Developing a Spatio-temporal Modeling and Forecasting Tool for a Subsurface Water-borne Toxin (1,4-Dioxane)

Groundwater systems are intrinsically heterogeneous with dynamic spatio-temporal patterns, which cause significant challenges in quantifying and mapping their complex processes. However, accurate forecasting of regional groundwater contamination is commonly needed to identify its spatio-temporal dynamic that helps the public anticipate the timing and severity of potential groundwater quality issues and possibly serve as an early warning system. This study focuses on modeling a plume of 1,4-dioxane originating from the Gelman site beneath the city of Ann Arbor, Michigan. It proposed a novel methodology to consider the spatially and temporally irregular and uncertain nature of groundwater contamination data to analyze the historical trends of dioxane concentration and predict its transportation, including the following key components:

A random forest interpolation model was deployed to fill in or extend fragmented time series data gaps among all the monitoring wells;

The Mann-Kendall test was applied to evaluate the trend of dioxane concentrations at various wells;

An automated time series machine learning (AutoTS) package was utilized to predict the best future values forecasts; and

An R-based Shiny web application was designed to allow visualization and quantification of dioxane contamination analytical data.

This research introduced a novel framework for filling spatial and temporal data sampling gaps in groundwater contamination to offer an effective and promising way to predict future plume concentration and spatial distribution.