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

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