Coverage of mobile methane leak surveys of gas distribution networks: The impact of wind-aware scheduling on the coverage distribution
Year: 2026
Type: preprint
Abstract: . Mobile leak surveys of gas distribution networks typically drive each street six times, two passes a night, and often run those nights in close succession. Each pass’s coverage (the pipe it is able to survey) depends on the wind direction, which is highly autocorrelated in time. This raises a concern about the standard practice: when passes are packed into a few nearby nights, does the wind direction vary enough for the survey to “see” the entire intended pipe, or do the repeated passes simply re-survey already-covered pipe and miss the rest? To investigate this, we use 3,992 real survey drives, each paired with a plume model that computes the fraction of the pipeline a drive could survey under a given night’s wind. For each drive we run a Monte-Carlo simulation, repeatedly drawing historical wind from other candidate nights and computing the resulting distribution of coverage under two deployment strategies: “windblind” surveys that drive on random nights within a week, and “wind-aware” surveys that choose nights for complementary wind directions. We find that six blind passes over three closely spaced nights cover a median of 81 % of pipe. The distribution has a low tail: a hard network, or any network caught in an unlucky week, is left with roughly a third of its pipe uncovered, and that tail deepens for the smaller leaks, while line-of-sight blockage (a heuristic but physically motivated treatment of obstructions’ impact on plume detectability) lowers coverage further still. Because the wind varies little over a single night, coverage is set by how many distinct nights a survey samples rather than by its pass count. Choosing those nights by the wind, using only conditions available within a one-to-two week window, removes the low tail and covers more with fewer passes: four passes over two chosen nights beat six driven blind, and three chosen nights reach about 93 %.
Language: English
Subfield: Global and Planetary Change
Field: Environmental Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7225613591
Cites:
Cited by:
Related to: