A predictive model of open illegal dumping in East Tennessee

While solid waste management issues continue to be a top priority for both state and local governments, little success seems to have been made in the area of open illegal dumping. While solid waste collection and disposal systems have been established at the county level to take care of local needs, illegal dumping continues to occur in many areas. And since dumping is widespread, it is nearly impossible for local officials to monitor an entire county, hoping to catch illegal dumpers in the act. Thus, this paper describes the fundamentals of a predictive model to locate those areas which are most vulnerable to illegal dumping, so that those areas can be more heavily patrolled. After sampling fifty-three dumps in Claiborne, Grainger, Hanock, Hawkins and Union counties in East Tennessee, this study showed the most common characteristics of illegal dump sites to be 1) roadside location (often with access from a pull-off area), 2) non-visibility (usually a steep embankment or dense forestation), 3) location in less-densely populated areas (rarely within view of neighboring houses), and 4) distant proximity to convenience center facilities. County boundary, transportation, hydrography, city and dump data were then placed in a digital Geographic Information System database, where it was possible to select for the above spatial criteria. First, TIGER road data was buffered to select all areas within 150 meters of a road. Then a two-mile buffer around convenience center sites was used to cookie-cutter the road buffer data. The resulting polygon coverage was used to overlay with 30x30 meter slope polygons, derived from Digital Elevation Model (DEM) data from the USGS. Polygons with slopes of 6.75 degrees (fifteen percent) or greater were reselected from this coverage and appended to a water buffer coverage which had been clipped by the road buffer to further identify steep water channel embankments alongside of a road not detectable by thirty-meter resolution DEM data. Thus, the model resulted in selecting steep, roadside locations more than two miles away from the nearest convenience center. While a predictive model can be a useful tool in monitoring areas for open illegal dumping, it is not the end-all solution to the problem. Cleaning up existing dumps is not necessarily the solution either, since people often continue to dump in those locations. The key is in changing people's behavior, and if a predictive model can help select those areas which need heavy surveillance, then perhaps officials can catch and reprimand more of the people responsible for dumping, thus creating a deterrent for others in the area.