Publication Date

2019

Document Type

Dissertation/Thesis

First Advisor

Luo, Wei

Degree Name

Ph.D. (Doctor of Philosophy)

Legacy Department

Department of Earth, Atmosphere and Environment

Abstract

The Central Valley (CV) Aquifer, California is one of the most productive regions of the United States, where large amount of nitrogen fertilizer has been applied for the last few decades to increase the crop productivity. The application of excessive fertilizer has increased the level of nitrate (NO3-N) in the groundwater to above EPA’s maximum contamination level (MCL) of 10 mg/L in several domestic, public and monitoring wells. The concentration of nitrate in the groundwater can vary spatially depending on the local nitrogen sources, aquifer characteristics and geochemical condition of the area. The changing hydrogeological conditions of the valley due to excessive groundwater pumping could further complicate the fate of nitrate in the aquifer. The statistical, index-overlay and process-based methods are commonly used to study the vulnerability of aquifer to nitrate contamination. The main purpose of this dissertation is to understand the spatial distribution of groundwater nitrate contamination in the CV which was achieved by applying a relatively new method called Geodetector. Geodetector analyzes the spatial distribution of groundwater NO3-N based on the spatial variance of groundwater nitrate in stratified geographic area of important explanatory variables such as fertilizer, cropland, permeability, slope, dissolved oxygen, etc. The assumption is that if an environmental factor contributes to a groundwater contamination, the spatial distribution of the groundwater contamination should be similar to that of the environmental factor and this spatial association is measured by the Power of Determinant (PD), which is derived based on local and global variances. This method identifies significant explanatory variables, vulnerable areas, relative significance of variables and interaction between variables to strengthen or weaken the effect. The watersheds in the Central Valley were used as the basic analysis units and the percent of wells with above 5 mg/L NO3-N (PWN>5) were calculated for each watershed for the period of 2002 to 2014 to represent its contamination level. Each explanatory variable was processed to different spatially stratified areas to quantify their spatial correspondence with PWN>5. The results of Geodetector method were compared with those from Principal Component Analysis (PCA) and Geographically Weighted Regression (GWR). Finally, maps of susceptibility to nitrate contamination of the CV were developed based on the groundwater basins using the optimized-DRASTIC index and Geodetector-Frequency Ratio Index (GFR). The quantitatively derived GFR index values resulted in better map as reflected by high PD values and correlation coefficient with observed nitrate contamination pattern. The Geodetector method makes no assumptions about the data and has the ability to process multiple data sets, which can be both categorical and continuous. Therefore, Geodetector is advantageous over PCA and GWR, which often suffer from the multicollinearity of data. The Geodetector method offers water resources managers and policy makers a general framework to assess groundwater contamination vulnerability in any other study areas.

Extent

182 pages

Language

eng

Publisher

Northern Illinois University

Rights Statement

In Copyright

Rights Statement 2

NIU theses are protected by copyright. They may be viewed from Huskie Commons for any purpose, but reproduction or distribution in any format is prohibited without the written permission of the authors.

Media Type

Text

Download

Included in

Geography Commons

Share

COinS