Impact of Aquarius and SMAP Satellite Sea Surface Salinity Observations on Coupled El Niño/Southern Oscillation Forecasts

This study demonstrates the positive impact of including gridded Aquarius and Soil Moisture, Active/Passive (SMAP) sea surface salinity (SSS) into initialization of intermediate complexity coupled model forecasts for the tropical Indo-Pacific. An experiment that assimilates conventional ocean observations serves as the control. In a separate experiment, Aquarius and SMAP satellite SSS are additionally assimilated into the coupled model initialization. Analysis of the initialization differences with the control indicates that SSS assimilation causes a freshening and shallowing of the mixed layer depth near the equator and enhanced Kelvin wave amplitude. For each month from September 2011 to September 2017, 12-month-coupled ENSO forecasts are initialized from both the control and satellite SSS assimilation experiments. The experiment assimilating Aquarius and SMAP SSS significantly outperforms the control relative to observed NINO3.4 sea surface temperature anomalies. This work highlights the potential importance of inclusion of satellite SSS for improving the initialization of operational ENSO coupled forecasts. Plain Language Summary El Nino/Southern Oscillation (ENSO) has far reaching climatic impacts over the globe so extending useful ENSO forecasts would be of great benefit for society. In response, NASA has developed satellite technology to observe the global hydrological cycle by measuring ocean sea surface salinity (SSS) from space. SSS, combined with temperature, helps to identify density changes and associated mixing near the ocean surface. Here we show results of two intermediate complexity coupled experiments designed to highlight the positive impact of SSS on ENSO forecasts. In the control, we assimilate all conventional satellite and in situ oceanographic information including satellite altimetry (matching current operational data assimilation schemes) but exclude SSS. In the second experiment, we add satellite SSS to our assimilation suite. Air/sea coupled model retrospective forecasts are then initialized from these two experiments and they show that satellite SSS assimilation improves coupled forecasts. For all lead times, the experiment with SSS assimilation has better correlation and root-mean-square difference with the ENSO metric (i.e., NINO 3.4 observed sea surface temperature anomalies). Density changes associated with SSS assimilation shoal the mixed layer near the equator and enhance the impact of large-scale ocean wave and wind changes that are associated with ENSO.