Improvement Of Sea Surface Turbulent Fluxes Parameterization Scheme In CAM3 And Its Impact On Climate Simulation

The parameterization scheme of sea surface turbulent fluxes in CAM3 (Community Atmosphere Model Version 3) is improved by the parameterization for wind gustiness, which considers the contribution to sea surface fluxes by boundary free convective and precipitation convective. Based on the improved CAM3, the simulation of generanl ciculation and climate is analysed. Furtherly, the simulation of boreal climatic change causing by El Nino events is studied. Main results are given as follows:(1) By means of the parameterization of convective gustiness, the mesoscale enhancement of surface fluxes is included in CAM3, thus, the parameterization scheme of sea surface turbulent fluxes in CAM3 is improved. The improved parameterization scheme is named CAM3 ME (Mesoscale Enhancement) scheme.(2) The sensitivity of CAM3 model to the parameterization of mesoscale enhangcement of sea surface turbulent fluxes showes that: the simulation of sea surfaces latent fluxes is more reasonable in improved model than original one. Compared to the original model, the absolute errors of the sea surface latent fluxes inkuroshio extension region in winter of the improved model is 31 W/m~2 less, and10～20 W/m~2 less in India Ocean and western Pacific in summer.(3) The CAM3 ME scheme has better performance in simulating the interannual variability of sea surface turbulent fluxes over tropical southern west Pacific, Kuroshio Current, mid-latitude North Paicific and eastern north Paicific in winter and south to Australia in summer.(4) The CAM3 ME scheme simulates the interannual variability of the winter general circulation more reasonable than the original scheme. The correlation coefficents of the interannual variability of Aleutine low in the ERA40 data and that in the CAM3 ME scheme is 0.564; it is 0.02 in the ERA40 and in the CAM3 scheme. The correlation coefficents of the interannual variability of subtropical high in northern Pacific in the ERA40 data and that in the CAM3 ME scheme is 0.475; in contrast, it is 0.237 between that in the ERA40 data and in the CAM3 scheme. The correlation coefficents of the index of PNA teleconnection in the ERA40 data and that in the CAM3 ME scheme is 0.791, while it is only 0.436 between that in the ERA40 data and that in the CAM3 scheme. The interannual variability of wind fields over North America and warm currents of North Pacific and North Atlatic is better simulated by the improved scheme. The correlation coefficents of the interannual variability of wind fields over the above region in the ERA40 data and that in the improved scheme is almostly 0.5, but the correlation coefficents of that in the ERA40 data and in the original sheme is less than 0.3.(5)The interannual variability of the general circulation over the tropical and southern Pacific is better simulated by the CAM3 ME scheme. The correlation coefficents of the index of East Asian monsoon in the ERA40 data and that in CAM3 ME scheme is 0.395, yet it is -0.093 of that in the ERA40 data and in the CAM3 scheme. Also the index of antartic oscillation (AAOI) is bettern simulated by the improved scheme, with the correlation coefficents of the AAOI in the ERA40 data and in the improved scheme is 0.428, and the correlation coefficents of the AAOI in the ERA40 data and the original scheme results is 0.078. Furthmore, the simulation of interannual variability of subtropical high and western jets over southern Pacific is also more reasonal than the original scheme.(6) The CAM3 ME scheme simulates the interannual variability of the winter precipitation over tropical sourthern west Pacific (TWSP) and tropical Atlantic (TA) more reasonable than the CAM3 scheme. The correlation coefficents of winter precipitation over TWSP (TA) areas in the CMAP data and in improved scheme is 0.578 (0.665), while it is 0.320 (0.428) in the CMAP data and in the orignal one. The interannual variability of the summer precipitation over TWSP simulated by the improved scheme is more closely to the CMAP data, the correlation coefficents of summer precipitation over TWSP in the improved scheme and in the CMAP data is 0.438, yet it is 0.124 in the original scheme and in the CMAP data.(7)The CAM3 ME scheme has better performance in simulating the relationship between El Nino and North Pacific/North America climate. The composite sea level pressure anomaly (SLPA) of Aluetine low is -7hPa in ERA40 data in strong El Nino years (1982, 1986, 1987, 1991 and 1997), it is -9hPa in improved model simulation, but it is only -3hPa in the original model results. The correlation coefficents of the index of observational nino3.4 SSTA and the index of 500hPa PNA teleconnection of the ERA40 data is 0.555. But the original scheme fails to simulation the closely relationship between the nino3.4 SSTA and 500hPa PNA teleconnection, with the correlation coefficents of the nino3.4 SSTA and the index of 500hPa PNA teleconnection is only 0.355. The correlation coefficents of the index of observational nino3.4 SSTA and the index of 500hPa PNA teleconnection of the improved scheme is 0.660. The improved model simulates the relationship between El Nino and climate changes over northern America much better than the original one. The composite 1000hPa air temperature anomaly is 3℃above zero in north and and 1℃below zero in south of North America in ERA40 data and in improved model simulation, but the air temperature anomaly is not well simulated in the original model results.