The ozone climate change initiative: Comparison of four Level-2 processors for the Michelson Interferometer for Passive Atmospheric Sounding (MIPAS)

摘要: Abstract The MIPAS spectrometer onboard the Envisat platform observed infrared emission from Earth's limb between 2002 and 2012. It recorded high-resolution spectra during day night, pole to 6 70 km altitude in nominal measurement mode or up 170 km special modes, producing daily more than 1000 vertical profiles of various trace gases. operational Level-2 data are processed by ESA/DLR but there exist three other, independent research processors that hosted ISAC-CNR/University Bologna, Oxford University, KIT IMK/IAA. All four rely on same Level-1b provided ESA their retrieval schemes differ. As part ESA's Ozone Climate Change Initiative project, an intercomparison took place, which ozone retrieved these measurements were compared for 2007 2008. We present results this comparison exercise, consisted five parts: information content study averaging kernels, zonal seasonal means spreads, a determination biases through ozonesonde lidar measurements, other satellite records (bias estimation precision assessment with respect ACE-FTS Aura-MLS data), geophysical validation error bars using MIPAS–MIPAS collocations. demonstrate similar performance. use ESA, apply global fits, microwindows instead full spectrum. main differences processing include choice microwindows, regularization approach, treatment negative values, cloud detection threshold. different lead trade-off noise resolution, without clear average advantage any particular set. resolution is typically 3–5 km single profile about 2–3%. In middle upper stratosphere, at 25–45 km, all clearly show high bias 2 5% relative reference instruments. similarity structure magnitude among sets indicates most likely linked AB band. intercomparisons furthermore dataset, onset starts somewhat higher (only above 35 km) datasets. This due restrictive band processor, comes cost coarser near volume mixing ratio (vmr) peak. troposphere, algorithms suppress values iterative process produce larger positive algorithm does not follow such strategy. Our conclusion each instrument. very instrument-related.