Space-Time Analysis of Systematic Errors and Biases in SMAP SSS
[27-Aug-2018] Melnichenko, O., Hacker, P., and Meissner, T.
Presented at the 2018 Ocean Salinity Science Team and Salinity Continuity Processing Meeting
The Soil Moisture Active Passive (SMAP) satellite based version V2.0 SSS data delivered by Remote Sensing Systems (RSS) are evaluated against in situ
salinity from Argo buoy measurements. To assess the effect of smoothing, two versions of the data have been evaluated- the so-called 40-km and 70-km resolution products. The difference statistics are computed by comparing Argo buoy measurements for a given week with SSS values at the same locations obtained by interpolation of the corresponding Level-3 SSS maps. The results can be summarized as follows.
The time-mean global bias is around -0.05 psu. There are time periods, however, when the global average biases are significant, reaching magnitudes greater than 0.1 psu. The geographical distribution of the time-mean bias exhibits clear zonality with large positive biases (up to 0.3 psu) in the tropics and negative biases at mid and high latitudes. The spatial variability in the bias field is also significant. The largest positive biases with magnitudes greater than 0.3 psu are observed in the western tropical Pacific and are likely a manifestation of the SST-dependent error. The largest negative biases are observed along the continental boundaries and appear to be related to the low level RFI contamination from adjacent land areas. Resampling from 40-km resolution to 70-km resolution has had no effect on the bias fields.
The time-mean RMSD for the 8-day Level-3 40-km SSS product is 0.38 psu. The geographical distribution of the RMSD shows that over most of the ocean the RMSD between the weekly SSS maps and collocated buoy data exceeds 0.2 psu. The largest RMSD (up to 0.5 psu) are observed in the subpolar North Pacific, the subpolar North Atlantic, and in the Southern Ocean. Resampling from 40-km resolution to 70-km resolution has had moderate effect on the RMSD reducing it by about 15%.