An OSSE Study of the Impact of the Micro Unmanned Robot Observation Network (MURON) Ocean Surface Observation on Tropical Cyclone Prediction Using GSI-Based EnKF

The Micro Unmanned Robot Observation Network (MURON) is a planned in-situ observation network over the surface of West Pacific Ocean, and it will provide observations of the ocean surface wind and mass variables. MURON is designed to sample tropic and subtropics with continuous spatial and temporal resolution under even severe weather conditions. Such design results in higher resolution in-situ observations over the ocean surface for tropical cyclones (TC) than current conventional ocean surface observations or satellite observations. The goal of this study is to investigate the impacts of MURON for TC intensity forecast using Observation System Simulation Experiments (OSSEs).

In this study, the GSI-based EnKF data assimilation system is used with the WRF-ARW model to conduct OSSEs for TC Haiyan (2013) while Haiyan went through rapid intensification. MURON observation improves TC intensity and track at both initial (analysis) time and forecast time. The circulation of TC is improved largely due to correction of initial vorticity and its convergence at the low-to-mid levels by the MURON observation. The improvement of intensity forecast is first attributed largely by assimilating wind observation when Haiyan is at the tropical disturbance stage, and then by the mass observation when Haiyan moves toward tropical storm stage. Our study also shows that the MURON and SATWIND upper air observations are complementary to each other. Assimilating MURON and SATWND decreases the error of vertical profiles of the wind and reduces the vertical wind shear, which facilitates TC intensification. In addition, our results show that, for the assimilation of MURON, it should be preferable to select the mixing ratio as a moisture control variable rather than relative humidity because the moist environment over the ocean causes non-Gaussian distribution for relative humidity or pseudo-relative humidity.