The Knowledge Transfer Process

A cooperative project between The Nature Conservancy, the USDA Forest Service, and the Department of the Interior led to the creation of the Fire Learning Network in 2002 in an effort to overcome implementation barriers to ecologically appropriate projects that reduce hazardous fuels and restore fire-dependent ecosystems (http://tncfire.org/training_fire.htm). A series of collaborative forums organized by the Network provided a framework and approach for knowledge transfer in the Border Lakes Partnership. The importance of a collaborative learning approach to knowledge transfer is illustrated later in this section by descriptions of each phase of the project. Detailed descriptions include the challenges encountered with each step and how those challenges were addressed. Key steps are summarized in Table 5.2. Although considerable progress has been made during the last 2 years, the first four phases represent a collaborative learning process that we hope will provide the foundation for a project that spans multiple years. The overarching goal is to engage partners in developing a long-term, large-scale vision that they will implement through the identification of strategic, collaborative opportunities.

5.6.1. Phase One: Initiate the Collaborative Learning Process

We initiated the collaborative learning process in the winter of 2003 with a knowledge transfer forum held under the auspices of the Fire Learning Network. Participants included representatives from the USDA Forest Service (Superior National Forest and North Central Research Station), the National Park Service (Voyageurs National Park), Ontario Ministry of Natural Resources (Quetico Provincial Park), the Minnesota Department of Natural Resources, and The Nature

Table 5.2. Timing of the key steps in the knowledge transfer process for the first two years of the Border Lakes Partnership

Knowledge transfer process Time frame Steps

Phase one: initiate the collaborative learning process Phase two: develop a pilot model

Phase three: early review of the pilot model

Phase four: build institutional support

Next steps: increase the scale and impact of the initiative

Winter 2003 Hold the first knowledge transfer forum, with an emphasis on ecological models and qualitative desired future condition.

Spring and Assemble data layers from partner agencies and gather partner summer 2003 input on assumptions for the pilot modeling exercise.

Fall 2003 Hold a second knowledge transfer forum, with an emphasis on developing a spatially explicit desired future condition.

Winter 2004 Present the results of the pilot modeling exercise to each of the four key agency partners for input and refinement.

Spring 2004 Begin building institutional support for expanding the collaborative modeling exercise to the whole landscape, including meetings with agency technical leadership and the development of communication products such as a fact sheet.

Summer 2004 Hold a third knowledge transfer forum, with an emphasis on refining a spatially explicit desired future condition and developing strategies for overcoming challenges to future implementation.

Winter and Assemble data layers for expanding the modeling to spring 2005 encompass the full 2-million-ha landscape.

Winter and Hold a meeting of technical team partners to discuss future spring 2005 uses of scenario modeling, including the selection of a desired future condition.

Fall and winter Hold a fourth knowledge transfer forum, with an emphasis on 2005 selecting a collaborative desired future condition and exam ining options for collaborative, cross-boundary projects.

Conservancy. Prior to the forum, we sought participant input on the forum's goals, the conceptual ecological models to be considered, and a qualitative description of the desired future condition. We also engaged the Minnesota Incident Command System, the lead interagency fire organization in the state. The early approval of the System's Prescribed Fire Working Team was crucial, and they assisted with publicizing the forum. As a result, agency attendance was excellent, even though there was some confusion about the purpose of the forum. Several conceptual ecological models formed the basis for the Border Lakes discussion with the goal of establishing a common language about ecological processes and successional pathways (e.g., Figure 5.3; Brown and White 2002). Although opinions varied on the broad application of the models, the models nonetheless provided an important conceptual starting point for developing a common understanding of natural disturbance regimes in the Border Lakes landscape.

How to develop a long-term (100 years), quantitative set of spatially explicit description of the desired future conditions for the landscape was a more difficult topic for the group to tackle. Although in principle all partners acknowledged the advantages of working collaboratively, they focused initially on programmatic benefits such as securing federal monies for cross-boundary projects. Before committing to collaboration on a common vision, partners requested that The Nature Conservancy facilitate the modeling of several alternative forest management scenarios across the Border Lakes landscape to inform their development of a desired future condition. They further recommended that we start by modeling forest management scenarios for a smaller, pilot project area to test the applicability of this modeling approach for the entire landscape. To move the modeling process forward, a small team of the partners volunteered to identify a suitable landscape for the pilot project, conduct the pilot assessment, and report back to the full Border Lakes Partnership on their progress and the initial results.

The high degree to which partners were engaged in the discussion emerged as the primary achievement during Phase One of the project. However, the discussion tended to leap ahead into strategy development rather than laying the groundwork for establishing a joint vision. Our experience is that this is a common tendency among land managers, particularly for a large landscape with complex ownership patterns. Rather than forcing the visioning process, we documented strategy suggestions throughout Phase One to assist in later phases of the collaboration.

5.6.2. Phase Two: Pilot Modeling Project

Subsequent to the decision to proceed with a modeling approach, the modeling team identified a number of "sideboards." First, they agreed that the pilot landscape should contain a sufficient range of initial conditions (e.g., forest structures, stand compositions, and stand ages) and management objectives to ensure that the model's projections would provide insights into the entire Border Lakes landscape. Second, they agreed that the modeling exercise should project the future of several management scenarios that reflected different sets of assumptions about the management activities undertaken within the pilot project area. The resulting landscape projections could

Highly fire-dependent forest: jack pine - black spruce forest dynamics

Disturbance intervals (years): Stand replacing fire 50-100: Wind 1000-2000 Succession Wind Fire

Disturbance intervals (years): Stand replacing fire 50-100: Wind 1000-2000 Succession Wind Fire

(Redrawn from Brown and While 2002)

Figure 5.3. Example of the conceptual state and transition model used for the knowledge transfer exercise, with natural resource managers as the intended audience. Each box describes a unique vegetation growth stage (numbered 1 through 10), with the stand age determined by time since last disturbance. Arrows show changes in forest age, composition, and structure resulting from succession and canopy-replacing wind and fire disturbances.

(Redrawn from Brown and While 2002)

Figure 5.3. Example of the conceptual state and transition model used for the knowledge transfer exercise, with natural resource managers as the intended audience. Each box describes a unique vegetation growth stage (numbered 1 through 10), with the stand age determined by time since last disturbance. Arrows show changes in forest age, composition, and structure resulting from succession and canopy-replacing wind and fire disturbances.

thus provide a range of outcomes for the Border Lakes partners to use in the development of the desired future condition for the landscape. These management scenarios were to be based on the planning work of the partner management agencies, although the modeling team also felt strongly that the management scenarios should not be constrained by these forest management plans if there was reason to model the use of alternative management tools and techniques.

The modeling team selected the Trout Lake land type association in Minnesota (hereafter referred to as the Trout Lake pilot area; Figure 5.4) as its pilot landscape. In this context, a land type association is a fine-scale ecological map unit within the ecological land classification system (Bailey 1995) developed by the USDA Forest

Knowledge Transfer Process

Figure 5.4. Land ownership within the Trout Lake pilot area. The Superior National Forest and Minnesota Department of Natural Resources have the largest holdings. The "Other" category includes county and private nonindustrial ownerships.

Figure 5.4. Land ownership within the Trout Lake pilot area. The Superior National Forest and Minnesota Department of Natural Resources have the largest holdings. The "Other" category includes county and private nonindustrial ownerships.

Service. This 160 000-ha area is centrally located within the Border Lakes landscape, and is similar to the larger landscape in that most of the area is publicly owned and is managed for multiple objectives ranging from wilderness to intensive timber production. The ownership pattern is fragmented, with a mixture of lands managed by the Minnesota Department of Natural Resources and the Superior National Forest dominating the landscape. The selection of a single element within the ecological classification system also worked well as a basis for the simulation model because it simplified the development of site-specific model parameters.

The team used the LANDIS forest dynamics simulation software (Mladenoff et al. 1996; Mladenoff and He 1999) to carry out the landscape simulations requested by the Border Lakes partners. This tool simulates succession, fire, harvesting, and wind damage over large areas (104 to 106 ha) and long time spans (10 to 1000 years), and is thus well suited for the type of analysis the Border Lakes partners desired. An overview of the model is presented in Table 5.3. The model allowed partners to

Table 5.3. An overview of the attributes and data requirements of the LANDIS software

Parameter

Specifics

Purpose A spatially explicit simulation model of landscape-level forest dynamics

Spatial domain 10 000 to >1 000 000 ha

Temporal domain 100 to >1000 years, in 10-year time steps

Processes simulated Seed dispersal Competition Fire Wind

Harvesting

Biological disturbances such as disease and insects Fuel accumulation and decay Biomass accumulation and decay Required data Maps of forest composition and age

A map of land types

For fire: return intervals and mean, minimum, and maximum fire sizes For wind: return intervals and mean, minimum, and maximum areas of storm damage

For harvesting: minimum harvestable area, species preferences, adjacency rules, reentry periods, and many other relevant parameters Species characteristics Model output Species composition and age

Forest stand maps Fire maps Wind maps Harvesting maps

Disease and insect outbreak maps Fuel maps Biomass maps

Model structure Raster-based, with variable cell size (10 to 100 m)

Within each cell, species cohorts are tracked by age Developers David Mladenoff (University of Wisconsin-Madison)

Hong He (University of Missouri-Columbia)

develop spatially and temporally explicit examinations of the goals of the partner agencies, and of the important interacting processes (timber harvesting, forest succession, and fire) that affect the Border Lakes landscape. Reinforcing the need for a spatial analysis was the recognition that some land management objectives are incompatible when arranged in certain spatial patterns (e.g., two adjacent parcels, one managed for a wilderness objective and the other for intensive timber production), and that a nonspatial assessment could not identify such incompatibilities. However, use of the LANDIS model created new challenges for knowledge transfer. Parameterization and implementation of the modeling scenarios in LANDIS requires expert modeling skills that take months for even a highly trained individual to learn. As such, it was not realistic to expect the Border Lakes partners to take this tool and run their own scenarios. Thus, we were challenged to simplify this complexity sufficiently that partners could remain active participants in the use of this modeling tool and that key stages of the modeling were as transparent as possible. At this stage, a core group of two staff from The Nature Conservancy and the USDA Forest Service's North Central Research Station took responsibility for moving the project forward. The Nature Conservancy focused on cultivating relationships with the primary partners, while the North Central Research Station took the technical lead in data acquisition and modeling.

The management scenarios selected by the modeling team included components of each of three core elements: the management strategy, the role of prescribed fire, and the role of wildfire. The management strategy contrasted the current management plans of the partner agencies with newly proposed (but not yet implemented) management plans. In 2004, the Superior National Forest finalized and implemented its revised forest plan (USDA Forest Service 2004). The model scenarios described here are, however, based on the Draft Forest Plan Revision (i.e., the best information available at the time of the modeling exercise; USDA Forest Service 2004). For clarity, the Draft Forest Plan Revision is referred to henceforth as the "proposed plan," and the now-outdated Superior National Forest Plan of 1986 (USDA Forest Service 1986) is referred to as the "current plan." Within the Trout Lake pilot area, the current and proposed plans differed substantially in their objectives (Figure 5.5a,b).

The second and third elements contrasted the effects of the current policy of fire suppression with alternative strategies in which prescribed fire and managed wildfire were used to achieve management goals. These elements reflected the modeling team's concerns about the 80-year history of fire suppression within the Border Lakes landscape. Of the eight possible combinations of options, three management scenarios were selected for analysis: the current management strategy combined with fire suppression (i.e., the status quo), the proposed management strategy combined with fire suppression (i.e., the strategy likely to be implemented on the ground in the near future), and the proposed management strategy combined with the use of both prescribed fire and managed wildfire. In addition to these three management scenarios, a fourth scenario paired a strategy of no timber harvesting with fire suppression. This scenario served as a control and was added to assess the impact of a

Figure 5.5. (a) Current management objectives within the Trout Lake pilot area. The "General forestry" objective is achieved through the use of a variety of silvi-cultural techniques, but relies most heavily on even-aged management practices, including clearcut harvesting. The "Scenic landscape," "Potential scenic river," and "Semi-primitive motorized recreation" objectives balance timber production with recreational goals, and make greater use of partial harvesting and uneven-aged management practices. No timber harvesting is allowed within the Boundary Waters Canoe Area Wilderness. Fire suppression is permitted to support all management objectives.

Figure 5.5. (a) Current management objectives within the Trout Lake pilot area. The "General forestry" objective is achieved through the use of a variety of silvi-cultural techniques, but relies most heavily on even-aged management practices, including clearcut harvesting. The "Scenic landscape," "Potential scenic river," and "Semi-primitive motorized recreation" objectives balance timber production with recreational goals, and make greater use of partial harvesting and uneven-aged management practices. No timber harvesting is allowed within the Boundary Waters Canoe Area Wilderness. Fire suppression is permitted to support all management objectives.

Figure 5.5. Cont'd (b) Management objectives proposed within the Superior National Forest's draft forest management plan, and the Minnesota Department of Natural Resources draft Border Lakes subsection forest management plan. The "Extended-rotation forestry" objective utilized longer timber harvest rotations than the "General forestry" objective, and also relied less heavily on even-aged management. The "Semi-primitive motorized recreation," Semi-primitive non-motorized recreation," "Scenic landscape," and "Potential scenic river" objectives balance timber production with recreational goals, and make greater use of partial harvesting and uneven-aged management practices. The "General forestry" objectives are the same as described in (a).

Figure 5.5. Cont'd (b) Management objectives proposed within the Superior National Forest's draft forest management plan, and the Minnesota Department of Natural Resources draft Border Lakes subsection forest management plan. The "Extended-rotation forestry" objective utilized longer timber harvest rotations than the "General forestry" objective, and also relied less heavily on even-aged management. The "Semi-primitive motorized recreation," Semi-primitive non-motorized recreation," "Scenic landscape," and "Potential scenic river" objectives balance timber production with recreational goals, and make greater use of partial harvesting and uneven-aged management practices. The "General forestry" objectives are the same as described in (a).

"hands-off" management approach, which provided a useful reference point for discussions between the management partners and external stakeholders concerned with the effects of forest management.

The initial data for the LANDIS modeling project (including forest composition, structure, and age; maps of management objectives; and descriptions of the techniques used to meet the objectives) were provided to the modeling team by the Border Lakes partners. These data are collected and maintained in a digital format by the largest public and private forest management organizations, and thus required relatively little manipulation before they could be fed into the model. Data on harvesting rates and techniques are somewhat more difficult to convert into model-accessible formats, but even this process was relatively easy because the members of the modeling team drew on the expertise within their own agencies to provide the necessary data. The parameterization phase of the modeling effort also proved helpful to the modeling team for a completely different reason: it allowed those unfamiliar with LANDIS to learn more about its structure, assumptions, and outputs. In turn, this knowledge helped the team define the types of data to be produced by the project, and the types of conclusions these data would support.

Once the parameterization was complete and the team had determined the types of outputs the model would be used to generate, the simulations were conducted and the raw data were compiled and presented to the modeling team for evaluation. Although the original purpose of the evaluation was to refine the outputs that would be presented at a meeting of the Border Lakes partners, it also served a second, and unintended, purpose: it allowed the modeling team members to compare the model projections with their own expectations. The model projections were uniformly seen as reasonable and logical outcomes of the four management scenarios, and thus were essentially "validated" by the modeling team. As a result, the team members were willing to promote the Border Lakes Partnership process within their own organizations, to arrange access to key decisionmakers, and to provide funding to continue the project. In retrospect, the opportunity to validate model projections was a critical step in the progress of the Border Lakes Partnership.

Model projections provided insights into the potential effects of forest management within the Trout Lake pilot area, and highlighted the importance of a land-scapewide assessment in measuring these effects. Among the most interesting results from the Trout Lake modeling process are those that place the ability to achieve the goals of the management agencies (as outlined within the agency planning documents) into a broader landscape context. For example, one of the goals contained within the management scenario based on the proposed management plans was an increase in the area occupied by red pine and white pine. This goal was shared by both the Superior National Forest and the Minnesota Department of Natural Resources, and their planning documents show that where they actively manage the landscape (i.e., the areas outside the Boundary Waters Canoe Area Wilderness) through timber harvesting and other tools, this goal is being met (MDNR 2004; USDA Forest Service 2004). However, when the unmanaged portions of the Trout Lake pilot area were also included in the analysis, there was no evidence of an increase in red pine and white pine stands at the landscape scale, and little difference between the current and proposed management plans (Figure 5.6a). At the landscape scale, therefore, the maintenance of existing red pine and white pine stands and the creation of new ones through active management was balanced by the loss of stands due to natural successional processes in unmanaged portions of the landscape. The goal defined for the actively managed portions of the landscape was thus not achieved when considered in the larger landscape context. However, using prescribed burns and wildfire as a management tool could increase the abundance of both pines across the landscape, both when stands of these species are considered and when the abundance of both species across all stand types is considered (Figure 5.6a,b).

Figure 5.7 provides a second example of the importance of the landscape context in developing the desired future condition by depicting a small portion of the pilot area under the proposed management plans of the Minnesota Department of Natural Resources and the Superior National Forest. As the figure demonstrates, management objectives can vary greatly across even a small area, and potentially conflicting management objectives may exist in close proximity. The juxtaposition of wilderness, general timber production, extended-rotation forestry, and recreational objectives within a small area may create challenges for land managers attempting to achieve goals related to wildlife, aesthetics, and recreation, not because of their own management actions, but because of management actions taken in adjacent land ownerships.

The examples of red pine and white pine stands and of conflicting management objectives in adjacent ownership parcels are not intended as a criticism of the planning processes of the Border Lakes project partners. Indeed, within the context of their planning activities, the agencies with ownership in the Trout Lake pilot area were successful in meeting their goals. Rather, these examples are intended to highlight two points: (1) understanding the context within which plans are developed and implemented is critical, as the ability to meet goals changes when the spatial or temporal scale of the analysis changes, and (2) a landscape-level analysis can add value to the planning processes of individual stakeholders by changing the scale, and thus the perspective, of the analysis. The experience of the Trout Lake modeling group suggests that these points were lost neither on the Border Lakes partners nor on key decisionmakers within the partner agencies. For example, when Figure 5.7 was shown to partners and decisionmakers, there was a ready acknowledgment of potential management incompatibilities and, more significantly, a willingness to consider strategies that would increase coordination of management actions across ownership boundaries and possibly to spatially rearrange the management objectives in order to reduce management conflicts.

5.6.3. Phase Three: Early Review of the Pilot Model

Flexibility in meeting with partners individually and in small groups was especially important during the review of the pilot model, in addition to during the formal forums held by the Fire Learning Network. During Phase Three, two distinct audiences for

Figure 5.6. The projected abundance of red pine and white pine across the Trout Lake pilot area, as a proportion of the landscape, under four management scenarios: (a) red and white pine stands and (b) red and white pine as component of all stand types.
Figure 5.7. Potentially incompatible land management objectives under the proposed management scenario. The inset highlights part of the Trout Lake pilot area to show the juxtapositioning of land management objectives.

knowledge transfer and communication emerged: natural resource managers and decisionmakers. Flexibility in the size and timing of update meetings was essential in order to obtain feedback from resource managers. Agency partners, including the Superior National Forest, the Minnesota Department of Natural Resources, and Voyageurs National Park, critiqued the pilot model results during two separate sessions. Representatives from the Department of Natural Resources and Quetico Provincial Park were unable to attend either session, necessitating individual updates for their staff. Results from the pilot project were well received, with several insights gained into the ways in which the approach could be helpful to each agency's mission, planning efforts, and land management. We considered this to be successful knowledge transfer, and a breakthrough in the project. The pilot project enabled us to establish internal advocates and identify early implementers among the technical partners.

Although well-received by the Voyageurs National Park staff, the pilot project was less directly relevant to their routine management and planning because no National Park lands were located within the pilot area. Moreover, the park's recent staff changes required the team to update them on the Border Lakes project's background and history. We expect that expansion of the model to cover the entire landscape will create opportunities for greater participation on the part of park staff in the future.

5.6.4. Phase Four: Build Institutional Support

Building on the renewed enthusiasm among the resource managers permitted additional outreach to institutional decisionmakers for continued, landscape-level modeling. We developed a fact sheet to explain the pilot modeling results and the potential for the information to inform existing management plans and future collaborative activities (TNC 2004). With the cooperation of a key Superior National Forest technical team member, outreach began with National Forest decisionmakers and the Minnesota Forest Resources Council. This Council, the primary large-scale goal-setting body for forests in Minnesota, encompasses government agencies, nongovernmental organizations, and timber-industry partners. The Council invited core team members to present their preliminary results to a broader audience, leading to a grant awarded to the team, matched by the National Forest, to continue the ecological modeling work. Again, we considered this to be a successful knowledge transfer, made easier by the existing technical backgrounds of National Forest and Council leadership.

Building institutional support was more difficult than expected within The Nature Conservancy as a result of communication and priority-setting challenges. Although the 3-year strategic plan of the Minnesota chapter of The Nature Conservancy identified the Border Lakes landscape as a priority, initial participation on the part of local program staff was limited, largely due to a lack of staff. Communication among key staff members also needed improvement, and the definition of roles and expectations needed to be clarified. Issues surrounding communication were resolved in part by creating a vision document for The Nature Conservancy's role in fire issues related to northern forests. To advance the project, it was also necessary to distinguish clearly between two main project objectives: developing ecological models to create a scientific underpinning for modeling and identifying appropriate strategies for implementation. A long-term goal of greater partner involvement and leadership in building institutional support within participating agencies was also developed.

Was this article helpful?

0 0

Post a comment