Applications of urban growth models and wildlife habitat models to assess biodiversity losses

Habitat loss and subsequent fragmentation due to urban development is part of a larger suite of anthropogenic impacts on biodiversity, but it now ranks among the principle causes of species endangerment in the United States. Several types of urban growth simulation models have been developed which can supply useful information for biodiversity planning. In many cases however, the data required for biodiversity planning may not be compatible with the urban models, leading to analytical inaccuracies and misleading conclusions. Here, we examine several lines of logic likely to be employed in biodiversity assessment and show how assumptions built into the data influence model outcome. Introduction Biodiversity can be described conceptually as a collection of indices, ranging from diversity of compositional, structural, and functional biotic elements (Noss 1990), to ecosystem, species, and genetic diversity (Soulé and Mills 1998). Representing such a broad ecological concept requires the combination of several models (Cogan in press), and central amongst these are predictions of species habitat quality and quantity. Habitat loss or fragmentation due to urban development is only one of many anthropogenic impacts on biodiversity today (Tilman and Lehman 2001), but it now ranks among the principle causes of species endangerment in the United States (Dobson et al. 1997, Vitousek et al. 1997, Wilcove et al. 1998). Urban growth models developed recently, incorporate a broad range of variables and spatial dependencies (Makse et al. 1995, Couclelis 1997, White et al. 1997, Batty 1998, Clarke and Gaydos 1998, Landis and Zhang 1998, Makse et al. 1998). These models do not themselves attempt to determine the environmental consequences of future urbanization. However, by providing a localized abstraction of the direction and magnitude of land use change, they offer a starting point for assessing future impacts on biodiversity. Planning efforts are not currently making full use of basic biodiversity information (Press et al. 1996, Crist et al. 2000), and improved linkages between urban growth models, biodiversity models, and land use planning is urgently needed. By using predictive models of urbanization and its effects on biodiversity, county planners and other stakeholders will be able to visualize and evaluate different future growth scenarios as an effective way to lessen the impact of urbanization on biodiversity (e.g., Landis 2000) Biodiversity assessments often rely on ecological models that simplify and abstract the biophysical world based on a series of assumptions about ecosystem functions. The set of assumptions may not be clearly understood by the model user and this can lead to inappropriate applications of the models. Problems of generalization and appropriate model selection are further compounded as more models are used in combination. When planning for both urban growth and biodiversity conservation today, a planner may be faced with the task of using two potentially complex types of models: one of urbanization processes and another of ecological processes. Many urban growth models use a simple binary classification of land use into urban or non-urban, although in fact urbanization includes a wide range of settlement patterns and human densities. Perhaps most significant for biodiversity conservation is the phenomenon of urban and suburban sprawl at the margins of existing metropolitan areas. Because the sprawl is incremental, the loss of habitats in these “front line” areas can be difficult to control. The boundary between urban and non-urban is fuzzy at best. Similar challenges are involved in predicting the impact of urbanization on biological species, involving simple habitat classification schemes and crude habitat suitability rating systems to predict whether a given land use or land cover class is or is not suitable habitatthe biophysical world based on a series of assumptions about ecosystem functions. The set of assumptions may not be clearly understood by the model user and this can lead to inappropriate applications of the models. Problems of generalization and appropriate model selection are further compounded as more models are used in combination. When planning for both urban growth and biodiversity conservation today, a planner may be faced with the task of using two potentially complex types of models: one of urbanization processes and another of ecological processes. Many urban growth models use a simple binary classification of land use into urban or non-urban, although in fact urbanization includes a wide range of settlement patterns and human densities. Perhaps most significant for biodiversity conservation is the phenomenon of urban and suburban sprawl at the margins of existing metropolitan areas. Because the sprawl is incremental, the loss of habitats in these “front line” areas can be difficult to control. The boundary between urban and non-urban is fuzzy at best. Similar challenges are involved in predicting the impact of urbanization on biological species, involving simple habitat classification schemes and crude habitat suitability rating systems to predict whether a given land use or land cover class is or is not suitable habitat for a species. As a first step, urban growth models sensitive to increasing human population density and urban expansion in rural regions can be usefully combined with biodiversity models. Since even relatively sparse development on a parcel-by-parcel basis can dramatically affect biodiversity, the growth models should be able to detect change over fairly small spatial areas (e.g. 100 meter grids). The biodiversity models should likewise be sensitive to land use change at a similar spatial grain. Even with the constraints of generalization and spatial grain compatibility, information gained from such a union of predictive urban and biodiversity models will be valuable in helping to anticipate and avoid biodiversity erosion caused by habitat loss and fragmentation. In this report, we examine several lines of logic likely to be employed in the combination of models used to plan for urbanization and biodiversity conservation. We also show how assumptions inherent to the data can influence model outcome. Examples are presented from a case study in Santa Cruz County, California, where multiple urban growth scenarios are combined with land cover data, and wildlife habitat relationship (WHR) models. Methods Habitat quality and quantity aspects of biodiversity were examined using three principle inputs: urbanization scenarios, wildlife habitat maps, and species habitat models. Output from the analyses is reported as loss of habitat area, or in some cases, in terms of impact to the vertebrate species associated with degraded habitats. Limited data availability did not permit an impact analysis for invertebrate species. A flow chart of the models and analyses provides an overview of the biodiversity sensitivity analysis (Figure 1). Three different previously developed models for predicting patterns of urban expansion were tested. The three models included the “urban buffer” (see below), “Landis” (Landis and Zhang 1998) and “Clarke” (Clarke and Gaydos 1998) scenarios. Outputs from the different growth models were then used in conjunction with coarse grain (100 ha minimum mapping unit) land cover maps from the California Gap Analysis Project (GAP, Davis et al. 1998). The Landis and Clarke models were also used with a finer grain (1 ha) land cover data set. This map layer was commissioned by the Association of Monterey Bay Area Governments (AMBAG) based on 30-meter Landsat Thematic Mapper (TM) imagery. Spatial distributions of 21 individual vertebrate species predicted to occur in the study area were made possible by applying wildlife habitat relationship (WHR) models (Airola 1988) to the courser grained GAP land cover data. Potential impacts of urban growth to these species were explored by intersecting scenarios of future urban growth from each of the three models with the WHR-based predicted distributions of the species. The onehectare finer grained land cover data was intersected with output from the three urban growth models to generate impact assessments for generic habitat types, such as “coastal oak woodland”, without evaluating potential to individual vertebrate species from WHR (Fig. 1). Figure 1. Flow chart for biodiversity sensitivity analysis. Three urban growth scenarios and two land cover models combine to evaluate vertebrate and habitat impacts in Santa Cruz County, California. GAP 100 ha land cover WHR models Landis growth scenario Clarke growth scenario urban buffer growth scenario County 1 ha land cover vertebrate species impacts habitat impacts Urban Growth Models: As a case study, we have employed three different urban growth scenarios, each portraying a possible future urbanization pattern in Santa Cruz County, California, USA. All three scenarios are based on the same initial urbanization pattern from the California Gap Analysis Project, derived largely from Landsat TM satellite imagery (Davis et al. 1998). The Gap urban data have a coarse (100 hectare) spatial grain, which can potentially cause problems when analyzing some land cover effects (e.g. absolute measures of fragmentation). In this study, we use the Gap urban data only for urban model starting points and relative area comparisons. The first growth scenario is based on a simple spatial buffer, which is generated by expanding current urban land use areas outwards by a distance of 500 meters. The 500meter forecasted growth area appears as a narrow red band around the current urban areas (Figure 2). A second growth scenario is based on a model of urbanization developed by Landis and Zhang (1998), which incorporates socioeconomic and physical data to predict areas of future urbanization. Using logit (natural log of the odds ratio) models of historical land use change, 100-meter grid cells ar