Repeated helicopter-based surveys of fugitive hydrocarbon emissions from tight oil operations: Persistence and driving factors

A recent study used helicopters and optical gas imaging (infrared-based) to study hydrocarbon leaks [1] across numerous oil and gas producing basins. The appearance of leaks appeared to have small statistical correlations with factors such as age of wells, production rate, or water production. However, there have not been few if any attempts to observe the presence of leaks over time. In this work, we used the same helicopter-based optical gas imaging techniques to revisit wells in the Bakken formation that had previously been measured in 2014 [1] In this study, we visited 353 of the total 683 well pads in Bakken from the original study, and we also visited 50 randomly selected well pads that were newly developed between studies. We first examine the presence of leaks as a function of different factors, including: operator, number of wells, tanks, and treaters on a well pad; and numerous other possible emissions drivers. In addition, using the previous years observations as our prior, we conducted a Bayesian analysis to determine the likelihood of finding a leaking or non-leaking well pad and compared the observations to a Monte Carlo simulation of randomly distributed leaks. We find that the overall prevalence of observed emissions was similar for each year (11% for 2014 observations and 10% for 2015 observations). We also find whether a pad was found to be leaking in the first observation year had a statistically significant influence (4.9σ) on the presence of a leak in the following observation year. Similar levels of persistence are also apparent for non-leaking well pads. [1] D. R. Lyon, R. A. Alvarez, D. Zavala-Araiza, A. R. Brandt, R. B. Jackson, and S. P. Hamburg, "Aerial surveys of elevated hydrocarbon emissions from oil and gas production sites," Environ. Sci. Technol., p. acs.est.6b00705, Apr. 2016.