Future Growth - Methodology

Modeling future growth starts with Landsat satellite imagery and high resolution GIS data centered on Washington, D.C. Using grids from 3 to 30 meters, the level of imperviousness for the study area was determined and aggregated at the 30 meter cell level for each of four time points, 1986, 1990, 1994, and 2000. Those cells with imperviousness of 10% or greater were defined as developed which then established the historic baseline of development patterns.

Using the historic footprint of development (1986-2000) in the Washington, D.C. area, University of Maryland researchers employed the urban growth model known as SLEUTH. The SLEUTH model projects the probability of cells becoming developed over time based on five parameters. Those parameters - dispersion (probability of any random cell becoming developed spontaneously), breed (tendency to become a growth center), spread (growth around the edge of development), road gravity (the influence of roads), and slope (the effect of topography) - are the guiding algorithms of the model.

Decision rules are then added to the model for the predictive phase which adjust the likelihood of development based on the scenario being projected. For example, current trends decreases the likelihood of development along tidal waters and some streams to reflect stream buffer policies.

More details about the methodology will be coming available on the University of Maryland website soon:
http://www.geog.umd.edu/resac/

The amount of data and the number of iterations and calculations required in the calibration phase posed new challenges to the capacity of the U.S. Geological Services lab in Colorado and required almost a week of machine time. The SLEUTH model was run on a 45 meter grid to balance between cell size and area so that the USGS computers could handle the data at one time. The predictive phase generates statistics and grid data images, which can be taken into a Geographic Information System for quantitative analysis, of the location and extent of probable development in the future. In the case of the Washington, D.C. area, only those cells which showed an 85% or higher probability of development were accepted as indicating new growth.

The ability of the model to project not only how much development is likely to occur, but the location of that change is a significant part of its power and usefulness.

More details about SLEUTH are available at: (http://www.ncgia.ucsb.edu/projects/gig/v2/Home/home.htm).