The projects below have been conducted by Don Morgan over the last five years

Introduction

The recent paradigm shift in forestry towards ecosystem management has created new challenges for forest managers (Bunnell 1995, Holt 2001). First, because ecosystem management takes guidance from naturally disturbed forests, forest managers must be able to characterise natural ecological patterns. Second, because ecosystem management focuses on the state or condition of the forest over the long term, while forestry operations serve as agents of change, forest managers must be able to predict the long-term ecological consequences of management. Related to the second point, because ecosystem management also places importance on socio-economic impacts, forest managers must be able to predict impacts on resource use. Third, because ecosystems function at various scales, forest managers must express and assess information at different scales. Fourth, because ecosystem management emphasises community participation, forest managers must include stakeholders in management. Data management underpins most of these objectives.

Data management

In the past, access to well organized data has limited the application of landscape analyses. Data came from a variety of agencies and in a variety of formats with little or no information (meta data) to describe it. The majority of the time in a spatial analysis project was dedicated to translating data in different formats, projecting spatial data to a common map projection, identifying and repairing missing or bad spatial and attribute information and stitching together 1:20,000 map tiles into a single coverage of an ecologically relevant area. With increasing computer processing speed and cheaper storage it is becoming far easier to manage large data sets.

Characterising natural landscapes

Ecosystems change naturally over ecological and geological time frames. The “range of natural variability” is the amount of variation exhibited by ecosystem characteristics (or ecosystem components) over an appropriate time frame (Holt 2001). The Biodiversity Guide Book provides a good initial description of natural disturbance patterns in different regions of British Columbia, but provides less information about natural variability. Forest managers, developing long term management plans feel that more detail about disturbance patterns would help planning. While some ecological patterns require further data collection, others may be analysed using existing information. In particular, the historic range of seral stage distributions, patch size distributions and connectivity can be characterised using computer models that use existing forest inventories and information about disturbance rate.

The scope of available geographic information can limit analysis of landscape pattern. Although relevant ecological features may not be explicitly represented in a Geographic Information System (GIS), some may be derived from existing data using rules based on expert interpretation.

Predicting long-term consequences

Predicting the consequences of forest management is not trivial for at least two reasons. First, management acts on complex and inter-connected systems where impacts extend beyond the target resource. Second, the impacts of forestry accumulate over time and over large areas. Landscape modelling is particularly useful when problems include aspects of natural disturbance, succession and management over long time frames and large areas (Baker 1989, Hunsaker et al. 1993, Sklar and Costanza 1991, Turner 1989).

Landscape modelling is still evolving. Non-spatial models, ranging from simple equations to simulations and optimisations have been used for the last three decades to aid forest management decisions. The application of GIS to forestry, since about 1985, has highlighted the importance of spatial relationships in forest management: resources are linked with the landscape, and extracting one resource from a specific location impacts others in that area and in surrounding areas. Ignoring spatial information may significantly influence predicted levels of sustainable forest use (Daust and Nelson 1993). Spatial models (emerging in the early 1990s) combined spatial simulation and temporal simulation to predict impacts of forest management at the stand scale (e.g., Alvarez-Bullya and Garcia-Barrios 1993) and the landscape-scale (e.g., Baskent and Jordan 1991, Nelson and Finn 1991, Sessions and Sessions 1991). These spatial forestry models tended to focus on harvesting economics and timber supply. Wildlife modelling has followed similar trends to forest modelling, moving from non-spatial models of populations and habitat to models that include spatial patterns (Daust and Bunnell 1994). Some modelling approaches integrate forestry models with models of wildlife habitat and forest structure to assess impacts of forest management on biodiversity (e.g., Daust and Bunnell 1992, Davis and Barrett 1992, Hansen et al. 1993).

While addressing biodiversity has emerged as an important goal for modelling within an ecosystem management context, several important ecological processes are typically poorly addressed. First natural disturbance is not modelled explicitly. Many models subtract a yearly average to account for episodic natural disturbance events and may only consider impacts on timber (e.g., FSSIM). Even if impacts on habitat are tracked, these models do not adequately capture the possible consequences of large disturbances. For example, a large disturbance may eliminate some key old forest habitat whereas an averaged disturbance maintains some habitat while disturbed areas recover. Second, succession is not modelled explicitly. Usually, changes in tree species composition are only modelled following harvesting. As well, succession that occurs with stand and ageing that occurs outside of the timber harvesting landbase is not modelled. Tree species composition, an important indicator of biodiversity (BCMOF and MOE 1995), is not adequately modelled.

Many of the currently available spatial forest models have limited ability to analyse large landscapes such as TSAs, because, in part, they include fine detail such as block boundaries and road locations, making model runs very slow and data compilation onerous. Non-spatial models that function well at the TSA scale do not include the spatial information necessary to assess general trends in harvesting location over time. An intermediate approach to modelling forest harvesting, based on the principles of natural disturbance modelling (e.g., harvesting location based on probabilities), may allow sufficient spatial detail without excess effort.

Linking scales

Ecosystems function at a variety of scales from microscopic to global. Presently land use decisions in BC are made at a number of different scales including Forest Region (7 to 30 million hectares), timber supply area (approximately 1 million hectares), landscape unit (30 to 100 thousand hectares), watershed, and forest stand. Higher level (larger area) decisions should establish a framework for land use planning at lower levels. Unfortunately, this is not always the case in BC. Often land use decisions assume an even timber flow for each landscape unit in a timber supply area (TSA), where in reality the timber supply targets are for even flow over the entire TSA—not within each landscape unit (Research Branch 1998). To provide proper guidance to landscape unit planning, the timber supply must be spatially and temporally partitioned to determine harvesting levels in each landscape unit.

Similarly, natural disturbance information is compiled by Biogeoclimatic Ecosystem Classification (BEC) zone (Meidinger and Pojar 1991) while biodiversity conservation planning occurs at the landscape unit scale. To provide proper guidance to landscape unit planning, natural disturbance patterns must be spatially and temporally partitioned to describe the range of natural disturbance in each landscape unit. The ecological classification system is an hierarchical system. BEC inventories are typically mapped at the 1:250,000 scale. Using predictive ecosystem mapping techniques more detailed ecological information (site series) can be provided at a 1:20,000 scale.

Including Stakeholders

Stakeholder participation is fundamental to ecosystem management. Growing public awareness and use of non-timber forest values has increased pressure, sometimes through boycotts and civil disobedience, on forest managers to treat all forest values equitably. This has resulted in increased public involvement in planning both in the US (Selin et al. 1997) and in Canada (e.g. Province of British Columbia 1999). Thus, many people, including staff, local experts, and representatives of industry and the environment, want their views to be considered and respected in decision-making. Since these stakeholders can influence the decisions that are considered or accepted, they need opportunities to learn about the problems and to contribute to the solutions.

Landscape modelling can provide a focus for stakeholders to share perspectives, learn, and integrate a variety of information. The results from a sufficiently realistic, yet tractable and verified, landscape model can improve comprehension of the consequences of various management options. Generating model results and presenting them to appropriate people, however, is not sufficient to increase understanding. The target audience must have confidence in model results, rather than blind trust or unwarranted scepticism, and the knowledge of how to interpret them. This implies that people must be involved in the process of landscape model development and analysis.

The notion that a group with appropriate expertise can combine talents to produce better models and solutions is not new (Clark et al. 1979, Holling 1978, Kyng 1991, Maxwell and Constanza 1997a, Selin et al. 1997). Participatory model design combines diverse sources of expertise and may generate novel solutions by exploring a wide range of scenarios, including alternatives that might otherwise not be considered (Gray 1989, Kyng 1991, Schuler and Namioka 1992). The benefits of a participatory design process extend well beyond the model produced (Kyng 1991, McLain and Lee 1996, Maxwell and Constanza 1997b, Selin et al. 1997), bringing benefits to participants such as improved understanding, communication and co-operation among stakeholders and designers; conflict resolution and consensus building; and an opportunity to influence the decision-making process. While the secondary benefits of collaborative modelling are well understood, the goal of most such processes still seems to be generating model results for a limited set of viewpoints (McLain and Lee 1996).

There is a need for a structured framework to guide the collaborative development of landscape models that has the clear goal of building the decision making capacity of stakeholders. This framework should be oriented to finding solutions for a given socio-economic and ecological problem, to maximise mutual understanding, and to improve decision-making.

Acknowledgements

DULP is a BC Ministry of Forest Research Branch project with funding assistance from Forest Renewal BC.