Underground natural gas storage facility operations and well leakage events in the United States

Abstract

Federal minimum safety standards for underground natural gas storage (UNGS) operations in the United States (U.S.) were set by the Pipeline and Hazardous Materials Safety Administration (PHMSA) in the aftermath of the 2015 Aliso Canyon well leakage incident. Beginning in 2017, these standards required operators of U.S. UNGS facilities to publicly report: 1) leakage events that meet established criteria for incidents; 2) annual details of well construction, maintenance, and integrity testing activities; and 3) the occurrence of leading indicators of leakage (i.e., safety-related conditions). We reviewed both PHMSA reports and previously published studies to compile a dataset of 53 well leakage events at UNGS facilities in the U.S. that align with the PHMSA reporting requirements. Casings (55.0%) and wellheads (32.5%) were the most common failed well components identified. Human intervention through incorrect operation or outside force damage was the most common (39.3%) cause of well leakage events. Well leakage events caused by corrosion resulted in significantly greater emissions than those with other causes. UNGS annual reports detailed operations at 406 UNGS facilities with 17,993 wells—14,469 were injection and the remainder monitoring. Of these wells, 11,446 (63.6%) were reported to have a design that is potentially susceptible to a single point of failure, which is more than three times the previous estimate. Additional information about the status and construction of individual wells is needed to confirm the number of operational wells with a single-point-of-failure design. These data would help regulators identify single-point-of-failure wells and ensure the elevated risks associated with them are appropriately mitigated by proposed risk management plans. Statistically significant differences in the characteristics of UNGS facilities that have and have not experienced well leakage events were observed among those operating in depleted hydrocarbon reservoirs. Depleted hydrocarbon UNGS facilities that experienced well leakage events were generally larger, which suggests that focusing regulatory inspections and reviews on these facilities may be justified. The UNGS data reported were not sufficient to build a predictive model that forecasts well leakage events. Modeling efforts were limited by the data reporting format, which requires UNGS operators to aggregate data by facility in their annual reports. Gathering and evaluating information for individual wells at UNGS facilities would be valuable and may provide additional insights into well leakage risks to better inform risk management planning efforts.