This project develops a technique that combines visitor and resource data through simulation software that uses intelligent agents, coupled with Geographic Information Systems (GIS), to simulate wilderness rock climber behavior at Joshua Tree National Park (JTNP).  The purpose of the JTNP simulation is twofold.   The first purpose of the JTNP simulation is to help resolve the decade old, controversial issue of fixed anchors in JTNP wilderness.  JTNP simulations are specifically designed to 1) evaluate the geographic extent of climbing impacts, 2) identify wilderness attributes that are most responsible for attracting heavy use, and 3) assess the merit of the proposed fixed anchor permitting process.  Second, the simulation provides the National Park Service with a powerful management tool to quantify visitor flow and evaluate policy effects.  Process-based, statistics-based, and hybrid simulations are tested for validity.  Process-based simulations rely on hierarchical, behavior rules induced from surveys, whereas statistics-based simulations model visitor use data collected by monitoring equipment and observation.  Hybrid simulations integrate both techniques.  The advantages and limitations of the simulations are examined in order to develop the optimal technique for simulating wilderness visitor flow relative to national park resources, facilities, alternative management plans, and wilderness use trends. Most importantly, simulation illustrates the extent of resource conflicts, and allows park managers to explore the cascading consequences of proposed management plans prior to field implementation.

    The need for understanding the relationship between resource impacts and wilderness visitor flow is a fundamental issue addressed by wilderness managers.  Wilderness use trends reveal that use levels, as well as biophysical and social impacts, are steadily increasing (Cole, 1996; Cole et al., 1997).  The conundrum of mitigating resource impacts, without negatively affecting the wilderness experience with excessive regulations, is a prevalent theme in wilderness literature (Nash, 1967; Hendee et al., 1990).
    Simulation of wilderness recreation is an interdisciplinary technique that combines aspects of geographic information science, environmental psychology, and artificial intelligence modeling, to resolve wilderness management issues.  The concept of simulating recreation use was conceived in the early 1970s due to the nexus of necessity and computer technology.  The efforts of social scientists at the USDA Forest Service Forestry Sciences Laboratory to explore the human dimension of wilderness resource management resulted in the Wilderness Travel Simulator and the Wilderness Use Simulation Model (Smith and Krutilla, 1976; Shechter and Lucas, 1978).  These two models were designed to understand the distribution of visitors in wilderness settings in order to estimate the number of encounters as a function of increased intensity of use.  Although these early attempts at wilderness recreation simulation were successful, the cost and time required to run simulations were prohibitive.  Wilderness recreation simulation models reemerged in the mid-1990s due to low-cost and high-speed computer technology.  Two philosophically different models were developed.  Researchers at the University of Vermont developed an empirical/statistical model using an object-oriented, dynamic simulation package typically used for industrial applications called Extend (Wang and Manning, 1999; Lawson and Manning, 2000).  The model effectively predicted encounters, but did not model human transactions with changing environmental conditions nor discrete behavior of individuals.  It also lacked a georeferenced visual representation of the modeled environment.  Nearly contemporaneously, Itami and Gimblett developed Recreation Behavior Simulation (RBSim), a model based on artificial intelligence principles to simulate discrete, temporospatial behavior (Gimblett et al., 2002).  Geographic Information Systems are used to represent the simulation landscape (Gimblett, 2002), and intelligent agents that behave according to hierarchical rules represent wilderness visitors (Itami, 2000; Daniel and Gimblett, 2000).           
    The Joshua Tree National Park fixed anchor controversy provides a perfect application for the simulation approach.  Many studies have addressed the impacts of rock climbing on biological resources (McMillan and Larson, 2001; White, 1999).  Studies at Joshua Tree National Park have specifically correlated popular rock climbing sites and impacts on vegetation and bird communities (Camp, 1998; Overlin et al., 1999).
    Sociologists and environmental psychologists have examined the perceptions, preferences, and motivations of mountaineers and rock climbers (Mitchell, 1983; Ewert, 1985).  The Adventure Model correlates individual attributes, such as experience level and frequency of participation, to activity setting attributes, such as potential risk and remoteness (Ewert and Hollenhorst, 1989).  This general model predicts the type of activity setting that climbers of various experience and commitment levels would prefer to visit.  A clear understanding of preferred activity settings, and the associated potential for environmental degradation, has significant wilderness management implications (Ewert and Hollenhorst, 1996).  This project extends this concept by using simulation to link statistically determined numbers of wilderness climbers to specific climbing sites.