| Description |
-
Abstract: Globally, 1.8 billion people use contaminated water, leading to childhood diarrhea. Ceramic water filters can reduce diarrhea when used consistently and correctly, although adherence can decline over time. These declines have previously been investigated using agent-based models (ABMs) in Limpopo, South Africa. In this thesis, the previous ABM was re-built with minimal field data, and ad... read moreherence and flow rate preference were incorporated as variables of interest. Additionally, a linear regression model was developed as a traditional comparison. Normalized water contamination matched the previous ABM within 200 days. Altering adherence produced statistically significant (p<0.0001) but non-field-relevant declines in diarrhea, while changing flow rate preference produced larger declines (p<0.0001). The linear regression model had R2 = 0.535, but contained bias in residuals. I found adherence was less influential than other studies suggest. Nonetheless, ABMs can reproduce water contamination accurately without full access to original data, thus deserving consideration alongside traditional modeling techniques.
Thesis (M.S.)--Tufts University, 2016.
Submitted to the Dept. of Civil Engineering.
Advisor: Mark Woodin.
Committee: David Gute, and Daniele Lantagne.
Keywords: Environmental health, Environmental engineering, and Systems science.read less
|
|---|
This object is in collection
