| This study represents a contribution to land use modeling through the creation of a framework to predict and assess the influence of future population growth on land use change in the California Mojave Desert. A series of development probability models were constructed at different spatial scales during a 20-year period based on proximity to existing development sites, proximity to primary and non-primary roads, percent of surrounding development, city boundary, and site slope.; By using U.S. population growth projections, three future land use scenarios—Current Development Trend (CDT), Open Space Conservation (OSC), and Pro-Development (PD)—were constructed and analyzed to demonstrate the application of the model. The scenarios represented different assumptions about growth in the region between the year 2000 and 2020. Under the CDT scenario, which considers the existing socioeconomic trends in the region and a new settlement density of 3.76 residents/ha, 177,120 new hectares would be developed to house 665,971 expected new residents. This represents a near doubling of area under development. Under the OSC scenario, which used the same population but a higher settlement density of 20 residents/ha, 33,298 new hectares would be developed—a decrease of over 80% from the CDT. The PD scenario used the new settlement density of 3.76 residents/ha with 50% increase in population. Under this scenario, 265,680 hectares would be developed—an increase of 50% and 698% as compared with the CDT and OSC scenarios, respectively. By altering model assumptions, an endless number of alternative futures can be constructed and analyzed.; This study also analyzed the effects of geographic scale on the development probability models. A model for each portion of the three counties partially included in the study area was constructed. The predicted development probabilities obtained with each county model were then compared with those of the regionwide model. The analysis of development at different scales allowed us to understand the variables that influence development at regionwide and county levels. Each scale may be appropriate for different aspects of interest. The analysis indicated that predicted future development patterns under the regionwide and county models are quite different. As a result, using large scale models to predict development at finer scales could lead to poor development predictions. Conversely, using county models across a multi-county region could also result in unrealistic predictions. Policy makers should consider a multiple scale approach to reach better understanding about the development process across the landscape. |
