Evaluation Of Sociodemographic Factors And Presence Of Oil And Gas Waste Disposal Wells In Kansas

Katherine Rose Wolf

This thesis is restricted to Yale network users only. It will be made publicly available on 08/28/2021


Background and objective: Kansas hosts the second-largest number of wells in the United States for the disposal of oil and gas waste (n = 13,396), which can contain chemicals toxic to human health. In this study we evaluate potential associations between oil and gas waste disposal well locations and sociodemographic characteristics in Kansas.

Methods: We obtained oil and gas production and disposal well data from the Kansas Geological Survey and sociodemographic data at the census block group level from the United States 2013–2017 American Community Survey. We used two-variable t-tests, logistic regression with and without thin plate spatial splines, and generalized additive modeling to evaluate associations between disposal well presence and sociodemographic variables including population density, sex, age (percent 0–17 and percent 65+), percent white, percent high school completion rate, percent without health insurance, renter occupancy rate, percent unemployment, median household income, median home value, income concentration at the extremes, percent in poverty, median individual earnings, and production well counts, among others.

Results: Bivariate analyses showed block groups with disposal wells (n = 489) compared to those without disposal wells (n = 1,850) to have lower population density (7 versus 978 person/km2), percent female (49.7% versus 50.7%), percent without health insurance (7.6% versus 8.6%), renter occupancy rate (19.7% versus 32.3%), unemployment rate (3.1% versus 4.1%), home values ($101,050 versus $112,800), and percent in poverty (9.2% versus 10.9%); and higher percent white (92.7% versus 80.9%) and percentage 65+ (18.5% versus 13.6%). After adjustment in multivariable models, the odds of disposal well presence in a block group were 22% lower per 100-person/km2 increase in population density (OR = 0.78, 95% CI: 0.74–0.83) and 4% higher per additional production well (OR = 1.04, 95% CI: 1.03–1.04). No other covariates significantly predicted disposal well presence.

Conclusions: Disposal well presence in Kansas was negatively associated with population density and positively associated with oil and gas production well count after adjustment for sociodemographic variables.