Human influences on soil pollution: Modeling soil lead with urban environmental characteristics using regression, support vector machine and random forest

Authors: An-Min Wu*, University of Southern California, Nic Jelinski, University of Minnesota
Topics: Soils, Spatial Analysis & Modeling, Geographic Information Science and Systems
Keywords: urban environmental health, soil lead, spatial modeling
Session Type: Paper
Day: 4/5/2019
Start / End Time: 3:05 PM / 4:45 PM
Room: Jackson, Marriott, Mezzanine Level
Presentation File: No File Uploaded


Soil lead in urban areas has been linked to elevated blood lead levels with long-term health effects on young children. Due to the accumulation of historical lead paint and leaded gasoline, lead-contaminated soil continues to pose a high risk to children in the urban environment (Mielke and Reagan 1998; Laidlaw and Flippelli 2008). This study aims to understand soil lead distributions throughout the Twin Cities, Minnesota over time using various spatial modeling methods. General linear model, random forest and support vector machines will be tested for mapping soil lead levels from samples collected in both 1997-1998 and 2015-2016. The predictor variables for soil lead models will consist of urban environmental characteristics related to the contamination sources, including traffic volume, building density and housing age, and the model performances will be compared. The resulting spatial distributions of soil lead will help house siting and land use planning in order to protect human health in the urban environment.

Abstract Information

This abstract is already part of a session. View the session here.

To access contact information login