PlumeQuant: Uncertainty-aware consistency assessment of methane plume masks and emission-rate estimates

Research Context and Approach

Imaging spectrometers like EMIT provide critical data for monitoring methane point sources. These products typically include a plume mask, integrated mass enhancement (IME), plume length, and an emission-rate estimate. Because these values are physically and algorithmically linked, the plume mask—the spatial boundary defining the plume—is a primary driver of the final emission estimate. This paper investigates whether these distributed quantities are internally consistent and whether the plume boundary is uniquely constrained by the reported scalar values.

To address this, the authors developed PlumeQuant, a framework that recomputes plume quantities from distributed product components using transparent, standardized conventions. The researchers analyzed 63 EMIT-derived methane plume records from the Permian Basin and surrounding regions. They evaluated four different mask representations: the original reference mask, a transparent 'CM-like' dynamic-threshold mask, a genetic-algorithm (GA) ensemble, and optional expert-edited masks.

Key Findings

The study reveals that the reported scalar quantities (IME, plume length, and emission rate) do not uniquely constrain the plume boundary. By using a genetic-algorithm ensemble to search for all masks that satisfy the published IME and length, the authors found that substantially different, yet plausible, masks can yield the same final emission estimates. This ambiguity is most pronounced for weak or low-overlap plumes, where the 'high-confidence core' of the plume footprint covers only a small fraction (median 13%) of the total plausible area.

Despite this spatial ambiguity, the PlumeQuant 'CM-like' mask demonstrated high consistency with published products. It reproduced published IME with a median difference of +0.72% and emission rates with a mean absolute error of 6.98%. The tool successfully matched the published uncertainty scales, confirming that while the spatial boundaries are often ambiguous, the resulting scalar quantities are generally consistent with the provided metadata.

As methane monitoring moves toward operational use in regulatory and verification frameworks, users must be able to distinguish between physically meaningful data and artifacts of the processing chain. PlumeQuant provides a standardized, reproducible way to flag ambiguous or weak detections that require expert review. By making the 'equifinality' of plume masks explicit, this research highlights the need for transparency in how plume boundaries are delineated and how that uncertainty propagates into emission-rate estimates.