Point Source Satellite Constellations Can Track Diffuse Tropical Methane Through Cloud Gaps

A comprehensive theoretical case study published in Atmospheric Measurement Techniques demonstrates that satellite instruments originally designed to detect localized industrial methane plumes can be effectively repurposed to quantify large-scale diffuse emissions across tropical Africa. Using observing system simulation experiments, researchers evaluated the Near Infrared Multispectral Camera for Atmospheric Methane, known as NIMCAM, an innovative satellite instrument concept designed to record atmospheric methane at a high spatial resolution of 60 meters. Persistent cloud cover and elevated aerosol loading from seasonal biomass burning frequently obscure spaceborne greenhouse gas observations across equatorial regions, severely limiting the usable data yield of traditional global survey satellites like TROPOMI. The new modeling demonstrates that by viewing the surface through the fine clearings between fragmented clouds and smoke layers, high-resolution instruments substantially increase clear-sky data retrieval rates. When aggregated into regional modeling grid cells, these dense pixel collections form super-observations capable of reducing regional methane flux uncertainties far beyond the capabilities of traditional broad-footprint sensors.

The research addresses a vital challenge in global climate science, as the atmospheric growth rate of methane has surged to unprecedented record levels over the past decade. This acceleration is driven by expanding natural tropical wetland emissions alongside unresolved variations in the atmospheric hydroxyl radical sink and growing anthropogenic contributions from agriculture, fossil fuel extraction, and waste management. In tropical Africa, dynamic wetland complexes like the Sudd and Congo Basin generate major seasonal pulses of biogenic methane that strongly influence the global atmosphere. However, establishing and maintaining long-term ground monitoring networks across these remote, politically complex regions presents severe logistical obstacles. Consequently, researchers rely almost exclusively on satellite remote sensing to track shifting regional emissions. Traditional survey instruments, such as the Japanese Greenhouse gases Observing Satellite and the European TROPOMI sensor, operate with coarse spatial footprints spanning several kilometers. During equatorial rainy seasons and intense agricultural burning periods, extensive cloud cover and thick aerosols obscure the landscape, creating persistent blind spots over key emitting wetlands.

To overcome these critical observational barriers, the study evaluates NIMCAM, an innovative Earth observation mission concept under active development at the University of Edinburgh since 2017. Originally tailored to detect and quantify localized fugitive point sources, the platform features optical and orbital characteristics that enable broader environmental monitoring across large geographic domains. Flying in a sun-synchronous low Earth orbit at an altitude of approximately 600 kilometers with an inclination of 97.8 degrees, NIMCAM uses a fixed nadir-viewing design that completely avoids complex satellite retasking maneuvers. Its optical payload features a compact telescope with a 135-millimeter primary aperture, a 2.9-degree cross-track field of view, and an operational swath width of approximately 30 kilometers. A commercial Raptor Photonics indium gallium arsenide detector array of 640 by 512 pixels delivers a fine 60-meter ground pixel resolution. Each satellite operates on a 10-day orbital repeat cycle, providing weekly revisit capabilities across most land locations. The authors modeled three operational deployment scenarios: a single in-orbit demonstrator, a three-satellite constellation, and a five-satellite constellation providing near-complete global coverage within each 10-day repeat cycle.

The core instrument architecture of NIMCAM relies on an efficient spectral filtering scheme that measures methane absorption while maintaining high radiometric performance. The imager incorporates narrow bandpass interference filters located near 1640 nanometers in the shortwave infrared, each having a 1.0-nanometer bandwidth. An on-band channel at 1637.7 nanometers targets methane absorption, bracketed by two off-band channels at 1636.6 and 1638.7 nanometers that fit background surface reflectance and confounding gases. To extend this system to diffuse emissions, where enhancements approach the instrument noise floor, researchers added a carbon dioxide proxy channel at 1610 nanometers. Because carbon dioxide is well-mixed across localized scenes, its observed spatial variability reflects common radiative transfer perturbations, such as aerosol scattering, thin cloud cover, and surface albedo variations. Normalizing the methane retrieval with the carbon dioxide signal cancels shared radiative transfer effects, isolating true methane variations.

To quantify observational availability under realistic tropical meteorology, the authors analyzed cloud and aerosol datasets across Africa between 10 degrees south and 20 degrees north. Cloud cover was evaluated using Sentinel-2 cloud probability data resampled to the 60-meter NIMCAM grid, enforcing a strict 10 percent cloud probability threshold. Diurnal cloud checks with GEOS Forward Processing data confirmed that observations at the 10:30 local solar time crossing provide representative daytime sampling across the region. Atmospheric aerosol loading was characterized using Moderate Resolution Imaging Spectroradiometer green-band aerosol optical depth at 1-kilometer resolution. These data were downscaled to 60 meters using a Gaussian cumulative distribution function that estimated sub-grid variability from adjacent cells. The authors evaluated quality cutoffs at aerosol optical depths of 0.2 and a conservative threshold of 0.1, reflecting the stringent filtering needed for precise greenhouse gas retrievals.

Evaluating data coverage across March, April, October, and November 2019 revealed significant differences between fine-scale imaging and coarse-resolution survey platforms. Under all-sky conditions, TROPOMI achieves daily continental coverage due to its 2600-kilometer swath, whereas the NIMCAM demonstrator, three-satellite, and five-satellite constellations cover 18 percent, 54 percent, and 89 percent of the domain over ten days, respectively. However, cloud and aerosol filtering substantially reduces usable data. In October, the cloudiest month studied, average cloud cover reached 77 percent. When applying the strict filter of cloud probability under 10 percent and aerosol optical depth below 0.1, TROPOMI clear-sky coverage fell to 4 percent across March, April, and October. In contrast, NIMCAM’s 60-meter pixels exploited small clearings between clouds and smoke plumes. A five-satellite constellation retained 39 percent coverage in April and 42 percent in October, providing up to six times more valid grid boxes than TROPOMI.

The study examined how aggregating dense arrays of high-resolution pixels into model grid cells mitigates the higher random noise of compact multispectral instruments. Individual NIMCAM retrievals have an estimated random error of 40 parts per billion for solar zenith angles under 40 degrees, compared to 1.1 parts per billion for TROPOMI. However, when mapped onto a 0.25-degree by 0.3125-degree grid over ten days, up to 273,780 NIMCAM pixels can contribute to a single super-observation, compared to at most 200 measurements from TROPOMI. Assuming independent random errors that scale with the inverse square root of sample count alongside a 5 parts per billion systematic floor, the baseline clear-sky error ratio of NIMCAM to TROPOMI super-observations is 2.31 for the demonstrator, 1.33 for three satellites, and 1.04 for five satellites. Under strict cloud and aerosol filtering, TROPOMI suffers rapid data loss, allowing roughly 75 percent of five-satellite NIMCAM super-observations to achieve lower random errors than TROPOMI.

To assess regional flux improvements, the authors conducted observing system simulation experiments using the GEOS-Chem nested atmospheric chemistry transport model paired with a Local Ensemble Transform Kalman Filter. The model simulated methane transport at 0.25-degree by 0.3125-degree resolution for April and October 2019, corresponding to the equatorial long and short rainy seasons. Prior emissions incorporated inventories for wetlands, biomass burning, termites, geological seeps, and anthropogenic sectors, balanced against an OH oxidation sink. Synthetic observations were generated by sampling simulated atmospheric fields derived from GOSAT-constrained fluxes, incorporating instrument averaging kernels and measurement noise. An ensemble of 91 surface flux perturbations with an assumed 100-kilometer error correlation length was assimilated to evaluate how effectively NIMCAM and TROPOMI observations could recover true surface fluxes. Assimilating aggregated super-observations allowed the researchers to isolate observational sampling density from model transport mechanics.

The assimilation experiments demonstrated substantial gains in flux estimation accuracy from high-resolution observations, particularly in regions where persistent cloud decks blinded coarse sensors. In April 2019, prior surface emissions had a root-mean-square error of 5.35 grams per square meter per year against the assumed truth. Assimilating TROPOMI reduced this error to 4.70 grams per square meter per year, a 12.1 percent reduction. In comparison, a single NIMCAM demonstrator reduced the error to 3.78 grams per square meter per year, achieving a 29.3 percent reduction. Three- and five-satellite NIMCAM constellations lowered flux errors to 3.50 and 3.41 grams per square meter per year, yielding reductions of 34.3 percent and 36.9 percent. In October, prior error stood at 5.82 grams per square meter per year. Assimilation reduced errors to 5.02 for TROPOMI, 3.66 for the demonstrator, 3.52 for three satellites, and 3.48 for five satellites, confirming that additional satellites provide valuable constraints during cloudy seasons.

Despite these positive findings, the authors emphasize important technical caveats and operational considerations within their inversion setup. The identical-twin framework used GEOS-Chem for both data generation and flux inversion, excluding systematic transport errors that occur in real atmospheric modeling. Furthermore, the aggregation assumed uncorrelated retrieval errors; in real flights, spatially correlated errors from surface reflectance and aerosol variations would diminish effective sample sizes, rendering the reported uncertainty reductions an upper bound on performance. Persistent cloud cover over the Congo Basin also left residual flux errors unconstrained across all configurations. To mitigate systematic instrument biases, which can reach 200 parts per billion before correction, mission plans include routine solar calibrations and validation against the Total Carbon Column Observing Network to achieve residual biases within 5 parts per billion. The authors conclude that high-resolution point-source imagers working alongside global survey mappers offer an effective strategy to resolve diffuse methane emissions in cloudy tropical environments.

Subject of Research: Theoretical evaluation of high-resolution satellite instruments for estimating tropical diffuse methane emissions

Article Title: Potential of point source imaging satellite instruments to infer diffuse methane emissions: a theoretical case study of the Near-Infrared Multispectral Camera (NIMCAM)

Article References: Schooling, C. N., Feng, L., Woodwark, A. J. P., Arnstein, A., Green, P. D., & Palmer, P. I. (2026). Potential of point source imaging satellite instruments to infer diffuse methane emissions: a theoretical case study of the Near-Infrared Multispectral Camera (NIMCAM). Atmospheric Measurement Techniques, 19(21), 7279-7303. https://doi.org/10.5194/amt-19-7279-2026

Image Credits: AI Generated

DOI: 10.5194/amt-19-7279-2026

Keywords: atmospheric methane, NIMCAM, TROPOMI, Observing System Simulation Experiment, GEOS-Chem, diffuse emissions, tropical Africa, cloud masking, aerosol optical depth, ensemble Kalman filter, super observations, near infrared spectroscopy, wetland emissions, Sentinel-2, MODIS, uncertainty reduction, remote sensing