The complexities of urban environments—like variable winds off of tall buildings or surfaces—significantly impair the ability to detect natural gas leaks from underground pipelines, a new study shows.
SMU researchers found that detection probability can be reduced by 16 percentage points for methane gas leaks from main pipelines and by as much as 38 percentage points for service line leaks in urban areas, compared to rural settings.
Main lines are the larger, high-pressure pipelines under streets that carry gas to neighborhoods, acting like a main artery, while service lines are smaller ones connecting the main line to an individual customer’s meter, delivering lower-pressure gas directly to a home or business.
The researchers have also identified which city-like conditions impact the ability to detect belowground natural gas leaks using different survey methods.
“This matters because methane is both a safety risk and a potent greenhouse gas, and these detection challenges mean leaks can go unnoticed and unaddressed longer than we’ve assumed,” said Kathleen Smits, one of the co-authors of the study published in the Journal of Pipeline Science and Engineering. Smits is the Solomon Professor for Global Development at SMU’s Lyle School of Engineering and a Fellow of the Maguire Energy Institute at SMU’s Cox School of Business.
Methane gas from pipeline leaks can potentially cause dangerous explosions. It is also the second-largest contributor to global warming after carbon dioxide. Detecting and repairing these leaks could meaningfully reduce climate impacts.
Urban complexity creates detection blind spots
The SMU study is one of the first to systematically compare the effectiveness of walking, driving and simulated unmanned aerial vehicle (UAV) surveys for detecting methane leaks in realistic urban settings, Smits said.
Led by SMU postdoctoral researcher Venkata Rao Gundapuneni, the research team conducted controlled natural gas pipeline leak experiments in Colorado, examining how factors such as leak rate, wind conditions and survey height influenced the probability of detection.
Key findings:
Leak rate matters significantly: The probability of detection improved dramatically with increased leak rates—by 15 to 37 percentage points for main lines and 24 to 35 percentage points for service lines when comparing low-rate leaks (1 standard cubic foot for hour) to high-rate leaks (21 standard cubic foot for hour). This means smaller, slower leaks are far more likely to go undetected in urban settings.
Distance from pipeline reduces detection: Surprisingly, even a small shift—15 meters downwind from the pipeline—dropped the detection probability by 65 percentage points for walking surveys, 49 percentage points for UAV surveys, and 10 percentage points for driving surveys, compared to surveys conducted directly over the pipeline.
Detection Threshold Sensitivity: The study found that detection threshold settings severely impact success rates, particularly for low leak rates and service lines. A mere 0.05 parts per million increase in detection threshold could reduce detection probability by 30 to 40 percentage points for main lines at low leak rates and for service lines across all tested leak rates.
“The important takeaway from our results is that no single method works in every condition,” Smits said. “The key is to know the conditions and know how the method performs in those conditions.”
The research team also included Josh R. Aldred, a research scientist at SMU, and Daniel J. Zimmerle, Director and Principal Director of the Methane Emissions Technology Evaluation Center at Colorado State University.
This work was funded in part, under the Department of Transportation, Pipeline and Hazardous Materials Safety Administration (693JK32210006POTA and 693JK32410007POTA). The views and conclusions contained in this document are those of the authors and should not be interpreted as representing the official policies, either expressed or implied, of the Pipeline and Hazardous Materials Safety Administration, the Department of Transportation, or the U.S. Government. This research was jointly funded by the Energy and Carbon Management Commission’s Mark Martinez & Joey Irwin Memorial Public Projects Fund.