Variability Scales of Sea Surface Salinity

The variability of sea surface temperature (SST) and salinity (SSS) were investigated from the monthly World Ocean Atlas (WOA) Climatology. First, the zonal means and standard deviations were constructed for each 10° latitude bins from the climatology. Both the SST and SSS zonal means are nearly symmetric about the ITCZ between the hemispheres. There are differences in the zonal standard deviation however, with SST variability being nearly symmetric about the ITCZ and the SSS variability having a much smaller variability south of 20° S. It was presumed that this lack of variability in the SSS is due to the lack of data that went in to constructing the climatology in the southern hemisphere. Results from a global ocean general circulation model were used to do similar analysis to investigate the variability in the southern hemisphere. As part of the NASA Aquarius mission sea surface salinity forward model, the fully global Hybrid Coordinate Ocean Model (HYCOM) at 2° and 0.72° horizontal resolutions is set up to provide surface salinity and temperature inputs to simulate the brightness temperatures at the satellite. After an initial 30 year spin-up to near-dynamic equilibrium using COADS forcing, a 9 year simulation is conducted using the NCEP reanalysis forcing and monthly river discharge to produce three 3 year realizations of surface salinity that will cover low, medium and high solar activity periods. Maximum variability in the model SSS occurs primarily along the ITCZ region and in the regions of significant river flow. While these results are generally consistent with climatological estimates, the variability magnitude is slightly lower on the order of 0.2 psu except in the regions of strong river inputs and in Bay of Bengal. In general, the variability in the model is lower than in the climatology. However, the zonal means estimated from the model also show a smaller variability in the southern hemisphere for SSS. Additionally, multiple runs from the 2° model forced with different datasets show a very similar nature of SSS variability magnitude south of 20° S of less than 0.2 psu (mostly on the order of 0.1 psu). Therefore, our conclusion is that the smaller SSS variability in the southern hemisphere is not the artifact of data voids, although the lack of data may create some local variability due to mismatch in the time of acquisition. This issue is further investigated with the use of the Argo near-surface salinity data.