Using Remote Sensing to Estimate Exposures to Flaring from
Unconventional Oil and Gas Operations
Abstract
Unconventional extraction technologies including directional well
drilling and high-volume hydraulic fracturing have led to a surge in
domestic oil and gas production in the US over the past decade. Flaring,
or the combustion of petroleum products into the open atmosphere, is a
common practice associated with oil and gas development. Flaring is
potentially a major source of hazardous air pollutants including carbon
monoxide, nitrogen oxides, sulfur dioxide, toxic heavy metals, and black
carbon. Quantifying exposures from flaring is challenging due to the
sheer number of drilling sites, irregularity in production, and a lack
of regulatory requirements to report flaring activities by the industry.
The few datasets reporting oil and gas production are proprietary and do
not contain sufficient information to reliably quantify emissions from
flaring needed for public health studies. Furthermore, oil and gas
operations typically occur in rural regions that are not covered by
routine air quality monitoring. Remote sensing provides a unique means
of monitoring flaring activities, as multi-spectral satellite
instruments collecting data at night are able to detect thermal
anomalies. Recent advances to algorithms have resulted in two global,
open source products that provide data on heat sources related to
flaring: the MODIS fire detection product and the VIIRS Nightfire
product. We show that VIIRS Nightfire uniquely provides sub-grid scale
(< 750m) identification of flaring sources in the Eagle Ford
Shale region of Texas, one of the largest and most active drilling areas
in the US. VIIRS Nightfire provides flaring source temperature, area,
and radiant heat intensity, along with estimates of methane and carbon
dioxide. Using spatiotemporal hierarchical clustering we pinpointed
sources of flaring, and by empirically fusing VIIRS radiant heat and
source area with oil and gas well permit data, we estimated flared gas
volume. These estimates provide a new and novel means of quantifying
exposure for the many residents of the region who have been impacted by
the surge in oil and gas extraction. While we show an application of
these exposures to examine the association of flaring with adverse birth
outcomes in Eagle Ford, our data and methods are easily generalizable,
having wide-reaching policy and public health applicability.