Strategy for high-accuracy-and-precision retrieval of atmospheric methane from the mid-infrared FTIR network
We present a strategy (MIR-GBM v1.0) for the retrieval of column-averageddry-air mole fractions of methane (XCH4) with a precision <0.3 %(1-¿ diurnal variation, 7-min integration) and a seasonal bias<0.14 % from mid-infrared ground-based solar FTIR measurements of theNetwork for the Detection of Atmospheric Composition Change (NDACC,comprising 22 FTIR stations). This makes NDACC methane data useful forsatellite validation and for the inversion of regional-scale sources andsinks in addition to long-term trend analysis. Such retrievals complementthe high accuracy and precision near-infrared observations of the youngerTotal Carbon Column Observing Network (TCCON) with time series dating back15 yr or so before TCCON operations began. MIR-GBM v1.0 is using HITRAN 2000 (including the 2001 update release) and 3spectral micro windows (2613.70-2615.40 cm-1, 2835.50-2835.80 cm-1,2921.00-2921.60 cm-1). A first-order Tikhonov constraintis applied to the state vector given in units of per cent of volume mixingratio. It is tuned to achieve minimum diurnal variation without dampingseasonality. Final quality selection of the retrievals uses a threshold forthe ratio of root-mean-square spectral residuals and information content(<0.15 %). Column-averaged dry-air mole fractions are calculated usingthe retrieved methane profiles and four-times-dailypressure-temperature- humidity profiles from National Center forEnvironmental Prediction (NCEP) interpolated to the time of measurement. MIR-GBM v1.0 is the optimum of 24 tested retrieval strategies (8 differentspectral micro-window selections, 3 spectroscopic line lists: HITRAN 2000,2004, 2008). Dominant errors of the non-optimum retrieval strategies areHDO/H2O-CH4 interference errors (seasonal bias up to ¿4 %).Therefore interference errors have been quantified at 3 test sitescovering clear-sky integrated water vapor levels representative for allNDACC sites (Wollongong maximum = 44.9 mm, Garmisch mean = 14.9 mm,Zugspitze minimum = 0.2 mm). The same quality ranking of the 24 strategieswas found for all 3 test sites with one optimum, i.e., MIR-GBM v1.0. Seasonality of XCH4 above the Zugspitze (47degrees N) shows a minus-sineshape with a minimum in March/April, a maximum in September, and anamplitude of 16.3 + 2.9 ppb (1.0 + 0.2 %). This agrees very wellwith newest-version scientific retrievals from SCIAMACHY (WFM-DOAS v2.0). A conclusion from this paper is that improved spectroscopic parameters forCH 4, HDO, and H2O in the 2613-2922 cm-1 spectral domainare urgently needed. If such become available with sufficient accuracy, atleast two more spectral micro windows could be utilized leading to anotherimprovement in precision. The absolute inter-calibration of NDACC MIR-GBMv1.0 XCH4 to TCCON data is subject of ongoing work. C Author(s) 2011.