Biases and Correlations in the Co-sampled OMPS and NOAA Ground-based Ozone Network

The National Polar-orbiting Partnership satellite (S-NPP) was launched on October 28, 2011 and has been in the orbit for almost 4 years. Measurements from the Satellite have been used to derive a total ozone and profile product (i.e. OMPS NM/EDR/INCTO and NP/EDR/IMOPO). Several long-term NOAA Dobson stations were continuously used for validation of total column ozone (TCO) and ozone profile (PRO) since the NPP launch. Dobson data processing is currently available in an almost real time with the Dobson automation system upgrade designed by the Japanese Meteorological Agency at five NOAA stations. Three Dobson stations with WinDobson automation are collecting data since 2010/2011, and are located at MLO, Hawaii (tropics), Lauder, New Zealand (Southern Hemisphere middle latitudes), and in Boulder, CO (Northern Hemisphere middle latitudes). Two more stations were added recently, i.e. Observatoire de Haute-Provence, France (Northern Hemisphere middle latitudes, since May 2014), and Fairbanks, Alaska (Northern Hemisphere high latitudes, since April 2015). Dobson direct sun observations are used to derive the best Dobson ozone product (precision is better than 1 %). The TCO observations are typically taken three times a day, excluding overcast conditions and weekends when Zenith sky measurements are used to derive TCO. The number of direct sun (AD-pair) observations at Boulder, MLO and Lauder observatories over period from 2012 and until the middle of 2015 are 2261, 2441, and 1724 respectively. These same three stations also make measurements of the Umkehr effect, from which an ozone profile over the station is derived (630, 1271 and 600 respectively). Profile comparisons show some biases, most likely due to treatment of the stray light in both satellite and Dobson Umkehr data. Results of daily and monthly averaged profile will be shown for the co-located pairs of satellite over pass and ground-based data. The overpass satellite ozone profile (column) product is determined within 12 hours and within area of +/-2 degrees (+/- 0.5) in latitude and +/- 10 degrees (1/cos(lat*pi/180)) in longitude centered on station location. During the first stage of the comparisons the OOTCO and INTCO datasets continued to be modified and adjusted through calibration and algorithm changes. It was noticed that the average difference between OMPS and Dobson at MLO prior to middle of June 2013 was at +15 DU, while after that and until the end of 2013, it changed to ~4 DU, in 2014 the offset was increased to ~ + 8 DU. At the same time difference between OMI and Dobson at MLO was on average at 13 DU before, and remained at 11 DU after the change in OMPS output through the end of 2013. In 2014 difference between OMI and Dobson at MLO was reduced to ~ 8 DU, which is similar to the offset in the OMPS data. At MLO the correlation between Dobson and INTCO (OMI) is 0.88 (0.89) for the period between January 1 2012 and December 31, 2014, where mean Dobson, OMPS and OMI TCOs are 266, 271 and 275 DU. Similar correlation results are found for Boulder, CO (0.92 and 0.95 respectively). However, over Lauder station the Dobson TCO daily correlations are 0.98 and 0.97 with OMI and OMPS overpass ozone respectively. The OMPS and OMI data sets appear to merge now and have similar offsets relative to Dobson station records. There is a seasonal cycle in Dobson/satellite comparisons that is likely related to the temperature sensitivity of the absorption cross sections used in each data set and the magnitude of this error will be discussed. Results of comparisons between OMPS and other NOAA and NASA satellite ozone products will be also shown. The biases and patterns in the total ozone column and profiles for zonal means and matching orbits will be assessed.