Item

Abstract

This study investigates the use of total column CH4 (XCH4) retrievals from the SCIAMACHY
satellite instrument for quantifying large scale emissions of methane. A unique
data set from SCIAMACHY is available spanning almost a decade of measurements,
covering a period 5 when the global CH4 growth rate showed a marked transition from
stable to increasing mixing ratios. The TM5 4DVAR inverse modelling system has been
used to infer CH4 emissions from a combination of satellite and surface measurements
for the period 2003–2010. In contrast to earlier inverse modelling studies, the SCIAMACHY
retrievals have been corrected for systematic errors using the TCCON network
10 of ground based Fourier transform spectrometers. The aim is to further investigate the
role of bias correction of satellite data in inversions. Methods for bias correction are
discussed, and the sensitivity of the optimized emissions to alternative bias correction
functions is quantified. It is found that the use of SCIAMACHY retrievals in TM5 4DVAR
increases the estimated inter-annual variability of large-scale fluxes by 22% compared
15 with the use of only surface observations. The difference in global methane emissions
between two year periods before and after July 2006 is estimated at 27–35 Tgyr−1. The
use of SCIAMACHY retrievals causes a shift in the emissions from the extra-tropics to
the tropics of 50±25 Tgyr−1. The large uncertainty in this value arises from the uncertainty
in the bias correction functions. Using measurements from the HIPPO and
20 BARCA aircraft campaigns, we show that systematic errors are a main factor limiting
the performance of the inversions. To further constrain tropical emissions of methane
using current and future satellite missions, extended validation capabilities in the tropics are of critical importance.