Detection of Long-Term Change in Methane Emissions Using Atmospheric Network Observations

Abstract:

The number of sites in the Arctic at which observations of atmospheric CH4 are collected has grown over the past several

decades. Some of these sites now have observation records that span several decades; from the early 1980s to present.

At the same time, Arctic temperatures have increased at double the rate of the global average increase. A recent comparison of

models of CH4 emissions from wetlands (the “WETCHIMP” study) found that many models predict increased emissions in

response to higher temperatures. Given the temperature sensitivity of the wetland emission models, and the observed

Arctic temperature increase, the change in annual CH4 emissions is likely small, however, the cumulative extra emissions

over this period may be at the detection level of the atmospheric network. Even so there is still no firm evidence from the

atmospheric network that emissions are changing.

In this study, we explore the sensitivity of the atmospheric network to changing emissions. How large do the emissions have to be

before they can be seen in the network observations? Can changes be detected by using spatial gradients of observed methane?

Can changes in the amplitude and phase of the annual cycle be used to detect increasing emissions during the growing season?

Furthermore, by using atmospheric assimilation techniques, network observations provide a strong constraint on total Arctic emissions.

Using results from a suite of atmospheric CH4 assimilations we show that total emissions from the Arctic are ~25 TgCH4/yr

(with a range from 18 to 29 TgCH4/yr). This is lower than many bottom-up analyses, and implies that emissions from Arctic lakes, the

Eastern Siberian Arctic Sea, wetlands and possible geologic sources cannot all be accommodated in the Arctic atmospheric

budget of CH4. Finally, we address the issue of how the observation network can be augmented to allow timely trend detection.