Atmospheric methane is the second most important long-lived greenhouse gas and its atmospheric abundance has been increasing since pre-industrial times. But over the past 15 years its growth rate has undergone large variations and the reasons for this remain a mystery.

Methane, along with CO2, has the capacity to change the global climate through the greenhouse effect. Since the pre-industrial era human activities have accelerated the growth of methane in the atmosphere. In more recent decades atmospheric growth has become less smooth, for example, between 1999 and 2006 it stopped almost completely only to resume again in 2007. The reasons for this recent intermittent growth are not yet understood.

It is not only human activities that produce methane; in fact, approximately half of all emissions are natural in origin. These include emissions from wildfires, oceans and even termites. The largest single source of methane to the atmosphere is wetlands (~30%); more specifically from microorganisms, called methanogens.  So changes to the methane emissions from wetlands could play an important role.

To help understand whether wetlands might be causing the recent changes in atmospheric methane growth our group used computer models, surface observations and satellite data. Computer simulations using the UK land surface model provided an estimate of global methane emissions from wetlands from 1993 to 2014. From this information alone we could see where and when emissions might have changed, and whether these might have influenced the atmospheric growth.

But there are two complications with this approach. Firstly, emission changes are not the same as changes in atmospheric growth; they only tell us how much methane is going in to the atmosphere and not what is happening to it once it is there. Secondly, computer models are not perfect and therefore need to be evaluated against observations.

To overcome the first issue we used an atmospheric computer model. This simulated methane in the atmosphere accounting for all types of emissions and the loss through chemical reactions, to help us understand changes to atmospheric methane growth.
To evaluate the models we performed comparisons with observations from surface stations and from the Japan Aerospace Exploration Agency (JAXA) satellite, GOSAT. GOSAT offers global coverage and provides atmospheric vertical column information. Our atmospheric model simulations compared well with observations, suggesting that the emission and loss estimates in the model were reasonable (see images below).