Error-Uncertainty Analysis of the TOMS Version 9 Total Ozone Record from the OMI and OMPS Mapping Instruments

We report on our error and uncertainty analysis of the TOMS Version 9 (V9) total ozone mapping records from the Ozone Monitoring Instrument (OMI) and the Ozone Mapping and Profiler Suite (OMPS) Nadir Mapper (NM) on Suomi NPP. Our results show the most important errors come from uncorrected instrument errors, high loads of absorbing aerosols, uncertainty in profile shape, and radiative transfer errors at large view angles. With a modern array of observations and model data, along with error sensitivity functions provided in the V9 data product for each retrieval, we can conduct a more careful and detailed evaluation of errors than has been possible before. We focus on errors that produce the drift, bias, and seasonal cycles seen in zonal averages, but we also examine pseudo-random errors introduced by uncertainties in ozone profile shape at shorter spatial and time scales, where data are not averaged broadly. These errors affect the retrievals particularly at high slant paths, which occur not only at high latitudes in LEO orbit, but also near sunrise and sunset at GEO. Finally, we explore possibilities to reduce our errors using enhanced a priori information that is more appropriate for making daily maps, and by applying additional corrections for certain types of instrument error.