Richard O. Flamm

Integrating spatial models into local land-use decision making

ABTRACT

In Florida, land-use decisions at the local level range from individuals operating watercraft to the designing of an infrastructure for urban development. These decisions, which individually are often considered relatively innocuous, accumulate to have measurable impacts on landscape structure and function. Landscape ecologists who seek to include the "human dimension" in their studies need to understand how humans use information to make decisions. For example, humans use relatively simple spatial models, typically maps of what's present, to make a multitude of land-use decisions. As modelers, we know that more complex manipulations of spatial data can provide additional knowledge about the system and its functions. Ecology, which has had tremendous success in model derivation, has had little impact on decision making at the local level, partly, because of the difficulty of introducing it's models into public arenas.

This paper discusses integrating ecological models into land-use decision making at the local level. Examples are presented of spatial models now being used in land-use decision making; a new, more complex model that has yet to be introduced; and scenarios in which model-based land-management alternatives for South Florida's natural system boundary are evaluated based on the landscape structure and function of that area. A computer-based modeling environment called the Land Use Change and Analysis System (LUCAS) is presented as an organizational framework for integrating ecological models into land-use decision making.

INTRODUCTION

In Florida, land-use decisions at the local level range from individuals operating watercraft to the designing of an infrastructure for urban development. These decisions, which individually are often considered relatively innocuous, accumulate to have measurable impacts on landscape structure and function. Landscape ecologists who seek to include the "human dimension" in their studies need to understand how humans use information to make decisions if they expect their models to influence future land use and ultimately have a positive impact landscape structure and function. Unfortunately, the ways humans make decisions is complicated and not easily modeled and is probably the major contributor to the error component of any predictive land-use model. Therefore, for several projects in Florida, we are examining land use decision making by applying spatial models of land use within sociological frameworks.

Three examples of integrating spatial models into local land-use decision making are presented: a spatial model now being used in local land-use decision making; a new, more complex model that has yet to be introduced; and scenarios in which model-based land-management alternatives for South Florida's natural system boundary are evaluated based on the landscape structure and function of that area. I conclude with a discussion of some issues that research teams will likely face during their attempts at introducing their spatial models into land-use decision making.

The most common spatial model used in local decision making is a map of what's present. Although not simple to construct, maps of what's present are probably the easiest models to integrate into land-use decision making. This is not very surprising given that they have been used for centuries, are well accepted, and although specifics of maps differ among cultures, overall, people's perception of what they mean is fairly uniform.

One large project that has been undertaken in Florida's Department of Environmental Protection is the mapping of seagrass scarring caused by boats (Sargent et al. 1995). Scarring is of significant environmental concern in Florida because of the continuous decline in seagrass coverage and the importance of seagrass as benthic habitat. In this project, the goals included (1) building a simple land-use model that informed the target users -- resource managers -- about land-use impacts and (2) influencing future land use in a positive way.

Maps of scarring were derived from aerial photos (Figure 1). Scarring was categorized as light, moderate, and severe. As maps of what's present, these models were easily understood, were widely distributed to resource managers, and have had a positive impact on local land use. They are being used on signs to inform boaters of the presence of seagrass, to determine where to place navigational markers directing boats away from seagrass beds, and to delineate closure zones so that seagrass beds can recover. The result has been that many seagrass beds are recovering, indicating beneficial changes in land use.

The second example, a more complex analytical model in the process of being introduced into land-use management, is associated with manatee protection. The manatee is an endangered herbivore that has been negatively impacted by a burgeoning human population along the coast of Florida. The dominant manatee mortality factor -- collisions with power boats -- is responsible for approximately 25% of their annual mortality. Thus, one important component of manatee protection is the designation of slow speed zones for boats in areas of high manatee abundance. Currently, these areas are delineated, in part, through visual inspection of aerial survey data plotted onto a map of Florida coastline (Figure 2). The problem with this approach is that the aerial survey maps can be difficult to interpret, which can cause readers to suggest differing delineations.

One task of the Marine Mammal program in the Florida Department of Environmental Protection was to devise a method to objectively delineate high-abundance areas from aerial survey point data. The method we developed and are experimenting with is a spatial filtering algorithm we call the fixed-area flexibly-shaped spatial filter.

Running the algorithm results in a contoured surface representing manatee abundance derived from aerial survey point data (Figure 3). The cell values in this map are interpreted as the likelihood of observing an animal while traveling along the aerial survey route. By applying rules in a GIS, such as "locate all areas with a manatee abundance value greater than 0.002 manatees/pixel," we can use these maps to help designate protection zones.

Using the spatial-filter model in decision making requires a different approach than that given in the seagrass scarring example because the target users are different and the model is less intuitive. The targeted audience for the seagrass-scarring model was primarily resource managers so the output of the model was catered to their uses. The spatial-filter model is targeted toward individuals or organizations that either make or are affected by policy. As such, it would be introduced in public forums such as local meetings and court challenges where the bases of decisions are subjected to close scrutiny. It is inevitable that stakeholders such as the members of the Marine Industry Association or of conservation organizations, and State and local officials will have various levels of mathematical or GIS expertise. Most of these people would probably be uncomfortable with decisions based on the model because they do not understand the underlying methodology. At present, this model is not used in manatee protection planning.

In an attempt to integrate this model into manatee protection planning, we adopted a strategy of educating policy makers and the public through an organization called the Manatee GIS Working Group. The working group consists of people interested in using GIS or understanding the role of GIS in manatee protection. Topics of interest to the Working Group include data-sharing issues, analysis techniques, and exploring decision making. The goal of the modelers is to use the manatee spatial-filter model to acquaint stakeholders in the working group with the filtering procedure and to facilitate the procedure's understanding and acceptance in public forums. We are just beginning to examine analytical techniques in the working group, so an evaluation of the group's effectiveness is premature. However, at least one stakeholder has demonstrated his committment to the process by hiring GIS expertise to represent them at the meetings.

The last example is a land-use project in its infancy called the Governor's Commission for a Sustainable South Florida. This project seeks to examine local land use and the influence of broad-scale policy on the landscape integrity of South Florida. Early phases of this program will involve simple manipulations of delineated "natural" areas. However, the project is expected to evolve into a more sophisticated integrated modeling system whereby users can evaluate alternative natural-system boundaries in terms of changes in the landscape structure of the South Florida. The project's value will depend on how well the knowledge encompassed by the modeling system is translated into policy that can be understood and accepted by land users.

The Sustainable South Florida project will probably not evolve into a comprehensive model, but rather will be a collection of alternative scenarios, each of which can be considered an integrated model in itself. The approach we are considering, which can both integrate complex ecological knowledge and represent it in a form that can be presented to policy makers and their constituents, is to develop a computer-based modeling environment in which to run the scenarios (Berry et al. 1996, Coulson et al 1989, Flamm & Turner 1994).

An example of a computer-based, scenario-driven modeling environment is the Land-Use Change Analysis System or LUCAS (Flamm & Turner 1995, Berry et al. 1996). LUCAS was developed during a U. S. Man and the Biosphere Project in the Computer Science Department at the University of Tennessee. It was designed to run management scenarios so that decision makers could visualize and analyze possible ramifications of today's actions. Figure 4 shows one of the windows that appears on the screen during a simulation exercise. The user selects a scenario to run, the number of timesteps and replicates per timestep, how many maps to save, and which effects to evaluate. Simulations are then run, and the results are displayed in the graphics window, and a variety of spatial indices are calculated.

Why use a system like LUCAS for a large-scale modeling effort like the Governor's Commission on a Sustainable South Florida? First, it's an excellent tool for interdisciplinary work because it can handle many types of information including spatial and tabular data, heuristics, and analytical models. LUCAS is also a warehouse of information, storing existing models and data that can be accessed during the development of future scenarios. Because LUCAS's structure is modular, it is not too difficult to incorporate new models or data into it. And finally, through its user-friendly graphical interface, we can represent simulation results in ways that allow users to explore management alternatives and ultimately to be more effective in formulating and implementing land-use policies.

The examples presented represent spatial models of different complexity and scale, and were designed for different target users. Each can be useful in influencing local land-use decisions, including those made by individuals, in public meetings, in permitting reviews, and in court challenges. However, introducing ecological models into land-use management is especially challenging because of the range of education and values of the people involved. The example of the propeller scarring was a simple map of what's present that was easily integrated into land-use management. The remaining models are less traditional, more complex, and more difficult to integrate into land-use decision making. Challenges to modelers associated with integrating these more complex land-use models into local land-use decisions might include the following: understanding social dynamics among stakeholders, understanding how humans value nature, being able to effectively communicate the concept of uncertainty, acknowledging the role of the courts in environmental management, and educating the public about their impacts on the landscape. In addition, landscape ecological research needs to expand its involvment beyond multidisciplinary land-use modeling into the sociology of how humans use ecological knowledge so that we can begin to investigate how land-use models influence land use.

References

Berry, M.W., Flamm, R.O., Hazen, B.C., and MacIntyre, R.M. (1996) The land-use change and analysis system (LUCAS) for evaluating landscape management decisions. IEEE (In Press).

Coulson, R.N., Saunders, M.C., Loh, D.K., Oliveria, F.L., Drummond, D., Barry, P.J., and Swain, K.M. (1987) Knowledge system environment for integrated pest management in forest landscapes: The southern pine beetle (Coleoptera: Scolytidae). Bull. Entomol. Soc. Am. 34:26-33.

Flamm, R.O. and Turner, M.G. (1994) GIS applications perspective: multidisciplinary modeling and GIS for landscape management. In: V. Alaric Sample [ed.], Remote Sensing and GIS in Ecosystem Management. Island Press, Washington, DC.

Sargent, F.J., Leary, T.J., Crewz, D.W., and Kruer, C.R. (1995) Scarring of Florida's seagrasses: assessment and management options. Florida Marine Research Insitute Tech. Rpt. TR-1.

Richard O. Flamm Associate Research Scientist
Florida Department of Environmental Protection, Florida Marine Research Institute
100 8th Avenue, SE, St. Petersburg, Florida, 33701-5095
Phone: (813)-896-8626, ext 3030
Fax: (813)-823-0166
email: flamm@manatee.fmri.usf.edu