Recreation habitat suitability indices: key concepts and a framework for application in landscape planning.

Abstract

The use of any landscape for recreational purposes means that natural processes will be impacted and that resource use conflicts may develop. This conflict will be exacerbated where the goals of the recreationist conflict with legitimate resource extraction activities. In order to deal with these issues land use planners must integrate recreation such that potential conflicts are minimized and in some cases some activities may be prohibited. This paper discusses the concept of recreation habitat suitability indices (rHSI) as a means of identifying opportunities for recreation and potential conflicts/impacts with ecological processes, other resource users and other recreation "species." Outdoor recreation takes a variety of forms, with each form requiring different environmental conditions. Habitat requirements of recreationists can be understood within the confines of separate HSI for individual recreational activities. As such different recreation types can be considered as distinct recreation species with distinct recreation habitat needs. The preferred habitat of specific recreationists can be measured and can be used to predict recreation species use of an area even if such use does not currently occur. In this way planners can take into account the needs of human beings in the same manner that they do other species and reduce or eliminate significant environmental effects as well as conflict between different recreation "species" and recreation and other forest uses. In this paper we present a brief assessment of the applicability of the HSI model through an examination of the boreal canoeist recreation "species." This study was undertaken as one element in the broad area planning initiative for the east side of Lake Winnipeg.

L'utilisation de tout paysage a des fins recreatives signifie qu'il y aura des repercussions sur les processus naturels et que des conflits pourront naitre quant a l'utilisation des ressources. Et ce conflit sera exacerbe lorsque les objectifs des adeptes de plein air entrent en conflit avec les activites legitimes d'extraction des ressources. Afin de regler ces questions, les planificateurs de l'utilisation du sol doivent integrer la dimension recreative afin de reduire les conflits eventuels et dans certains cas, certaines activites pourraient etre interdites. Les auteurs de cet article analysent le concept de l'indice de qualite de l'habitat recreatif recreation habitat suitability indices (rHSI) comme moyen de reperer les occasions en matiere d'activite recreative et les conflits et repercussions potentiels relativement aux processus ecologiques, aux autres utilisateurs de ressources et aux autres << especes >> d'amateurs de plein air. Les activites recreatives de plein air prennent diverses formes qui requierent toutes des conditions environnementales distinctes. On peut comprendre les exigences des adeptes de plein air en se servant d'indices de qualite de l'habitat differents pour chacune des activites de plein air. Comme tel, on peut considerer les differents types d'activites recreatives comme des especes d'activites distinctes avec des besoins distincts en matiere d'habitat. On peut mesurer quel serait l'habitat de premier choix pour des adeptes de plein air particuliers et on peut se servir de ces mesures pour predire l'utilisation d'un secteur, meme lorsque cette utilisation n'a pas lieu actuellement. Les planificateurs peuvent ainsi tenir compte des besoins des etres humains de la meme maniere que pour les autres especes et reduire ou eliminer les effets significatifs sur l'environnement, de meme que les conflits entre les differentes << especes >> d'amateurs de plein air et les activites recreatives et autres usages de la foret. Dans cet article nous presentons une breve evaluation de l'applicabilite du modele d'indice de qualite de l'habitat par le biais d'une analyse de l'<< espece >> d'amateur de plein air canoeiste boreal. Cette etude a ete entreprise dans le cadre de l'initiative plus generale de planification de secteurs pour la rive est du lac Winnipeg.

Keywords

recreation, wilderness, visitor management frameworks, habitat suitability indices

Introduction

Research directed at recreation management in parks, wilderness, and protected areas has expanded greatly over the past three decades, reflecting the fact that much outdoor recreation takes place in landscapes under such jurisdiction. In Canada, however, a great deal of outdoor recreation occurs in undesignated and unprotected crown land. Currently, about 50% of Canada is forested, 57% of this is considered commercial and much of it is largely inaccessible (Rutledge and Volde 2001; Environment Canada 2003). Less than two per cent of this is protected through national parks, provincial parks or other legislated designations and yet is what many would consider "true wilderness" (here defined as vast tracts of roadless areas where natural processes predominate but where aboriginal people may still live and practice traditional resource harvest). The current definition of wilderness is often too focused upon the concept of protected from human incursion and often ignores the existence of human habitation, or inhabited wilderness. In addition, many areas used by recreationists--and considered as wilderness by them--lie outside of protected areas.

The management frameworks of ROS, LAC, Carrying capacity, VIM, VAMP, and VERP (1) have all been developed to assist in planning and managing recreation within areas that have some form of protected areas status (Payne and Nilsen 2002). However, few models have been developed to plan for outdoor recreation in the broader landscape outside of protected areas where recreation is only one of many uses. The aforementioned models do not provide adequate inputs for non-recreational use nor do they specifically identify particularly important recreational features, so they have minimal suitability for planning in the broader landscape.

Over the past 25 years, as commitment has grown for the inclusion of broader values in land planning, there has also been an attempt in Canada to develop a uniform national ecological approach to terrestrial ecosystem classification and mapping (Ecological Stratification Working Group 1995). This approach to classification incorporates ecological values such as biodiversity and habitat quality as an integral part of sustainable forest management. On a worldwide basis, concern for the conservation of biological diversity has become a major priority (Baydack et al. 1999). Researchers in Canada are cognizant of this fact, but also positioned to take forest management one step further by incorporating social values, such as recreation, as an intrinsic component of habitat assessment and measurement. The intent of this paper is to outline an approach for such integration.

The need to understand how humans use parks and natural areas for recreational activities is growing in importance. Research on why people use parks has been exhaustive but linking this information to in situ behaviour has been difficult. The social psychological approaches that have dominated leisure behaviour research are motivation and satisfaction theory. Major investigations into these concepts can be attributed to Neulinger in 1974 and Iso-Ahola in 1979. They exposed motivation, and specifically intrinsic motivation, as a key way of determining leisure behaviour (Mannell 1987). Their findings demonstrated that the situation or place that people were at had as much to do with self-actualization of a state of leisure as the activity itself. In essence, human beings depend on environmental settings as much as an activity in order to reach a state of satisfaction. Satisfaction measures, however, have been criticized as management tools due to incremental changes in visitor expectations and the displacement of "purists" (Manning, 1999). Furthermore, satisfaction is only measurable in a post-hoc fashion and seems best suited to pre-existing recreation areas where user characteristics are well known. As such, satisfaction--like other social psychological variables such as substitutability (e.g. Shelby and Vaske, 1991), specialization (e.g. Bryan, 1977) and purism (e.g. Stankey, 1972)--may have limited utility in multiple-use broad area planning where data about potential users is non-existant.

An alternative approach suggested here is to extend the concept of Habitat Suitability Indices (HSI, U. S. Fish and Wildlife Service, 1981) already used in legacy GIS databases. Habitat suitability indices model the critical resources provided by a range of habitats, at a particular scale, that are required by individual species (Uhmann et al. 2001). In 1996, Brunson argued that this approach could be used to model recreationists, by viewing the landscape as consisting of 'recreational habitats' of varying suitability for particular recreational activities (i.e. 'species'). As with its traditional application in wildlife management, several 'species' can be modeled simultaneously on the landscape, providing a means to independently assess the needs of different users within a common framework. Because areas on the landscape may meet different suitability requirements for each species, HSI can then predict the extent to which 'species' needs overlap and are conflicting or compatible. Such an approach could greatly extend our ability to minimize the impact of resource extraction activities in multiple use areas, in much the same way as HSI is used in planning for wildlife. (For wildlife examples see, U. S. Fish and Wildlife Service, 1981). Indeed, competition (sensu Adelman 1982) that arises when different recreational activities occur on the same landscape can be modeled following this approach.

Despite the method's promise, only the broad concept of HSI has been discussed in the recreational literature (e.g. Brunson 1996) and only a preliminary assessment has been made defining fundamental parameters such as 'recreational species' (e.g. Hamilton 1996). Furthermore, these HSI models lack a framework identifying the scale of the recreational activities and the specific types of recreation to be modeled, thus violating the conditions necessary when applying HSI. Ideally, the best analog for a 'species' would be an activity and group of individuals that are specialists and thus selective of habitats they use. Shafer (1995) suggests that the requirements of 'purists' are the best way to determine the baseline support needed for recreational activities. Correspondingly, purists may also be viewed as 'species' within the context of HSI. For example, a river landscape is obviously required for a canoeing activity, but boreal river canoeists are very selective; only a narrow range of specific conditions provide a high quality recreational experience (Boxall 1995).

Objective

In this paper, we define and describe an HSI-based approach to modeling a 'purist' (sensu Schaefer 1995) 'recreational species,' the boreal river canoeist. As well, we will develop a parameterized recreational (r)HSI model within a hierarchical framework. This framework defines the scales over which the HSI model can be applied and is compatible with legacy GIS databases used by forestry and broad-area planning and management.

Currently the Manitoba government is studying a broad area planning report for the east side of Lake Winnipeg. The study includes an evaluation of the construction of an all-weather road and further hydroelectric development potential for the area. The East Side planning area encompasses 12.5 million hectares of boreal shield forest, is dotted with innumerable lakes and several large pool and plunge rivers, two of which are Canadian Heritage Rivers. It is considered a premier "wilderness" canoeing destination (Wilson 1999). A key concern for recreationists and ecotourism operators is that their needs are considered in the planning process. As such, the area represents an ideal location to test the model. The intent of this paper is to describe the model and to report on the early stages of its application to the East Side of Lake Winnipeg.

Methodological Approach

HSI Background

The Habitat Suitability Index is a simple empirical method of evaluating wildlife habitat preferences for individual species on the landscape (Wheatley 2001). Models using this approach were originally developed in the late 1970's as part of the U.S. Fish and Wildlife Services (FWS) Habitat Evaluation Procedures (U.S. Fish and Wildlife Service 1981). The method derives a single species suitability index value (HSI) for each habitat-type in a region based on the abundance and/or quality of several habitat resources or conditions. Choice in the number of resources and conditions to use must be determined and assessed relative to the wildlife population in question (Layer and Maughan 1985). As a practical consequence, the same habitat can receive a different HSI value for each modeled species, making it an appealing modeling approach for multiple-use management areas. In addition, the selected model parameters are not restricted to quantitative data and may include qualitative or ordered classes, providing that each component can be given a suitability ranking (Schaumberger et al. 1982). Because of the flexibility of this approach, HSI have been widely applied in the management of economically important or indicator wildlife species (MacMahon et al. 1984; Allen 1982; Jasikoff 1982) and specific life-history stages of a species (Stanley and Trial 1995). Overall, several key features of HSI modeling approaches are helpful in conservation management of wilderness landscapes. These include the capability of HSI to:

* model specific target or indicator species,

* incorporate 'sub-models' or multiple components,

* generate rules from expert-based information and opinion (heuristics),

* predict critical habitat on the landscape with or without empirical data,

* provide a common framework for managing multiple species on the landscape, and

* determine where species competition or conflicts might arise.

There is no intrinsic reason that habitat approaches need to be restricted to wildlife (Haskell 1940). The core foundations of the method are the principles that species' habitat preference relates in some way to the availability of suitable resources and that these can be modeled using a formal framework based on habitats (Wheatley 2001; Layer and Maughan 1985; Schaumberger et al. 1982). Many choices and preferences of recreationists are influenced by local environmental conditions that can, potentially, be modeled using HSI (Brunson 1996). In this context, natural landscapes can be seen to provide a range of recreationally 'suitable' habitats (rHSI) specific to various recreation 'species' over a range of spatial and temporal scales (Hamilton 1996). For instance, a boreal landscape provides recreational habitats for purist canoeists, fly-in fishing and moose hunters (Table 1). Each has particular requirements; the advantage of using HSI is that it can predict where competition and conflict might arise over time and space. A location with high suitability for fly-in fishing during the spring season or locations with clearcuts that provide moose hunters with ideal 'habitat' might reduce the sense of wilderness for the boreal canoeist. A positive HSI for one species in this instance reduces the suitability for another. Alternatively, a positive HSI for one species might be negative for another species or--in the case of the air traffic associated with fly-in fishing, for example--use of an area by one species might reduce the suitability for another. In contrast, the infrastructure needed for fly-in fishing (e.g. lodges etc.) may be ideal for moose hunters as well, yielding a positive suitability for both. This common framework for defining habitat is what makes HSI valuable as a method for landscape management.

HSI Method Details

A standard HSI methodological approach can be followed in developing an rHSI model. An overview of this approach is presented in Figure 1. Model development begins with the scoping of critical variables that might explain 'species' habitat selections on the landscape (Korman et al. 1994; Terrell et al. 1982). For wildlife populations, these variables consist of conditions at a site and the resources available. The choice of resources and conditions to use in the model and indeed the response of the species to these resources is often determined through interviewing experts on that species (Wheatley 2001). In a typical scenario, experts are selected and contacted using a Delphi survey approach (Crance 1987). For this approach, questionnaires are circulated that direct the experts to select those attributes that are critical in determining suitable habitat. As well, experts must assess how changing levels of availability of the selected resources or conditions influence habitat suitability often using suitability curves. The survey results are summarized and if discrepancies are encountered, a series of iterative steps may be initiated to 'fine-tune' or clarify the differences (Schuster et al. 1985). The process ends when a consensus emerges on the number of resources or conditions to be used in the HSI model and the level of each required for habitat to be deemed suitable (Crance 1987). The final model is developed and then directly applied to habitats in a management area (Wheatley 2001). Because the HSI models based on a Delphi approach are often strongly heuristic, direct testing of the model in the field is essential (Wheatley 2001; Cock 1978). For mobile wildlife species, it is assumed that habitat will be preferentially utilized in proportion to its suitability (Cock 1978). Several formal statistical tests based on abundance relative to availability have been developed thus allowing us to differentiate the relative weighting of commonness versus importance (Aldredge and Ratti 1986; 1992).

Defining 'Suitability' and Calculating HSI Values

The iterative process used in identifying the levels of suitability of resources and conditions is often the most challenging part of the process. Species response to resources can vary substantially with availability, such that experts must define a 'suitability index curve'. A suitability curve gives the range of suitability 'scores' appropriate for each level of a resource encountered in the habitat (Korman et al. 1994). For example, preliminary results of experts developing a recreational HSI for boreal river canoeists have identified availability of suitable campsites per 100 km stretch of river as a critical regional resource (Figure 2). In addition, at lower availability (e.g. greater than 25 km apart) the recreational 'habitat' is deemed not suitable to attract or support recreationists (Figure 2a.). At some intermediate level then, suitability may be seen as directly proportional to availability of sites (Figure 2b.). At very high availability the habitat is deemed ideal' with respect to availability of suitable sites, with enough to accommodate the recreationists present (Figure 2c). A common property of these suitability index curves is that the suitability 'score' is unit-less and expressed between zero (poorest) and 1 (ideal) (Schaumberger et al. 1982). The x-axis however can be expressed in any natural unit depending on the resource being examined or it may be expressed as a proportion of the total availability of the resource (Wheatley 2001). The relationship between resource availability and habitat suitability can be different for each resource or factor, and often requires 'fine-tuning' by the panel of experts (Crance 1987).

[FIGURE 1 OMITTED]

[FIGURE 2 OMITTED]

Several habitat resources and conditions may be included in the HSI model, each with its own suitability index curve. To obtain the final HSI value, the multiple SI values (represented by [V.sub.i]) must be numerically combined. The most common approach is to calculate the geometric mean of all of the SI scores (Korman et al. 1994; Layher and Maughan 1985). The general formula is:

HSI([V.sub.1],..., [V.sub.p])= ([p.[product].[i=1]] [V.sub.i])[.sup.1/p] [Equation 1]

where [V.sub.i] is the suitability index score (as determined from the suitability curve) for habitat resource i. If any of the resources or conditions are considered to be limiting (where 'limiting' implies that its availability determines overall habitat suitability) then an alternative to Equation 1 is to simply choose the minimum [V.sub.i] value obtained from among the limiting resources (Korman et al. 1994). For example, in the boreal river canoeing rHSI it has been suggested by experts that access or egress are limiting factors and that their consideration should override all others. In essence, if access and egress are unavailable, it is moot to determine the suitability as it relates to campsites. A less restrictive approach is to calculate the rHSI after weighting the variables:

HSI([V.sub.1],..., [V.sub.p] | [w.sub.i],..., [w.sub.p])= ([p.[product].[i=1]] [V.sub.i.sup.[w.sub.i]])[.sup.1/p] [Equation 2]

where [w.sub.i] in Equation 2 represents the 'weight' for each resource ([SIGMA][w.sub.i] = 1). These weights are usually provided by an expert panel and may reflect user preferences for particular resources that aren't completely limiting. In our study, purist boreal river canoeists are the species and the experts capable of commenting on the model parameters. Either equation once parameterized will obtain a single combined HSI value reflecting the unique combination of resources at that location supporting the recreation species of interest.

Defining Recreational Suitability

While human recreation systems can be conceptually modeled using a wildlife habitat approach (Haskell 1940; Greer 1990), there are fundamental differences. In particular, there is a profound link between suitable habitat and individual survival in wildlife populations (Hamilton 1996; Korman et al. 1994). However for obvious reasons this criterion is not applicable to human systems. Instead, recreation habitat is defined by the structural and functional components of the environment that support the 'success' of a particular activity (Greer 1990; Brunson 1996). An rHSI represents the combination of all of the component activities that contribute to a positive recreation experience. Because these activities occur over a range of spatial and temporal scales, the framework for modeling human recreation systems must be intrinsically hierarchical and context driven. To further explore the utility of an rHSI approach for the management of a multi-user landscape, we will outline a framework being developed for boreal river canoeists in eastern Manitoba, Canada.

Boreal river canoeist rHSI framework

The boreal river canoeing rHSI framework is intended to provide a consistent and complete model for 'typical' boreal river systems. As such the structure of the framework was designed to reflect these landscapes as they are encountered by recreationists. The framework is also being developed as part of a larger project on sustainable forest management. The goal of the overall project is to incorporate ecological and social values within a hierarchical ecological land classification system at the level of the ecosite (ESWG 1995). An ecosite is a landscape unit on the scale of 10-100 ha defined by local physiography, soil conditions and biotic components that are often associated with specific human activities such as resource extraction and recreation. In defining the rHSI framework we wanted to ensure consistency with the ecosite classification, but also have the flexibility of integrating this model with a comprehensive recreation database also under development. For these reasons, the boreal river canoe rHSI framework needed to have the following properties:

* hierarchical and multiscaled to reflect the principles used in ecological land classification and in particular the physiography of the boreal river systems;

* consistent with forest resource GIS and other management databases;

* reflect the stratification used in the sampling and statistical analysis of recreation datasets; and,

* structured to permit the development of a recreation GIS database and to ensure that attribute tables are linked directly with rHSI physical measurements.

An idealized boreal river recreational landscape is presented in Figure 3. The hierarchical framework is comprised of four basic scales consisting of: 1) the river system; 2) physiographic strata; 3) sections; and 4) sites. This framework closely matches the landscape context of a typical boreal Precambrian shield watershed system. The upper portion of the watershed (Figure 3A, strata iii) is dominated by formations of granitic rock interspersed with elongated and irregularly shaped lakes. Rivers here are characterized as pool and plunge, with lengthy sections of essentially lakelike character interrupted with short dramatic drops. Where the water flow is constricted, rapids or waterfalls are encountered. The whitewater strata (Figure 3A, strata ii) is channelized and characterized by rapids, steep banks and jackpine dominated ecosites. Depending on water flows, many of the rapids can be run by experienced canoeists, making white-water strata one of the greatest draws of the recreation experience. The lower strata (Figure 3A, strata i) is also channelized, but boggy flat terrain more commonly forms the riverbank, leaving fewer choices for campsite locations. Travel along this portion of the river is typified by little current with few rapids.

It is at the lower levels of the rHSI framework that much of the physical data used in the model is collected and modeled. For utilitarian and statistical reasons we define the landscape sequence encountered over a typical travel day (approximately 20 km) as a section of the route (Figure 3B). These sequences can consist of whitewater runs, interspersed with portages and attractions such as scenic views or cultural features. The value of defining or grouping these as sections allows for the evaluation of individual parts of a route using a standard scale of comparison. It also establishes a framework in which sites (Figure 3C) can be embedded and tracked within the GIS. For these sites, four basic types have been defined: 1) campsites--consisting of several attributes such as tent-pads, fire pits, landings etc.; 2) rapids--measured by several physical variables and level of challenge; 3) portages--recorded with length and substrate conditions, etc.; 4) attractions--consisting of view sheds, cultural and ecological features, etc. Several of these site types consist of multiple components for which rHSI sub-models may be developed. For example, the suitability of an individual campsite in our model appears to depend on an SI model that incorporates the quality of egress from/to the river, the landings available, the rock furniture, adequacy of fire pits and tent pads, availability of firewood, etc.

[FIGURE 3 OMITTED]

Parameterizing and Testing the Model

The strongest distinction between wildlife and recreational HSI is in how they are parameterized and evaluated. Recreational model development is actually far more direct than for wildlife: to develop an elk HSI human experts comment on what they believe elk require, for the rHSI our purist 'species' are also the experts that suggest the parameters required to best represent their needs. To that end, interviews were conducted with purist boreal river canoeists in order to parameterize the model and identify the physical attributes that contribute to a positive experience (Table 2). Participant selection was done through a snowball sampling technique and once repetition occurred the sampling was judged complete (Babbie 2001). The hierarchical framework provides experts with a structure that allows them to suggest parameters appropriate to each scale and to comment on how they can be combined across scales. For example, available resources in one campsite site will determine its suitability, but several sites influence the suitability of a section. Sites over several sections can help evaluate the suitability of the strata, while the combination of many sites, sections and strata determines the rHSI for a complete river system. Purist-experts not only could identify those features that were critical for a given scale, but whether the combination across scales was representative of overall experience. In essence, expert interviews also establish the foundation for evaluating the model, which, given its recreation context, may prove difficult using approaches developed for wildlife frequency such as residency in a habitat, frequency or optimal life history (Aldredge and Ratti 1992; Brunson 1996). Because we can interview our 'species,' alternative methods of evaluation using 'non-parametric rank-order correlations' (see Legendre and Legendre 1998 for definitions) between model-derived HSI values and expert-user rankings are being developed.

Preliminary Application and Discussion

The application of the model is in its early stage, as a complete field study and verification of the sites identified by experts was essential to build on the information they could provide. In total, over 260 field sites have been visited and documented with analysis still underway. Preliminary interviews with boreal river canoeists indicated that the general lay of the land (how it facilitates camp life--landing, tent pad, rock "furniture") and botanical/physiographic characteristics were impcrtant. However, it was determined through site visits that very specific environmental attributes are necessary for a positive recreational experience. In particular, "Roche Moutonees" geological features are always used when present, followed by open rock sites and lastly jackpine dominated sites (open rock and jackpine are elements of the first). Ongoing fieldwork will contribute to understanding the variation (both type and frequency) of physical attributes desired by boreal river canoeists.

In addition, with regards to the possibility of linking HSI with ecosite classifications, initial results indicate a very strong association between ecosite classification and highly valued campsite locations, thus allowing forest operations to avoid activities in important recreational habitat.

Conclusion

The intent of this paper was to describe a new model capable of integrating recreational values into broad-scale planning. While the approach has yet to be fully applied and evaluated, the preliminary study conducted on boreal rivers along the east side of Lake Winnipeg shows that the method has promise. Specifically, the application to boreal canoeists was able to identify a number of important habitat components that are highly correlated with specific ecosites. Given that the ecosites form a significant element in the stand level (operational) forest resource inventory, the model presents a unique method of informing forest planning to reduce conflict between extractive activities and recreation. Furthermore, elaboration of the model should allow identification of competing habitat needs for different recreationists on the landscape.

The development of the boreal river canoeist HSI is timely as new landscape scale models are needed to ensure that resource planning integrates the diverse needs of First Nations, recreationists, and resource extraction interests. The East Side planning area discussed in this paper is an example. Until recently, broad area land planning and land cover identification has been done by industry and governmental agencies with specific responsibilities and goals that were not necessarily shared by all potential forest users. In 1997, the Canadian Council of Forest Ministers committed their agencies to find a means of incorporating a broader range of values into forest management that includes recreational use. While there is now greater willingness to involve recreationists in planning, the landscape-level tools available are still primarily those developed for industry and government. One of the key challenges of incorporating recreational values in a broad area planning exercise is finding a 'common language' with these legacy databases and tools (Brunson 1996). The recreation habitat suitability index used in concert with an ecological classification for the region will enhance the planning process by allowing predictions about potential land use suitability prior to possible impacts from development.

While preliminary, the model and the application discussed here, illustrate the potential the rHSI model holds as a tool for broad area planning. The rHSI allows for the identification of key landscape features or "habitat" necessary for any number of recreational activities and can be combined with legacy GIS databases to map critical habitat areas and inform forest management. However, this requires that there is openness to considering these values in the broad area planning process. Fortunately, many forestry companies have become more attuned to the need to involve recreationists in planning their operations and in this case welcomed a potential solution that was scale appropriate for their operations. While the method poses several challenges, notably the need for regional ecosite GIS layers, cooperation of stakeholders and potentially time consuming fieldwork, it represents an effective planning tool that "fits" the databases and tools used by industry and government.

References

Adelman, B.J.E., Heberlein, T.A. and Bonnicksen, T.M. (1982). Social psychological explanations for the persistence of a conflict between paddling canoeists and motorcraft users in the boundary waters canoe area. Leisure Sciences 5: 45-61.

Aldredge, J. R., and J. T. Ratti. 1986. Comparison of some statistical techniques for analysis of resource selection. Journal of Wildlife Management 50:157-165.

Aldredge, J. R., and J. T. Ratti. 1992. Further comparison of some statistical techniques for analysis of resource selection. Journal of Wildlife Management 56:1-9.

Allen, A.W. 1982. Habitat suitability index models: Marten. USDI. Fish and Wildlife Service. FWS/OBS-82/10.11.

Babbie, E. R. 2001. The practice of social research. Wadsworth Thomson Learning: Belmont, California.

Baydack, R.K., H. Campa III, and J.B. Haufler. 1999. Practical approaches to the conservation of biological diversity. Island Press, Washington, DC.

Boxall, P.C., Watson, D.O., and Englin, J. (1996). Analysis of the revealed preferences of backcountry recreationists for forest and park management features in the Canadian shield region. Can. J. For. Res. 26(6): 982-990.

Brunson, M. 1996. Integrating Human Habitat Requirements into ecosystem management strategies: a case study. Natural Areas Journal 16: 100-107.

Bryan, H. 1977. Leisure value systems and recreational specialization: the case of trout fisherman. Journal of Leisure Research 9: 174-187.

Canadian Council of Forest Ministers. 1997. Criteria and Indicators of Sustainable Forest Management in Canada. Technical Report.

Cock, M. J. W. 1978. The assessment of preference. Journal of Animal Ecology. 47: 805-816.

Crance, J.H. 1987. Guidelines for using the DELPHI technique to develop habitat suitability index curves. U.S. Fish and Wildlife Service. Biological Report 82(10.134).

Ecological Stratification Working Group (ESWG). 1995. A National Ecological Framework for Canada. Agriculture and Agri-Food Canada, Research Branch, Centre for Land and Biological Resources Research and Environment Canada, State of the Environment Directorate, Ecozone Analysis Branch, Ottawa / Hull.

Greer, Jerry D. 1990. Recreation habitats: A concept for study and management. Proceedings of the National Outdoor Recreation Trends Symposium 111. J. T. O'Leary et al., eds. Indiana University Leisure Research Institute, Indianapolis, IN. 339-49.

Hamilton, H.R. 1996. A conceptual approach to recreation habitat analysis. Natural Resources Technical Note REC-02, U.S. Army Corps. of Engineers.

Haskell, E.F. 1940. Mathematical systemization of "environment," "organism," and "habitat." Ecology 21: 1-16.

Iso-Ahola, S. 1979. Basic Dimensions of Definitions of Leisure. Journal of Leisure Research 11: 28-39.

Jasikoff, J.M. 1982. Habitat suitability index models: Ferruginous Hawk. USDI. Fish and Wildlife Service. FWS/OBS-82/10.10.

Korman, J., C.J. Perrin and T. Lekstrum. 1994. A guide for the selection of standard methods for quantifying sportfish habitat capability and suitability in streams and lakes of British Columbia. Report prepared by Limnotek Research and Development Inc. Vancouver, for B.C. Environment, Fisheries Branch, Research and Development Section, Vancouver, B.C.

Layher, W.G. and O.E. Maughan. 1985. Spotted bass habitat evaluation using an unweighted geometric mean to determine HSI values. Proc. Okla. Acad. Sci. 65: 11-17.

Legendre, P. and Legendre, L. 1998. Numerical Ecology, 2nd English edition. Elsevier Science BV, Amsterdam.

McMahon, T.E., J.W. Terrell, and P.C. Nelson. 1984. Habitat suitability information: Walleye. U.S. Fish and Wildlife Service. FWS/OBS-82.10.56.

Mannell, R. and Iso-Ahola, 1987. Psychological nature of leisure and tourism experience. Annals of Tourism Research 14: 314-331.

Manning, R.E. 1999. Studies In Outdoor Recreation: Search and research for satisfaction, 2nd edition. Oregon University Press, Corvallis OR.

Neulinger, J. 1974. The Psychology of Leisure: Research Approaches to the Study of Leisure. Springfield, IL: Charles C. Thomas.

Schaumberger M., A.H. Farmer and J.W. Terrell. 1982. Habitat suitability index models: introduction. USDI. Fish and Wildlife Service. FWS/OBS-82/10.

Schuster, E.G., S.S. Frissell, E.E. Baker and R.S. Loveless, Jr. 1985. The Delphi method: application to elk habitat quality. U.S. Dept. of Agr., For. Serv., Intermountain Research Station. Research Paper INT-353.

Shafer, C.S. and Hammitt.W.E. (1995). Purism revisited: Specifying recreational conditions of concern according to resource intent. Leisure Sciences 17: 15-30.

Shelby, B. and J.J. Vaske. 1991. Using normative data to develop evaluative standards for resource management: A comment on three recent papers. Journal of Leisure Research 23(2): 173-187.

Stanley, J.G. and J.G. Trial. 1995. Habitat suitability index models: non-migratory freshwater life stages of Atlantic Salmon. USDI. Biological Science Report 3.

Stankey, G.H. 1972. A strategy for the definition and management of wilderness quality. In Natural Environments: studies in theoretical and applied analysis, J.V. Krutilla, ed. Johns Hopkins University Press. Baltimore, MD.: 88-114.

Terrell, J.W., T.E. McMahon, P.D. Inskip, R.F. Raleigh, and K.L. Williamson. 1982. Habitat suitability index models: Appendix A. Guidelines for riverine and lacustrine applications of fish HSI models with the Habitat Evaluation Procedures. U.S. Dept. Int. Fish. Wildl. Ser. FWS/OBS-82/10.A.

Uhmann, T.V., Kenkel, N.C. and Baydack, R.K. 2001. Development of a habitat suitability index model for burrowing owls in the eastern Canadian prairies. Journal of Raptor Research 35: 378-384.

U.S. Fish and Wildlife Service. 1981. Standards for the development of habitat suitability Index models. 103 ESM. USDI. Fish Wildl. Serv., Div. Ecol. Serv.

Wheatley, M. 2001. An Empirical Evaluation of a Habitat Suitability Index Model for the North American Red Squirrel in West-Central Alberta. University of Alberta,Edmonton, Alberta.

Wilson, H. (1999). Angering the water gods of Manitoba's Pigeon River. Kanawa: 40-48.

RESEARCH NOTES include preliminary or summary descriptions of research methods and/or results that do not yet address their full theoretical, policy or practical implications. The intention is to provide researchers and practitioners with a forum for presenting interesting but preliminary ideas, methodologies, or results in the spirit of fostering reflection and dialogue.

Dr. J. Michael Campbell is Associate Professor of Recreation Management and Community Development, Physical Education and Recreation Studies, University of Manitoba. His research focuses primarily on social and physical impacts of human activities in parks and protected areas and the human dimensions of fish and wildlife and their links to sustainable tourism. He is founder and past chair of the Parks and Protected Areas Research Forum of Manitoba. He can be contacted at Michael_Campbell@umanitoba.ca, 112 Frank Kennedy Centre, University of Manitoba, Winnipeg MB, R3T 2N2

Dr. David Walker is an Assistant Professor in Environment, Earth, and Resources, University of Manitoba. His research and teaching focus on geomatics in biological systems, remote sensing, bio-metric applications, and ecosystem investigations in grassland and boreal forest environments. He can be contacted at djwalkr@cc.umanitoba.ca

Borden Smid is a Ph.D. candidate in Environment, Earth and Resources, University of Manitoba. He is a sessional instructor and research assistant in the Health, Leisure and Human Performance Research Institute. His interests include understanding natural resource use as habitat use, traditional resource uses, management of natural areas with agricultural techniques and periphery tourism. He can be reached at borden_smid@umanitoba.ca

Dr. Richard Kenith Baydack is Associate Dean and Professor in Environment, Earth, and Resources, University of Manitoba. His research focuses on developing practical applications of ecosystem management of habitat to conserve biodiversity. Recent emphasis has centered on the North American Waterfowl Management Plan, North American Grouse Partnership, Manitoba Ecosite Project, and determination of the effects of human activities on species survival. He can be contacted at baydack@ms.umanitoba.ca

(1) Recreational Opportunity Spectrum, Limits of Acceptable Change, Visitor Impact Management, Visitor Activity Management Process, Visitor Experience and Resource Protection Table 1: A comparison of several boreal river recreational species.

Conflict

Resource Potential

Spatial Temporal Specificity with other

Requirements Requirements and Intensity species at Recreational for main for main of Use for moderate Species activity activity main activity levels of use Canoeist Site level to Spring or high/low High

regional during intensity

high water Moose Site level Fall/but high/medium Medium

Hunters depends on intensity

legislated

timing Fly-in Site to local All seasons medium/medium Low

Fishing level but poor intensity

late fall

and spring Table 2: An overview of the river system, strata and section parameters identified by purist experts for the boreal canoeist rHSI framework Hierarchy-level Description River system Accessibility The accessibility as it relates to getting access/egress to/from the river system; access and egress

may be de-coupled Trip Time Length of trip relative to start location Combined Strata Combined strata Si's weighted by preference Suitability and proportion Combined Impacts Degree of on-river and off-river impacts

(litter, roadways etc.) Strata Alternative Routes The suitability/availability of alternate

routes (lake strata only) Combined Rapid Overall rapid suitability weighted by Suitability interspersion Combined Campsite Overall Suitability of campsites derived from Suitability the sections Combined Portage Overall ranking or suitability of portages in Suitability a strata Combined Impacts The overall impacts to a strata from sections Section Campsite Suitability Weighted suitability of all campsites along

the section Nearest neighbor or The interspersion of rapids of various classes interspersion of rapids along a section Combined Suitability of Weighted suitability of all portages along the Portages section Combined Viewshed of the Weighted suitability of the viewsheds section Combined Suitability of Weighted suitability of attractions (cultural Attractions and ecological) Site Campsite Tent Pads Tent pads at a site Fire pits Availability of fire pits Firewood Availability of firewood Rock Furniture Rock furniture presence and suitability Landings and site access Accessibility: egress to/from river and boat

moorings Impacts Impact class of the site Campsite Area Total area of campsite Aesthetics Viewshed, sightlines from campsite Rapid Class Class of rapid Length Length Distance from/to next Context/sequence of rapids rapid Portage Length Total length of portage Substrate Conditions Ground obstructions, moist soils, etc. that

impede movement Canopy Obstructions Low canopy and branches that impede movement Attraction Viewshed An SI under development based on structure and

perception Cultural Feature Aboriginal features (e.g. rock paintings,

petroglyphs) Ecological Feature Ecologically unique feature (rare plants,

edibles, etc.)