Comparison of the third-generation Japanese ocean flux data set J-OFURO3 with numerical simulations of Typhoon Dujuan (2015) traveling south of Okinawa

Insufficient in situ observations in high winds make it difficult to verify climatological data sets and the results of tropical cyclone (TC) simulations. Reliable data sets are necessary for developing numerical models that predict TCs more accurately. This study attempted to compare the third-generation Japanese Ocean Flux Data Sets with Use of Remote-Sensing Observations (J-OFURO3) data, with TC simulations conducted by a 2 km mesh coupled atmosphere-wave-ocean model. This is a case study of Typhoon Dujuan (2015) and the area of approximately 20̊N, 130̊E, south of Okinawa, was selected. The comparison reveals that J-OFURO3 data are reliable for verifying the atmospheric and oceanic components of TC simulations with two different initial sea surface temperature (SST) conditions, although the blank area remains within the inner core area for air temperature, specific humidity, and latent heat flux owing to issues with the construction method. Simulated maximum surface wind speeds (MSWs) are significantly correlated with J-OFURO3 MSWs. The asymmetrical distribution of simulated surface wind speeds within the inner core area can be reproduced well in the J-OFURO3 data set. In terms of the oceanic response to the TC, TC-induced sea surface cooling was reproduced well in the J-OFURO3 data set and is consistent with the simulation results. Unlike simulated SST, simulated surface wind speeds, surface air temperature, and surface specific humidity are still inconsistent with the J-OFURO3 data, even when the J-OFURO3 SST is used as the initial condition. New algorithms, more satellite data used, and model improvement are expected in the future.