12 March 2025 in Geoscientific Model Development
There has been a proliferation of dense observing systems to monitor greenhouse gas (GHG) concentrations over the past decade. Estimating emissions with these observations is often done using an atmospheric transport model to characterize the source–receptor relationship, which is commonly termed the measurement “footprint”. Computing and storing footprints using full-physics models is becoming expensive due to the requirement to simulate atmospheric transport at high resolution. We present the development of FootNet, a deep-learning emulator of footprints at the kilometer scale. We train and evaluate the emulator using footprints simulated with a Lagrangian particle dispersion model (LPDM). FootNet predicts the magnitudes and extents of footprints in near real time with high fidelity. We identify the relative importance of input variables of FootNet for improving the interpretability of the model. Surface winds and a precomputed Gaussian plume from the receptor are identified as the most important variables for footprint emulation. The FootNet emulator developed here may help address the computational bottleneck of flux inversions using dense observations.
25 October 2024 in Geophysical Research Letters
NASA's Orbiting Carbon Observatories (OCO‐2 and OCO‐3) provide measurements of column‐averaged carbon dioxide concentrations (XCO2) with sufficient spatial resolution and precision to constrain bottom‐up estimates of CO2 fluxes at regional scales. We use Bayesian inversion methods assimilating satellite retrievals to improve estimates of CO2 fluxes in the South Coast Air Basin (SoCAB) which surrounds Los Angeles, and in the San Francisco Bay Area Air Basin (SFBA). We study 2020 to understand the impact of the COVID‐19 lockdowns and an active wildfire season. Our results indicated that a 50% (30%) reduction in CO2 emissions relative to 2015 during the COVID‐19 lockdown period was consistent with OCO measurements for SFBA (SoCAB). We find that posterior wildfire emissions differed significantly from the prior at the scale of individual wildfires, though with large uncertainties, and that wildfire emissions in SFBA are significant, attributing 72% of the region's CO2 emissions during August 2020 to wildfires.
21 May 2025 in Atmospheric Chemistry and Physics
Quantifying greenhouse gas (GHG) emissions is critically important for projecting future climate and assessing the impact of environmental policy. Estimating GHG emissions using atmospheric observations is typically done using source–receptor relationships (i.e., “footprints”). Constructing these footprints can be computationally expensive and is rapidly becoming a computational bottleneck for studying GHG fluxes at high spatio-temporal resolution using dense observations. Here, we demonstrate a computationally efficient GHG flux inversion framework using a machine learning emulator for atmospheric transport (FootNet) as a surrogate for the full-physics model. The footprints generated by FootNet are at approximately 1 km resolution. We update the architecture of the deep-learning model to improve the performance in a GHG flux inversion. We find that the posterior fluxes estimated with FootNet footprints are in good agreement with the posterior fluxes estimated with STILT footprints. We observe that the more simplistic representation of transport in the machine learning model helps to mitigate transport errors. This flux inversion using a machine learning surrogate model requires only meteorological data, GHG measurements, and prior fluxes. Constructing footprints using FootNet is 650 times faster than the full-physics atmospheric transport model on similar hardware. This speedup allows for the computation of footprints “on the fly” during the GHG flux inversion (i.e., computed as needed, rather than archiving for future use) and makes near-real-time emission monitoring computationally possible. This work alleviates a major computational bottleneck with inferring GHG fluxes with next-generation dense observing systems.
8 January 2026 in Atmospheric Chemistry and Physics
Accurately estimating greenhouse gas (GHG) emissions from atmospheric observations requires resolving the upwind influence of measurements via atmospheric transport and dispersion models. However, the computational demands of physics-based models limit the scalability of flux inversions, particularly for dense in situ and satellite-based observations. Here, we present FootNet v3, a deep-learning emulator of atmospheric transport based on a U-Net architecture, which improves generalization and inversion fidelity over prior U-Net-based models. FootNet v3 is trained on 500 000 footprint examples across the contiguous United States. It predicts surface and column-averaged source-receptor relationships at kilometer-scale resolution and operates 650 × faster than traditional Lagrangian models. Critically, FootNet learns the underlying physical relationship between meteorology and atmospheric transport. We show that it produces consistent source-receptor relationships when driven by GFS meteorology, even though it was trained with HRRR inputs. FootNet generalizes to unseen regions and meteorological regimes, enabling accurate flux inversions in domains withheld during training. Case studies using GHG measurements in the San Francisco Bay Area and Barnett Shale show that FootNet matches or exceeds the performance of physics-based models when used in a flux inversion and evaluated against independent GHG observations. This is achieved despite FootNet having never seen meteorological inputs from Northern California or North Texas. Feature importance testing identifies physically meaningful drivers that are consistent across both surface and column models. These findings show that machine learning models can learn the underlying physical relationships governing atmospheric transport, allowing them to extrapolate to out-of-sample scenarios and support real-time, high-resolution GHG flux estimation in novel domains without the need for retraining or precomputed footprint libraries.
26 September 2024 in EGUsphere
Quantifying greenhouse gas (GHG) emissions is critically important for projecting future climate and assessing the impact of environmental policy. Estimating GHG emissions using atmospheric observations is typically done using source-receptor relationships (i.e., "footprints''). Constructing these footprints can be computationally expensive and is rapidly becoming a computational bottleneck for studying GHG fluxes at high spatio-temporal resolution using dense observations. Here we demonstrate a computationally efficient GHG flux inversion framework using a machine learning emulator for atmospheric transport (FootNet) as a surrogate for the full-physics model. The footprints generated by FootNet are at approximately one-kilometer resolution. We update the architecture of the deep-learning model to improve the performance in a GHG flux inversion. Paradoxically, the updated FootNet model out-performs the full-physics model when used in a flux inversion and compared against independent observations. This improved performance is likely because atmospheric transport simulated with a full-physics transport model is not necessarily more accurate. The more simplistic representation of transport in the machine learning model helps to mitigate transport errors. This flux inversion using a machine learning surrogate model only requires meteorological data, GHG measurements, and prior fluxes. Constructing footprints using FootNet is 650× faster than the full-physics atmospheric transport model on similar hardware. This speedup allows for computation of footprints "on-the-fly'' during the GHG flux inversion (i.e., computed as needed, rather than archiving for future use) and makes near-real-time emission monitoring computationally possible. This work alleviates a major computational bottleneck with inferring GHG fluxes with next generation dense observing systems.
24 July 2025 in EGUsphere
Accurately estimating greenhouse gas (GHG) emissions from atmospheric observations requires resolving the upwind influence of measurements via atmospheric transport models. However, the computational demands of full-physics models limit the scalability of flux inversions, particularly for dense in situ and satellite-based observations. Here, we present FootNet v3, a deep-learning emulator of atmospheric transport based on a U-Net++ architecture, which improves generalization and inversion fidelity over prior U-Net-based models. FootNet v3 is trained on 500,000 pseudo-observations across the contiguous United States. It predicts surface and column-averaged source-receptor relationships at kilometer-scale resolution and operates 650x faster than traditional Lagrangian models. Critically, FootNet learns the underlying physical relationship between meteorology and atmospheric transport. We show that it accurately predicts source-receptor relationships when driven by GFS meteorology, despite being trained on HRRR data. FootNet generalizes to unseen regions and meteorological regimes, enabling accurate flux inversions in domains withheld during training. Case studies using GHG measurements in the San Francisco Bay Area and Barnett Shale show that FootNet matches or exceeds the performance of full-physics models when evaluated against independent GHG observations. This is achieved despite FootNet having never seen meteorological inputs from Northern California or North Texas. Feature importance testing identifies physically meaningful drivers that are consistent across both surface and column models. These findings show that machine learning models can learn the physics governing atmospheric transport, allowing them to extrapolate to out-of-sample scenarios and support real-time, high-resolution GHG flux estimation in novel domains without the need for retraining or precomputed footprint libraries.