The Effects Of The Non-breaking Wave-induced Vertical Mixing On The Forecast Of Typhoon Track

Abstract The Weather Research and Forecasting model (WRF) is configured thesame domain with the operational forecast system MASNUM (Key Lab. MarineScience&Numerical Modeling) wave-tide-circulation coupled model. Two scenariosare designed using MASNUM, the first scenario include Bv (non-breakingwave-induced mixing) parameterization, the second scenario without Bv. A series ofatmospheric numerical experiments are conducted to test the sensitivities of radiationfluxes and typhoon tracks to high spatial and temporal resolution SST products outputfrom two scenarios of MASNUM. The results are compared with in-situ observedradiation flux data, satellite cloud fractions data and Japan Meteorological Agency’sbest tracks data. By analyzing the result, we found:The cloud fraction has significant effects on the short-wave radiation, moisturecontent can also affect short-wave radiation when the sky is clear. The long-wave isvery sensitive to moisture content. The experiment which use SST data fromMASNUM system give a more accurate estimation of short-wave and long-waveradiation than the control experiment. High spatial and temporal resolution SST canchange the vertical moisture distribution, and then affect cloud fraction. The cloud andradiation parameterizations of WRF should be improved.Two experiments are conducted to simulate typhoon JANGMI during September2008. The high resolution SST data from MASNUM system can improved thetyphoon track forecast. The bias of center position of the typhoon reduces about11%compared with the control experiment especially At the second half of typhoonprocess, The bias of center position of the typhoon reduces about37%. Theexperiment by using SST from MASNUM model is more realistic thanNCEP/NCARG SST from the control run. The results also can show that typhoonintensity has different sensitivity to sea surface temperature in different phase. TheSST drop induced by typhoon decrease the heat fluxes transport from ocean to atmosphere and then weaken typhoon.Three experiments are also conducted to simulate all the33tropical cyclonesfrom2008to2011at the model domain. The first experiment use NCEP/NCAR SSTwhich is default for WRF, the second experiment use high spatial and temporalresolution SST products output from the scenario of MASNUM which include Bv, thethird experiment use high spatial and temporal resolution SST products output fromthe scenario of MASNUM which exclude Bv. The results show that: The secondexperiment hindcast24hour typhoon track much better than the other twoexperiments, which give an averaged typhoon position error reduction about9%relative to the first experiment, give an averaged typhoon position error reductionabout7%relative to the third experiment, the third experiment give an averagedtyphoon position error reduction about2%relative to the first experiment. The secondexperiment also hindcast48hour typhoon track much better than the other twoexperiments, which give an averaged typhoon position error reduction about10%relative to the first experiment, give an averaged typhoon position error reductionabout4%relative to the third experiment, the third experiment give an averagedtyphoon position error reduction about6%relative to the first experiment.