Abstract
In the recent work of Liu and Minnett (2016), we estimated the sampling errors in Moderate Resolution Imaging Spectroradiometer (MODIS) Sea-Surface Temperatures (SSTs) due to clouds and other causes, and characterized the global error dependence on the variability of clouds and SST. Here we report sampling error sensitivity to the choice of reference field and the error variation when data from a different year are used. We also developed an empirical model to parameterize sampling errors. Our sensitivity tests show that the sampling error quantification method developed is robust and can reveal the consequences of missing infrared SST observations primarily due to clouds. Since the previously found pronounced negative sampling errors along the Tropical Instability Waves are largely dependent on the SST gradients, here these regional sampling errors are quantified using data from an El Niño year, confirming that the weakened meridional SST gradient due to El Niño can reduce the negative sampling errors. Furthermore, the climatology-derived sampling errors are found to be a primary component that can be utilized to estimate and parameterize the sampling errors, especially for the spatial sampling errors. For the temporal sampling errors, good estimates are obtained especially in the high latitudes and stratocumulus regions, by incorporating an empirical model proposed in this study and the previously found sampling error dependence.
•Global MUR SST fields and HYCOM-NCODA SST reanalysis have marked differences.•The global sampling error generated from MUR and HYCOM SSTs are very similar.•El Niño events reduce the negative sampling errors in TIW region.•The SST seasonality can only explain <30% sampling errors in the monthly SST fields.•The sampling error parameterization based on previous results shows promise.
