Strong temporal variability in methane fluxes from natural gas well pad soils

Highlights

• Short-term measurements may be inadequate to characterize soil fluxes at well pads.

• We measured methane fluxes from well pad soils over several days and seasons.

• Flux variability was due to soil, atmospheric, and bacterial processes.

• Fluxes varied by more than an order of magnitude over hourly and daily intervals.

Abstract

We measured methane and carbon dioxide fluxes at natural gas well pad soils and undisturbed soils in the Rocky Mountain and Gulf Coast regions of the United States, including producing and gas storage wells. We collected both short-term (15 min) and multi-day (between 3 and 8), continuous measurements at 47 well pads and two undisturbed locations. Methane fluxes varied by more than an order of magnitude over periods as short as 30 min (e.g., 19–593 mg m⁻² h⁻¹ in one instance), and diurnal and seasonal variability was also significant (e.g., spring-to-fall change from 509 to 14174 mg m⁻² h⁻¹). We hypothesize that short-term flux variability was caused by pulsed flow of methane during its migration through the subsurface. Barometric pressure and well conditions likely impacted fluxes, but we found only weak evidence for this. Bacterial methanotrophy appeared to impact methane flux magnitude and variability. We injected methane into the subsurface at one well, and we found that, while fluxes of methane and carbon dioxide, and combustible soil gas concentrations, increased in response to the injection, the response was not uniform, and fluxes exhibited high hourly-scale variability, in spite of a constant injection rate. Methane fluxes tended to be higher at well pad soils compared to background soils (often much higher), and fluxes tended to be higher at well pad locations closer to the well head.

Introduction

Many studies have investigated emissions of methane from the oil and natural gas industry to the atmosphere, especially during the past decade. Oil and gas-related methane emissions have been estimated at the global (Schwietzke et al., 2016), national (Alvarez et al., 2018; Miller et al., 2013; Omara et al., 2018), regional (Foster et al., 2019; Riddick et al., 2019; Schwietzke et al., 2017; Smith et al., 2017), facility (Brantley et al., 2015; Robertson et al., 2017) and component (Allen et al., 2013; Hendler et al., 2009; Mansfield et al., 2018) scales. This work has provided information that is being used to reduce methane emissions (Konschnik and Jordaan, 2018; Lamb et al., 2015; Lyman et al., 2019) and evaluate the benefits and costs of oil and gas production (Brandt et al., 2014; Howarth, 2015; Muresan and Ivan, 2015).

Few studies, however, had investigated the prevalence and impacts of natural gas leaks from subsurface oil and gas infrastructure, which can result in emissions from soils to the atmosphere. A prominent example is a well casing failure at the Aliso Canyon natural gas storage facility in California. The resultant subsurface leak resulted in one of the largest natural gas leaks in the history of the United States (Conley et al., 2016). Also, Kang et al. (2016) found that methane emissions from plugged and abandoned wells in Pennsylvania are a significant part of total emissions in the state. In other cases, methane emissions from the subsurface have been found to be unimportant relative to other oil and gas sources (Lyman et al., 2017; Townsend-Small et al., 2016). Even small methane emissions, however, can indicate subsurface infrastructure failure (Bachu, 2017) and can be associated with groundwater contamination (Cahill et al., 2017, 2018; Cheung et al., 2010; McMahon et al., 2018).

Well integrity failure is the generally-recognized cause of subsurface natural gas leaks, and was discussed in detail by Cahill et al. (2019), Jackson (2014), and Davies et al. (2014). Well integrity failure is caused by problems with well casing or cement that allow natural gas to escape from the well and into the surrounding rock, groundwater, or surface soil. Casing corrosion or failure can compromise well integrity (Davies et al., 2014; Jackson, 2014). Shrinkage of cement can compromise the seal between the well casing and the surrounding rock or the casing, allowing gas to migrate along the rock-cement or cement-casing interface (Dusseault et al., 2000), including from methane-rich areas above the producing reservoir (Dusseault and Jackson, 2014). Bachu (2017) and Watson and Bachu (2009) found that the majority of subsurface leaks from wells in Alberta, Canada, originated from methane-rich areas above the producing reservoir. Montague et al. (2018) used statistical methods to determine that well age, deviation of the well from vertical, and casing size are the most important predictors of well integrity issues. Some wells in which cement does not extend throughout the full length of the casing have been shown to leak (Lindeberg et al., 2017; McMahon et al., 2018), but Bachu (2017) and Lyman et al. (2017) found that cement depth could not explain the observed distribution of soil methane emissions. Operators are not required to monitor for subsurface leaks at oil and gas wells in the United States.

While it is well known that methane fluxes at undisturbed soils and landfills exhibit temporal variability (Christophersen et al., 2001; Le Mer and Roger, 2001) due to bacterial activity (Koch et al., 2007) and atmospheric conditions (Czepiel et al., 2003; Simpson et al., 1999), less is known about the temporal variability of fluxes at oil and gas well pads. Kang et al. (2016) and Lyman et al. (2017) made short-term repeat measurements at the same well pads. Kang et al. found relatively little variability in repeat measurements (for high-emitting locations; variability was higher for low-emitting locations), but Lyman et al. found high variability. The difference between the two findings could be due to the fact that Kang et al.‘s flux chamber enclosed well bores that were open in some cases, allowing for free flow to the atmosphere, while the measurements of Lyman et al. were on soils without an opening to the subsurface. Because of this difference, the portion of emissions measured by Kang et al. that were due to subsurface leakage, rather than direct well bore emissions, is unknown. Other well pad soil flux measurements collected to date have included only one-time, short-term measurements (Boothroyd et al., 2016; Day et al., 2014; Forde et al., 2019a; Townsend-Small et al., 2016). While these studies have provided an understanding of the causes and prevalence of subsurface leakage, additional work is needed to elucidate the characteristics and magnitude of this source across the industry.

Here we present long-term measurements of methane and carbon dioxide soil flux at producing and gas storage well pads in the Rocky Mountain and Gulf Coast regions of the United States. We collected both short-term and continuous measurements over several days from multiple locations at these well pads during multiple seasons. We investigate here the extent and causes of observed temporal and spatial variability.