Detecting changes in Arctic methane emissions: limitations of the inter-polar difference of atmospheric mole fractions

Abstract

Abstract. We consider the utility of the annual inter-polar difference (IPD) as a metric for changes in Arctic emissions of methane (CH4). The IPD has been previously defined as the difference between weighted annual means of CH4 mole fraction data collected at stations from the two polar regions (defined as latitudes poleward of 53∘ N and 53∘ S, respectively). This subtraction approach (IPD) implicitly assumes that extra-polar CH4 emissions arrive within the same calendar year at both poles. We show using a continuous version of the IPD that the metric includes not only changes in Arctic emissions but also terms that represent atmospheric transport of air masses from lower latitudes to the polar regions. We show the importance of these atmospheric transport terms in understanding the IPD using idealized numerical experiments with the TM5 global 3-D atmospheric chemistry transport model that is run from 1980 to 2010. A northern mid-latitude pulse in January 1990, which increases prior emission distributions, arrives at the Arctic with a higher mole fraction and ≃12 months earlier than at the Antarctic. The perturbation at the poles subsequently decays with an e-folding lifetime of ≃4 years. A similarly timed pulse emitted from the tropics arrives with a higher value at the Antarctic ≃11 months earlier than at the Arctic. This perturbation decays with an e-folding lifetime of ≃7 years. These simulations demonstrate that the assumption of symmetric transport of extra-polar emissions to the poles is not realistic, resulting in considerable IPD variations due to variations in emissions and atmospheric transport. We assess how well the annual IPD can detect a constant annual growth rate of Arctic emissions for three scenarios, 0.5 %, 1 %, and 2 %, superimposed on signals from lower latitudes, including random noise. We find that it can take up to 16 years to detect the smallest prescribed trend in Arctic emissions at the 95 % confidence level. Scenarios with higher, but likely unrealistic, growth in Arctic emissions are detected in less than a decade. We argue that a more reliable measurement-driven approach would require data collected from all latitudes, emphasizing the importance of maintaining a global monitoring network to observe decadal changes in atmospheric greenhouse gases.