What drove the record surge in climate-heating methane?

New paper! One of the last projects from my PhD, on global methane emissions trends and seasonality, is now open-access at Atmospheric Chemistry and Physics. A well-mixed greenhouse gas second only to CO₂ in climate impact, methane concentrations surged in 2020 and 2021 for reasons which are debated in the literature.

Satellites and ground-based monitors can measure pollutants in the air but cannot infer their sources directly; pollutants may have blown downwind or reacted away. Knowledge of emissions comes instead from inventories, which are maps of pollutant sources along with their expected emissions magnitudes, constructed from information such as infrastructure plans and economic reports. Inventories thus link pollution to processes, and processes can be made targets of mitigation efforts. However, inventory construction takes several years and is subject to error. During my PhD, I developed CHEEREIO (https://cheere.io), open-source software which fuses observational data with the actionable information of inventories through Bayesian optimization.

In the paper, we apply CHEEREIO to global methane emissions trends and seasonality using satellite methane observations from the TROPOMI instrument. We attribute the 2020-2021 methane surge to wetland activity from flooding in eastern Africa, especially in the river-fed wetlands of South Sudan. This flooding is associated with a positive anomaly in a climate pattern called the Indian Ocean Dipole, which leads water to be transported west across the Indian Ocean, away from Australia and Indonesia and towards eastern Africa. The record fires in Australia in 2020 were, in effect, a climate twin of the record flooding in Africa. This kind of anomaly is expected to increase with warming — an unexpected climate feedback, because it leads to more climate-warming methane to release from newly flooded wetlands. We find that wetland models perform poorly in the tropics, but that remote sensing data can both correct this and offer near-real-time insights on climate feedbacks.

How did we do this attribution? We used a very cool satellite instrument called GRACE, which looks for anomalies in the Earth’s gravitational field. The leading short-term driver of these gravitational anomalies is water storage, because water is so heavy. We found that the traditional approach, which used existing physical wetland models to attribute methane to particular sources and sectors, seemed to be missing many key events in the river-fed wetlands of eastern Africa. We think this is because the record rainfall which drove these floods actually took place quite far from the wetlands themselves, and were transported by swollen river flows which are not adequately captured in models. As you can see in the figure below, the gravity-based measure of water content (in purple) in the northern tropics aligns with what CHEEREIO thinks methane emissions are (in solid black). The dotted and dashed lines show the methane seasonality that leading wetland models expect, putting the seasonal peak earlier in the year in part because they seem to be missing the late-year flooding in river-fed wetlands. Much more of this sort of analysis is in the paper itself.

An aside: you may have heard about the “methane bomb” from melting permafrost. So far, this particular feedback has yet to materialize, in part because microbes called methanotrophs seem to eat much of the stored methane before it reaches the atmosphere. There is an emerging consensus that tropical wetlands are acting as a key methane feedback, but their role wasn’t recognized until very recently.

Read the paper for free here! And stay tuned for more research using CHEEREIO coming soon from other groups around the world, including tracking global wildfires and monitoring greenhouse gas emission trends in China.

I’ll be back soon with a short description of the another paper I led, which was also recently published.