Assessing Suomi-NPP OMPS Nadir Mapper Reflectance Accuracy Using SNO Observations with GOME-2

The Ozone Mapping Profiler Suite (OMPS) onboard Suomi-NPP (S-NPP) and Joint Polar Satellite System (JPSS) series of satellites has two nadir-viewing instruments: Nadir Mapper (NM) and Nadir Profiler (NP). Reflectance calculated from their Sensor Data Record (SDR) data are used to retrieve a series of ozone environmental data records (EDR) products such as total ozone atmospheric content and vertical atmospheric ozone profile. The quality assurance of the reflectance data is thus essential for retrievals of ozone EDR products. S-NPP OMPS data reached provisional maturity status in March 2013, so it is important to provide a new assessment of the data quality since the review. One of approaches to assess the SDR data quality is to use inter-sensor comparison based on simultaneous nadir overpass (SNO) observations with another satellite instrument. Recently, the STAR Integrated Calibration/Validation System (ICVS) is developing a new inter-sensor comparison function for independently assessing long term calibration accuracy of S-NPP NM reflectance data by using SNO observation differences between the NM and the Global Ozone Monitoring Experiment-2(GOME-2) on board METOP-B satellite. We use METOP-B GOME-2 to produce SNO pixel pairs since METOP-B became European Organization for the Exploitation of Meteorological Satellites (EUMETSAT) 's prime polar-orbiting satellite since 24 April 2013.

In this study, we will demonstrate our preliminary results about SNO reflectance difference time series since 2013 between the SNPP OMPS/NM and the METOP-B GOME-2 observations. Firstly, we analyze the impacts of two criteria on selecting SNO pairs to create an optimal quality control (QC) criterion for the SNO pairs. To do this, both the NM and GOME-2 data are processed to produce two different types of SNO reflectance pixel pairs: 1) Pixel center distance is less than 45km and observation time difference is less than 120 seconds, and 2) Pixel center distance is less than 30km, observation time difference is less than 120 seconds, reflectance standard deviation from OMPS pixel and surrounding 8 pixels is less than 2%. The second criteria ensures that the GOME-2 pixel is fully covered by the big cluster made up by 3x3 OMPS pixels and each cluster has less impacts from heterogeneous surface or atmosphere. For most of the SNO events, there are more than 90 pixel pairs satisfying the first criteria for each SNO event, while less than 30 pixel pairs satisfy the second criteria. Furthermore, we use the SNO difference time series from 2013 to 2019 under different QC criteria to assess the long term calibration accuracy of the OMPS/NM reflectance against the GOME-2 observations. Preliminary results show that OMPS/NM reflectance data are more stable compared to GOME-2 measurements. The mean relative bias steadily increases from 2013 to present but the rate or increase decreases with wavelength. In addition, for each SNO event, the relative bias does not show obvious dependency on reflectance. Additionally, standard deviation of their relative bias does not show obvious dependency on time and is within 2% for most of SNO events, which is more or less related to the second SNO criterion. More analyses about those features and the root cause for the increased SNO differences will be presented in this study. To the best of our knowledge, the long-term monitoring of OMPS SDR reflectance accuracy using METOP-B GOME-2 data has not been well documented. It is thus expected that those SNO difference time series will be accessible to a broad user community through the ICVS website, after we better understand the SNO difference features.