alternative E (BLM) and Alternative 2 (Private)
          Landscape Alternative 3: Alternative E (BLM) and Alternative 3 (Private)

Results: Discussion on Whether the Toggle Program Achieved the Design Goals

Discussion: Discussion of BLM and Landscape Alternatives

Discussion: Additional Planning and Analysis for the BLM
          Developing Additional Alternatives
          Performing a Spatial Analysis
          Increasing and Adjusting the Planning Area
          Implementation, Monitoring, and Adjustment

Discussion: Other Applications for the Model
          Including Additional Spatial Information in an Analysis
          Multiple Ownership Collaboration

Discussion: Preliminary Evaluation of the Multiple-Toggle Decision Support Tool
          Strengths
          Weaknesses

Conclusions

Bibliography

Appendix A: Figures for BLM, Private, and Landscape Alternatives

Appendix B: Information Concerning Modeling of BLM, Private, and Landscape Alternatives in LMS and Toggle

List of Figures

Introduction

Forest management has evolved through the 20^th Century from focusing on commodity production on a stand-by-stand basis to meeting many objectives, including sustaining ecosystem structures, functions and processes. The new paradigm, ecosystem management or landscape management, broadens the scope of planning and analysis spatially and temporally. Common analysis units now include watersheds and landscapes in addition to individual stands.

Despite a range of interpretations, ecosystem management typically requires consideration of a wider range of social and environmental objectives and greater understanding and application of ecological and silvicultural knowledge than previously. Decision support has been identified as a critical component to ensure values, available information, and current scientific knowledge are included in the ecosystem management decision-making process (Oliver and Twery, 1999). Existing and new technologies are necessary to facilitate analysis and provide decision support for landscape planning. These technologies include computer applications such as geographic information systems (GIS) and the Landscape Management System (LMS) (McCarter et al., 1998).

A difficulty of planning at the landscape or watershed scale is that frequently the area is divided among multiple ownerships. Sample (1992) stated that:

There are few areas of the U.S. in which the delineation of these ecosystems at an ecologically-significant scale does not encompass a mixture of both public and private lands, often in an intermingled pattern inconsistent with ecological boundaries. This suggests the need for a higher level of coordination and cooperation among adjacent public and private landowners in the planning and management of forest and range lands for the protection of biological diversity, water quality, and other ecosystem values.

This study will demonstrate the use of GIS, LMS, and a prototype decision support tool named Toggle for development and analysis of a management plan by one owner in context of estimated management activities by a neighboring owner. Multiple objectives will be analyzed at both the single-ownership level and the multiple-ownership level. The Toggle program will be demonstrated in the context of decision support for landscape management and for use in the rational-iterative decision making process. Results for the study site, a landscape in Oregon comprised of United States Department of Interior Bureau of Land Management (BLM) and private industry lands, will also be discussed.

Literature Review

Spatial and Temporal Complexity in Forest Ecosystems

The Ecosystem Concept

Developing and implementing management plans for a forested area is difficult because of the complexity of the natural world. This complexity is defined with the concept of the ecosystem. Tansley first used the term in 1935, including both organisms and the surrounding physical factors (Tansley, 1935). Odum later defined the ecosystem as "any unit that includes all of the organisms (i.e. the "community") in a given area interacting with the physical environment so that a flow of energy leads to clearly defined trophic structure, biotic diversity, and material cycles (i.e. exhange of materials between living and nonliving parts) within the system (Odum, 1971).

The ecosystem concept indicates a systems approach to dealing with the complex natural world. This approach reduces complexity by grouping common entities and dealing with the interactions between groups (Oliver et al., 1992). It also allows the ecosystem to be delineated in many ways by organizing groups differently (Kimmins, 1987). Tansley and Odum implied this variability by not indicating a specific spatial scale in their definitions. Later definitions explicitly identified the hierarchical nature of ecosystems (O'Neill et al., 1986). The appropriate delineation depends on the issue being addressed and can include an individual organism, a landscape, or a larger region.

Stand Development

Complexity is further increased by the temporal dimension of forest ecosystems. Trees and other vegetation grow through time, and stands and landscapes experience competition related mortality and other small and large scale disturbances. These include fire, wind, insects, and disease. Oliver and Larson (1996) identified four stand development stages: stand initiation (open), stem exclusion (dense), understory reinitiation, and old growth (complex).

The open stage exists from the time of a stand replacing disturbance, during regeneration, until further regeneration is excluded by competition from established trees. Stem exclusion then occurs while trees grow and compete, until competition related mortality and other disturbances create openings in the previously dense main canopy. Increased growing space allows trees, shrubs and other herbaceous plants in the understory to become established during the understory reinitiation stage. Finally, large overstory trees die and multiple canopy layers develop during the old growth stage, creating an uneven-aged structure (Oliver and Larson, 1996).

The names for these processes also apply to corresponding stand structures, which are the physical arrangement of trees and other vegetation. At any point in time, a forest and landscape are comprised of various amounts of one or several stand structures. These proportions vary through time (Oliver, 1992).

Ecosystem Management in a Single Ownership

Landscape Management

A common objective of ecosystem management is maintaining ecosystem functions and processes (Vogt et al., 1997). This can include maintaining biodiversity and forest health, such as maintaining stands resistant to local disturbance agents (fire, wind, insects, and disease). Other management objectives include optimizing harvest volume, revenue, recreation, aesthetics, and many more.

Many ecosystem processes and functions occur at broad spatial scales. Examples include biodiversity and the home range of some species (Oliver et al., 1992). Common planning units in forest management now include landscapes and watersheds. The watershed has been identified as a necessary planning level by the Federal Ecosystem Management Assessment Team (FEMAT) and the Washington State Department of Natural Resources (Sessions et al., 1997).

To achieve the various objectives, a landscape approach has been proposed that maintains a mix of all stand structures across the landscape at all times (Oliver, 1992; Hunter, 1990). This approach utilizes systems theory by identifying five hierarchical management levels, determining appropriate objectives to be achieved at each level, and coordinating activities between each level. The five levels of forest management, from the most "specific" to the most diffuse, are: silvicultural operations, silvicultural regimes (silvicultural pathways), landscape patterns, forest plans, and broader policy (Oliver et al., 1999). Salwasser (1991) identified the necessity to conduct ecosystem management on similar scales (Stand/Site, Watershed, Landscape, and Region), and similar corresponding management levels (Project, District, Forest, Regional, and National/Congressional) have been identified in the Forest Service (Carwse, 1994 as cited in Hobbs, 1998).

Silvicultural operations are the most specific level. These include harvesting and planting trees, pruning, fertilizing, road construction, and others. Decisions requiring the most site specific information must be made at this level. For some objectives, operations can be designed to mimic natural disturbances.

Silvicultural regimes are the stand level treatments applied through time to achieve desired objectives, including future stand characteristics. Each stand structure provides a unique set of values, including habitats, aesthetics, recreation, and resistance to various disturbance agents (Oliver et al., 1992). Knowledge of stand dynamics is important to understand how trees and other vegetation will respond to treatments in order to achieve desired future conditions. "Windows", or periods of time when stands respond effectively to treatments, should be identified for each stand (Oliver and Larson, 1996). These windows of opportunity exist for many operations, including planting and thinning (Oliver et al., 1999).

The landscape level coordinates silvicultural treatments to all stands at all times to ensure a desired mix of stand structures. Through a "coarse filter" approach, managing at this level ensures habitat will be provided for most species, since most habitats can be associated with one or more stand structures (Oliver, 1992). Spatial arrangement of structures may also be important for reasons such as habitat, corridors, and operation implementation. Planning at the landscape level can ensure that objectives with spatial criteria are achieved and spatially feasible plans are developed (Oliver et al., 1999).

The next hierarchical level, the forest, manages the flows of outputs from various landscapes. It is not concerned with spatial feasibility. This level coordinates landscapes and utilizes economies of scale to manage implementation costs and marketing of commodity products and other outputs (Oliver et al., 1999).

Finally, the policy level coordinates the flow of values within a uniform political unit. This can include companies or state and federal governments. Instruments at this level include incentives and regulations. Incentives can be tax reductions, market incentives, education, research, and monetary grants. Regulations can be laws and other procedural rules (Lippke and Oliver, 1993).

Through coordination between the hierarchical levels, management objectives are achieved. Coordination can include analysis, decision-making, and implementation; and information can move from the specific level to the diffuse level as well as from diffuse to specific. For example, final selection of a chosen management plan to implement occurs at more diffuse levels; however, implementation and the effects of implementation occur first at the specific level. Identifying the appropriate level at which to manage for a particular task is important to ensure efficient management of the system (Oliver et al., 1999).

Rational-Iterative Decision Making Process

The systems approach reduces ecosystem complexity and improves management for desired objectives. Additional problems with managing ecosystems include the variability in the spatial and temporal dimensions of ecosystem boundaries and the related variability of associated management objectives defined by those boundaries. Societal weights or values associated with those objectives can also be difficult to determine (Oliver and Twery, 1999). An appropriate decision-making process must be utilized which incorporates this degree of uncertainty.

A variety of decision-making processes exist. The Expert/Intuitive method relies on the judgment of an expert or group of experts, and can result in "groupthink" and decisions based on charisma. The "Muddling Through" approach addresses needs on a case-by-case basis to reduce present conflicts without extensive analysis of values and consequences. With the crisis approach, a manager assumes broader authority than needed to avert a perceived impending catastrophe. The Normative/Rational non-iterative approach is appropriate when all objectives, weights, and interactions between modules are known or uncertainty can be quantified. This approach allows optimization to determine the best solution. Finally, the Normative/Rational iterative method is appropriate when objectives and weights are not well understood, but interactions between modules or uncertainty can be predicted (Oliver and Twery, 1999).

The rational-iterative decision-making process is the most appropriate for ecosystem management. Through this process, multiple alternatives are presented, and alternatives can be refined by working "iteratively" with the decision-maker. This approach allows decision-makers to understand trade-offs and select a chosen alternative through iteration. The steps of the rational-iterative decision-making process are:

1) Identify the decision-makers,

2) Identify the problem, define the objectives, and develop measurable criteria,

3) Develop alternatives,

4) Compare alternatives,

5) Choose an alternative,

6) Implement the chosen alternative,

7) Monitor and evaluate.

The person with authority to make a decision must be identified first. Although many stakeholders may have a very high interest in the selection of a particular alternative, the person with legal authority or jurisdiction must be specified. Next, the problem must be defined, and objectives determined. This step includes scoping the planning area to determine initial conditions, identifying objectives, and converting objectives to measurable criteria. Scoping the planning area is necessary to identify the appropriate spatial scale and to examine initial landscape conditions to determine if perceived problems actually exist (Oliver and Twery, 1999). Management objectives must then be identified and converted to measurable criteria. Measurable criteria convert vague objectives to specific conditions defined numerically to indicate the degree of success each management alternative has in meeting objectives. Measurable criteria in forest management can frequently be defined by stand characteristics, such as number of trees per acre of a given size or percent canopy closure to define stand structures. The spatial and temporal dimensions of the analysis must also be determined. This step may also include development of models to describe vegetation growth, habitat for various species, or models for many other objectives (Oliver and Twery, 1999).

Next, a range of management alternatives must be developed. This is a creative step and can include many role-playing "games" to avoid "groupthink." All interested stakeholders could be allowed to develop an alternative, ensuring a wide range to analyze and compare (Oliver and Twery, 1999). A useful baseline alternative is no action, where the forest vegetation is allowed to grow through time with no silvicultural operations.

Each alternative is then compared to each objective to determine trade-offs. A decision matrix, which lists the impact of each alternative on each objective in a single table, can be useful for the decision-maker. Also, using normalized values, which lists results as a proportional score to a maximum value, can ensure decision-makers are not biased between objectives which commonly have large numbers, such as harvest volume, and objectives which commonly have small numbers (Oliver and Twery, 1999).

Next, the decision-maker must select a chosen alternative to implement. Prior to selecting an alternative, an analyst may need to explain the results, including assumptions embedded in models and measurable criteria, to ensure the decision-maker makes an informed decision. Because the process is iterative, the decision-maker may require additional alternatives to be developed after considering the initial set. The decision-maker selects an alternative based on the trade-offs between objectives, and can ignore one or many objectives at this point. Management objective weights and social values, previously not accounted for during the rational-iterative process, are implied with this step in the selection of a chosen alternative.

The remaining steps are implementation of the chosen alternative and monitoring and evaluation. Implementation includes coordinating the hierarchical management levels to ensure overall policy objectives are achieved. Because planning and implementation are rarely perfect, monitoring and evaluation must be conducted to determine when objectives are not being achieved and why. In forest management, because implementation of an alternative generally occurs over many years, early monitoring can allow alternatives to be adjusted or implementation of later activities to be improved (Oliver and Twery, 1999). Adjustments may include improving models and measurable criteria, increasing scientific understanding, and adjusting scientific and management paradigms.

Technical Tools: GIS, LMS, and Toggle

Analysis of ecosystems over broad spatial scales for many objectives requires large amounts of data and many computations. Projecting landscapes through time and analyzing each objective at periodic increments increases the number of computations further. Advancements in computing power and development of several computer applications have been important in making landscape management practical. These computer applications include geographic information systems (GIS) and the Landscape Management System (LMS).

GIS is a commonly used forest management tool that displays and analyzes spatially referenced information (Star and Estes, 1990). In forestry, this information commonly includes location of stand boundaries, streams, roads, and soils as well as many other features. GIS is capable of performing spatial analyses required for ecosystem management.

LMS is a Microsoft Windows® application that combines existing growth models, including the Forest Vegetation Simulator (FVS), visualization tools, and analysis tools to conduct analyses rapidly for a landscape comprised of a number of forest stands (Stage, 1973). The program organizes tree list inventories for each stand, stand boundaries, and a digital elevation model. Landscapes are analyzed as the aggregate of all stands. Silvicultural operations can be modeled and applied to one or many stands, and treated inventories projected through time. Tables can be produced which analyze any treated and projected inventories (McCarter et al., 1998). Growth models allow LMS to perform temporal analyses required for ecosystem management.

An additional companion program for LMS is Toggle. Toggle is a Microsoft Excel® spreadsheet program that allows users to conduct a multiple objective, landscape level analysis for many time periods. Toggle is a strata-based area allocation model, in which the user manually adjusts the percentage of acres in each stratum (group) subject to a particular silvicultural pathway. Each adjustment affects the outputs provided for each objective. Graphs, summary values, and normalized values for all objectives update immediately as percentages are changed in the program. Groups are typically defined by common significant ecological characteristics, such as dominant species, stand density, and stand age. Silvicultural pathways for each group are then modeled in LMS. Output tables which analyze objectives are obtained from each group from LMS and are input into Toggle for analysis. Toggle programs with the capacity for 6 groups and 15 pathways for each group have been used in studies by Johnson (2001) and Hall (2001).

Analysis in Toggle is based on the concepts that any stand can follow a range of silvicultural pathways leading to different stand structures; each stand structure provides a unique set of outputs and values; and achieving a desired mix of ecological, economic, and social values can be attained by providing some mix of all stand structures across the landscape. To provide a more detailed explanation of the program, the most significant steps for using the spreadsheet will be described. The Toggle spreadsheet organizes 58 separate worksheets, with functions including input data storage, preliminary and summary calculations, graphical and tabular output, storing alternatives, and storing adjustable model values. Microsoft Visual Basic for Applications® code is also used to add functionality. A screen capture of the opening Toggle page is provided in Figure 1. From this sheet, most other Toggle functions can be accessed.

Modeling alternatives in Toggle begins with identifying groups and representative stands for the landscape. For each group, a range of silvicultural pathways are modeled in LMS. These pathways represent different potential management options based on identical initial conditions. Tables that report outputs for all objectives at each point in time are obtained from LMS and pasted into Toggle. An example of a portion of an input table is provided in Figure 2. Each silvicultural pathway to be included in Toggle requires an input table. A screen capture of the input page for one group is provided in Figure 3. In addition to loading Toggle with tables from LMS, the total number of acres in each group must also be inputted (Figure 4).

Certain aspects of Toggle can be adjusted by the user, including threshold values and maximum values. Threshold values are components of measurable criteria that indicate the point of success or failure for achieving an objective. It may be necessary to adjust threshold values depending on local conditions or the performance of a particular growth model. For example, a landscape on more exposed aspects may have a different wind safety threshold value (height/diameter ratio) than a landscape on less exposed aspects. For Toggle to report accurate normalized values, maximum values must also be determined and inputted. Maximum values are necessary to scale current values for an objective to calculate the proportional normalized value. Screen captures for worksheets where threshold values and maximum values can be adjusted are provided in Figure 5 and Figure 6, respectively.

Analysis can then be completed in Toggle. Beginning with group one, the user allocates a percentage of the total group area to follow each potential silvicultural pathway for that group, until 100% of the area has been allocated. As area is allocated to a pathway, output values, previously per acre values, are multiplied by the number of committed acres to calculate total values. These calculations are accomplished automatically by the spreadsheet and immediately when the user adjusts any pathway percentages. The same process of area allocation to silvicultural pathways is completed for each group. To obtain output values for the entire landscape, the spreadsheet sums the outputs for all pathways (after being multiplied by the allocated acreage factor) for all groups. When the user has committed 100% of the area in all groups, the landscape alternative is complete. Figure 7 shows the Toggle output page where percentages allocated to silvicultural pathways can be controlled, and output values are graphed. Figure 8 shows a portion of the output page before and after acres are allocated to illustrate how graphs change immediately in response to new output total values.

After an alternative is developed it can be stored in the program, allowing the user to develop additional alternatives. The spreadsheet stores pathway percentages for each group, and can automatically reconstruct alternatives. Toggle also generates a decision matrix of normalized values from each saved alternative. A screen capture of the worksheet where alternatives can be saved is provided in Figure 9.

Technical Tools: Other Computer Programs

Many other computer applications have been developed to facilitate forest management and ecosystem management. Two programs will be briefly discussed to provide context for LMS and Toggle. These are FORPLAN (Johnson et al., 1986) and SNAP (Sessions and Sessions, 1992). This should not be considered an extensive critique of the tools, or an in-depth comparison between any of the applications. FORPLAN is a forest level, strata-based decision support tool developed for the Forest Service which utilizes linear programming to optimize an objective under given constraints. Criticisms of FORPLAN include: it utilizes the rational non-iterative decision-making process, which is not considered appropriate where objectives and weights are not well known (Oliver and Twery, 1999); both the model and the outputs are extremely complex and difficult to understand (O'Toole, 1983); and model outputs fail to account for spatial criteria or cumulative effects and are thus difficult to implement successfully (Johnson, 1992). SNAP is a GIS based harvest scheduling model. This program can identify near-optimal solutions for location of harvest units based on spatial constraints such as adjacency ("green-up"), maximum opening sizes, minimum habitat levels, and road networks. SNAP is more limited in the spatial scale it can analyze. Also, neither FORPLAN nor SNAP maintains tree-level resolution for each stand, limiting the ability of these programs to analyze additional objectives as necessary.

Ecosystem Management Across Multiple Ownerships

A difficulty of practicing ecosystem management at broad spatial scales is that watersheds and landscapes are frequently divided among multiple ownerships or agencies. Many potential barriers exist for collaboration, including state and federal laws. Where successful collaboration has occurred, key components have been identified.

Examples of Comanagement

Examples of comanagement include the Shelton Cooperative Sustained Yield Unit and the Plum Creek Habitat Conservation Plan. The Shelton Cooperative Sustained Yield Unit (CSYU) was formed in 1946 through an agreement between Simpson Timber Company and the Forest Service. The CSYU is on the Olympic Peninsula of Washington State. The CSYU was intended to be comanaged for sustained yield timber volume to stabilize the local communities of Shelton and McCleary and to ensure the general forest health of the contiguous area (U.S.D.A. Forest Service, 1946).

The Plum Creek Cascades Habitat Conservation Plan (HCP) was developed in 1996 for company land in the central Cascade Mountains in Washington State. The planning area was of the "checkerboard" configuration, with alternating sections of Plum Creek ownership and National Forest. Although the HCP did not establish a cooperative agreement, consideration of the contiguous landscape, rather than only the fragmented company land, was critical for successful application for an incidental take permit (Plum Creek, 1996). An incidental take permit allows a company to conduct operations in areas with endangered species without penalty for incidentally harming or killing an individual of the species. This consideration included assumptions of the activities that would occur on federal land during the 50-year plan, resulting forest conditions, and cumulative impacts when analyzed with projected Plum Creek activities (Plum Creek, 1996).

Finally, an analogous cooperative situation outside forestry may be the Clean Air Act of 1990, where many industries in an area must coordinate to reduce pollution below a certain level. This cooperation includes market functions, such as the selling of excess pollution quotas by companies efficient at pollution reduction to companies which are inefficient (Bryner, 1995). It also includes regulations, in the form of the collective pollution limit.

Legal Barriers to Comanagement

Legal barriers to comanagement include the Federal Advisory Committee Act (FACA) and the Sherman Anti-Trust Act. FACA requires any advisory committee formed by the federal government that includes individuals who are not federal, state, tribal, or local officials to include a balanced membership in terms of points of view represented. Comanagement on landscapes with federal and private land will be subject to FACA. This may result in inappropriate federal control and bureaucratic delay and cost (Meidinger, 1997). Between private companies, the Sherman Anti-Trust Act prohibits cooperation or exchange of information that would result in any form of price fixing (Meidinger, 1997).

Other laws can indirectly influence comanagement as well. Many private landowners may not want to improve wildlife habitat for fear an endangered species would inhabit the ownership. The Endangered Species Act could then severely limit potential activities (Sample, 1995). Many landowners are also hesitant to enter into a comanagement agreement because flexibility in future decision-making may be limited (Sample, 1995). For example, inheritance taxes may eventually require a landowner to harvest timber without regard to ecological values or the effect to a comanaged landscape (Sample, 1995).

Overcoming these laws to conduct comanagement may require clarification of certain points specific to natural resource management. For example, concerning the Sherman Anti-Trust Act, cooperating companies may be prohibited from sharing harvest volume information, but instead could share stand structures and agree to manage for a desired mix. Other laws, such as the Endangered Species Act, may require reauthorization to ensure a consistent national policy for forest ecosystem management (Sample, 1995). Finally, new legal mechanisms also exist which encourage collaboration for ecosystem management, such as conservation easements (Meidinger, 1997). These legal issues are important but beyond the scope of this paper.

Key Components of Successful Collaboration

Disregarding legal deterrents and other disincentives, several important criteria have been identified in successful partnerships. These include need, presence of a catalyst organization, peer-to-peer networking, communication, and trust (Sample, 1995). Because comanagement requires additional effort, a perceived need is critical for successful partnership. Even with a need, Sample (1995) identified a catalyst organization as the most important element for success. The topics covered in this study, including application of specific computer programs within an appropriate decision support system framework, are a component of communication. Sample (1995) also identified that communication includes sharing of experiential, historical, and cultural knowledge as well as technical knowledge.

Figure 1. Toggle main page. When the Toggle spreadsheet is opened, this is the active page. From this page, most program functions can be accessed, including silvicultural pathway data input, group acreage input, maximum value data input, adjusting threshold values, and performing the analysis.

Figure 2. Toggle input table for one silvicultural pathway. Values are provided for all objectives every projection cycle. Outputs are based on calculations performed on standing and cut inventories by LMS. Output values include stand structure, standing, harvest, and growth volumes, species mix, height/diameter ratio, average stand height, trees per acre, harvest volume by species, basal area by species, and others.

Figure 3. Toggle input page for silvicultural pathway data for group one. One input table (see Figure 2) is pasted into the appropriate cells on the page for each silvicultural pathway modeled for the current group. A similar page exists for each group to store pathway input data.

Figure 4. Toggle group acreage input page. The total number of acres in each group must be entered. Analysis in Toggle is based on allocating percentages of groups to silvicultural pathways. These percentages are multiplied by the total acres in that group to calculate the actual number of acres committed to that pathway. All per acre input values can then be multiplied by the number of committed acres to calculate total output values.

Figure 5. Toggle maximum value input page. Maximum values must be input to calculate accurate normalized values. Maximum values can be input for complex structure (acres), standing, growth, and harvest volume (mbf), cash flow (dollars), spotted owl nesting, foraging, and dispersal habitat (acres), and marbled murrelet habitat (acres). Normalized values are calculated by dividing the current value for an objective by the maximum value to determine the proportional score.

Figure 6. Toggle threshold values page. Threshold values for many objectives can be adjusted in Toggle. Threshold values indicate the point of success or failure for an objective. Adjusting these values may be necessary to calibrate the model for local conditions or performance by a particular growth model variant. A sensitivity analysis can also be performed by altering threshold values slightly and restoring a previously modeled alternative to determine the degree of change for Toggle results.

Figure 7. Toggle output page. From this page, the user can allocate group acres to silvicultural pathways, and graphs update immediately to display the new mix of outputs. Each graph is displaying output units (acres or MBF) over time (years). Larger versions of these graphs can be examined in later figures. The list of the silvicultural pathways appear in column A, with the corresponding percentage in column B. Each group utilizes the same space, with only one group active at a time. By clicking on the 'Back' and 'Next' buttons on this page, other groups can be selected as the active group and toggled.

Figure 8. Toggle output page before and after area allocation. This figure demonstrates how outputs change when area is allocated to silvicultural pathways and how alternatives are developed. The top screen capture shows all pathways for all groups set to 0%. The bottom screen capture shows 5% of group 1 allocated to pathway 2. The output graphs for stand structure (left) and harvest volume (right) change immediately. Larger versions of these graphs can be examined in later figures. All graphs not shown also change. As the user allocates more area to this pathway or other pathways, the graphs change in response to the new set of output values.

Figure 9. Toggle matrix page. On this page, alternatives can be saved or restored. Saving alternatives saves the percentages allocated to all silvicultural pathways for all groups, and copies the normalized values for all objectives into a decision matrix containing the normalized values for all other saved alternatives. When an alternative is restored the spreadsheet automatically reallocates the percentages to the appropriate pathways for each group, and outputs are graphed as before.

Methods

Study Area Description

Location

The study area is located in western Oregon, approximately 20 miles southwest of the city of Eugene in Lane County (Figure 10). Most of the study area is located in the Upper Siuslaw watershed, a 5^th field drainage. The ownership pattern is of the "checkerboard" configuration, with alternating sections of primarily BLM and private industry ownership (Figure 11). The study area is located in the Coast Ranges physiographic province (Franklin and Dyrness, 1973).

Site Discription

The Coast Ranges province contains often steep mountain slopes with ridges because of streams and west-flowing rivers (Franklin and Dyrness, 1973). The study area is generally defined by the drainage for the west running Siuslaw River, which runs through the site. Many streams on both the north and south facing slopes run into the river. Site quality within the study area ranges from a high site class 3 to a low site class 2 (DeMoss, June 10, 2002). The region is characterized by a wet, maritime climate (Franklin and Dyrness, 1973). The Upper Siuslaw watershed occurs in the Tsuga heterophylla vegetational area, with common species including Douglas-fir (Pseudotsuga menziesii), western hemlock (Tsuga heterophylla), and western red cedar (Thuja plicata). More specific vegetational characteristics will be discussed in the Methods section.

Applicable State and Federal Laws

Management activities on BLM land are subject to many federal laws. The most significant of these are the Federal Land Policy Management Act (FLPMA), the National Environmental Policy Act (NEPA), and the Northwest Forest Plan. The FLPMA was enacted in 1976, and requires the BLM to manage its land for multiple-use to best meet the needs of present and future generations (U.S.D.I. Bureau of Land Management, 2001). NEPA, enacted in 1969, requires the preparation of an Environmental Impact Statement for any proposed management action that will impact the natural environment (Bass and Herson, 1993). NEPA also encourages ecosystem management by requiring federal agencies to use a systematic, interdisciplinary approach to decision-making (Black and Herrington, 1974).

The Northwest Forest Plan was enacted in 1993. It was developed from the work of the Forest Ecosystem Management Assessment Team (FEMAT), which analyzed the viability of species associated with old growth forests, particularly the northern spotted owl (FEMAT, 1993). This law allocated all federal land within the range of the spotted owl into one of seven designations. These are Congressionally Reserved Areas, Administratively Withdrawn Areas, Late-Successional Reserves (LSR), Riparian Reserves, Adaptive Management Areas, Managed Late-Successional Areas, and Matrix lands. Generally, Congressionally Reserved Areas and Administratively Withdrawn Areas are permanent reserves. Adaptive Management Areas are designed to test a landscape management approach providing social, economic, and ecological values. Late-Successional Reserves and Managed Late-Successional Areas allow some silvicultural activities to produce and maintain forests with complex structures. The remaining land is designated as Matrix, and allows more typical forest management activities, including thinnings and harvesting for timber production (Tuchmann et al., 1996). The BLM land comprising the study area is in the LSR category.

The private land is owned by several industrial timber companies. Management on these lands is subject to laws including the Endangered Species Act, the Oregon State Forest Practices Act (State of Oregon, 1998), and company policy. These lands are managed for commodity production primarily through intensive silviculture.

Modification of the Toggle Program

To complete the multiple ownership and cumulative effects analysis, the Toggle tool was modified and functionality was expanded. Generally, flexibility was increased in all existing aspects of the model, and organization of the calculations and spreadsheets was improved to allow construction of a larger model with a smaller file size. An additional component was also added to sum outputs from individual Toggle programs and display aggregate results for each objective as graphs, summary values, and normalized values.

Existing functionality expansions included increasing the number of groups and pathways to allow development of a more complex model. Potential groups were increased from six to 20. Potential pathways were increased from 15 to 50.

More significant for this study, the Toggle program was redesigned