Global Impacts of Assimilating CYGNSS Winds on Surface Wind Fields Using a 2-Dimensional Variational Analysis Method

The CYclone Global Navigation Satellite System (CYGNSS) produces observations of the global ocean surface wind speed within the tropics and subtropics. The coverage of CYGNSS data in space and time are complementary to existing polar orbiting microwave sensors (active and passive) whose measurement swaths have gaps in space and time that increase from the pole to the equator. So the question arises: what are the benefits of the complementary CYGNSS sampling in space and time to existing global surface wind analyses that incorporate current polar-orbiting sensors of ocean surface wind speed and ocean surface vector winds?

For this study, GFS operational and NASA MERRA-2 global analyses are used as baseline ocean surface wind analyses (i.e., current practice). First, the time evolution of tropical meteorological features of interest in these respective analyses are examined, including African Easterly Waves (AEWs) in the north Atlantic and episodes of the Madden-Julian Oscillation (MJO) in the Indian Ocean and western tropical Pacific Ocean. The differences between the representations of meteorological features such as these in the two baseline analyses provide an estimate of uncertainty of these kinds of features in current analysis systems.

To assess the impact of CYGNSS winds on the analysis of these meteorological features, retrieved CYGNSS Level 2 winds are assimilated every 6 hours by a Variational Analysis Method (VAM, Hoffman et al. 2003) using the GFS operational or MERRA-2 global analyses as the background field. The various quality control (QC) and data thinning filters applied to the CYGNSS retrieved winds will be reviewed. The statistics of observation innovations (i.e., observation-background) and analysis increments (i.e., analysis-background) help show the impact of CYGNSS data on the two baseline analyses. The impacts of CYGNSS data on meteorological features of interest will also be shown by the geographic distribution of analysis increments.