Potentials in improving predictability of multiscale tropical weather systems evaluated through ensemble assimilation of simulated satellite-based observations

As a follow-up of our recent paper on the practical and intrinsic predictability of multiscale tropical weather and equatorial waves, this study explores the potentials in improving the analysis and prediction of these weather systems through assimilating simulated satellite-based observations with a regional ensemble Kalman filter (EnKF). The observing networks investigated include the retrieved temperature and humidity profiles from the Advanced TIROS Operational Vertical Sounder (ATOVS) and global positioning system radio occultation (GPSRO), the atmospheric motion vectors (AMVs), infrared brightness temperature from Meteosat-7 (Met7-Tb), and retrieved surface wind speed from the Cyclone Global Navigation Satellite System (CYGNSS). It is found that assimilating simulated ATOVS thermodynamic profiles and AMV winds improves the accuracy of wind, temperature, humidity, and hydrometeors for scales larger than 200 km. The skillful forecast lead times can be extended by as much as 4 days for scales larger than 1000 km. Assimilation of Met7-Tb further improves the analysis of cloud hydrometeors even at scales smaller than 200 km. Assimilating CYGNSS surface winds further improves the low-level wind and temperature. Meanwhile, the impact from assimilating the current-generation GPSRO data with better vertical resolution and accuracy is comparable to assimilating half of the current ATOVS profiles, while a hypothetical 25-times increase in the number of GPSRO profiles can potentially exceed the impact from assimilating the current network of retrievedATOVSprofiles. Our study not only shows great promises in further improving practical predictability of multiscale equatorial systems but also provides guidance in the evaluation and design of current and future spaceborne observations for tropical weather.