Sensitivity of the GCM driven summer monsoon simulations to cumulus parameterization schemes in nested RegCM3

The regional climate model (RegCM3) from the Abdus Salam International Centre for Theoretical Physics has been used to simulate the Indian summer monsoon for three different monsoon seasons such as deficit (1987), excess (1988) and normal (1989). Sensitivity to various cumulus parameterization and closure schemes of RegCM3 driven by the National Centre for Medium Range Weather Forecasting global spectral model products has been tested. The model integration of the nested RegCM3 is conducted using 90 and 30-km horizontal resolutions for outer and inner domains, respectively. The India Meteorological Department gridded rainfall (1° × 1°) and National Centre for Environment Prediction (NCEP)–Department of Energy (DOE) reanalysis-2 of 2.5° × 2.5° horizontal resolution data has been used for verification. The RegCM3 forced by NCEP–DOE reanalysis-2 data simulates monsoon seasons of 1987 and 1988 reasonably well, but the monsoon season of 1989 is not represented well in the model simulations. The RegCM3 runs driven by the global model are able to bring out seasonal mean rainfall and circulations well with the use of the Grell and Anthes–Kuo cumulus scheme at 90-km resolution. While the rainfall intensity and distribution is brought out well with the Anthes–Kuo scheme, upper air circulation features are brought out better by the Grell scheme. The simulated rainfall distribution is better with RegCM3 using the MIT-Emanuel cumulus scheme for 30-km resolution. Several statistical analyses, such as correlation coefficient, root mean square error, equitable threat score, confirm that the performance of MIT-Emanuel scheme at 30-km resolution is better in simulating all-India summer monsoon rainfall. The RegCM3 simulated rainfall amount is more and closer to observations than that from the global model. The RegCM3 has corrected its driven GCM in terms of rainfall distribution and magnitude over some parts of India during extreme years. This study brings out several weaknesses of the RegCM model which are documented in this paper.     

中国夏季不同强度降水模拟对不同积云对流参数化方案的敏感性研究

利用PσRCM9区域气候模式, 分析了中国夏季不同强度降水模拟对不同积云对流参数化方案的敏感性。结果表明, 采用四种积云对流参数化方案, 模式能够模拟出小雨、 大雨和暴雨的雨量百分比和雨日百分比空间分布的一致性特征, 但不能模拟出中雨雨量百分比和雨日百分比空间分布的相似性, 这是由于模式不能模拟中雨雨量百分比的空间分布形式所致。还发现模拟的我国夏季降水以小雨和中雨为主, 四种积云对流参数化方案均低估了中国夏季大雨和暴雨对总降水的贡献, 尤其是在我国西部、 东北和华北地区更明显。不同积云对流参数化方案下模拟的极端强降水阈值的空间分布形式基本与观测一致, 但强度与观测存在较大差异。相比较而言, Grell方案较Kuo、 Anthes-Kuo和Betts-Mille积云对流参数化方案更适合中国东南部地区夏季极端强降水的模拟。     

Performance of Convective Parameterization Schemes in Asia Using RegCM:Simulations in Three Typical Regions for the Period 1998–2002

This study discusses the sensitivity of convective parameterization schemes(CPSs) in the Regional Climate Model(version 4.3)(Reg CM4.3) over East/South Asia. The simulations using different CPSs in Reg CM are compared to discover a suitable scheme for this region, as the performance of different schemes is greatly influenced by region and seasonality. Over Southeast China and the Bay of Bengal, the Grell scheme exhibits the lowest RMSEs of summer precipitation compared to observed data. Moreover, the Emanuel over land and Grell over ocean(ELGO) scheme enhances the simulation, in comparison with any single CPS(Grell/Emanuel) over Western Ghats, Sri Lanka, and Southeast India. Over the Huang–Huai–Hai Plain(3H) and Tibetan Plateau(TP) regions of China, the Tiedtke scheme simulates the more reasonable summer precipitation with high correlation coefficient and comparable amplitude. Especially, it reproduces a minimum convective precipitation bias of 8 mm d-1and the lowest RMSEs throughout the year over East/South Asia. Furthermore, for seasonal variation of precipitation, the Tiedtke scheme results are closer to the observed data over the 3H and TP regions. However, none of the CPSs is able to simulate the seasonal variation over North Pakistan(NP). In comparison with previous research, the results of this study support the Grell scheme over South Asia. However, the Tiedtke scheme shows superiority for the 3H, TP and NP regions. The thicker PBL, less surface latent heat flux, the unique ability of deep convection and the entrainment process in the Tiedtke scheme are responsible for reducing the wet bias.     

积云参数化方案对夏季东亚季风区海气系统位相关系模拟的影响

本文利用区域海气耦合模式RegCM-POM,分别选取Grell积云参数化方案和Emanuel积云参数化方案对北半球夏季(5—10月)的东亚气候进行模拟,研究不同积云对流参数化方案(CPS)对东亚夏季季风区海气系统位相关系模拟的影响。结果表明:不同CPS模拟的陆地降水具有一定的不确定性,而海洋降水和海温的模拟受CPS选择的影响更大。其中,Emanuel方案对海洋降水和海温的分布形势模拟总体上要好于Grell方案,且可以更好的模拟中国近海各海区的海气系统位相关系,特别是大气对海温的负反馈过程。原因在于Emanuel方案模拟的对流降水与海温的位相关系更接近观测总降水与海温的位相关系;而Grell方案对南海和孟加拉湾的对流降水模拟偏少,对黑潮对流降水的模拟偏多,错误地模拟了这几个海区积云对流过程发挥的作用,故其模拟的海气系统位相关系不如Emanuel方案。     

Simulation of Indian summer monsoon circulation and rainfall using RegCM3

Summary The Regional Climate Model RegCM3 has been used to examine its suitability in simulating the Indian summer monsoon circulation features and associated rainfall. The model is integrated at 55 km horizontal resolution over a South Asia domain for the period April–September of the years 1993 to 1996. The characteristics of wind at 850 hPa and 200 hPa, temperature at 500 hPa, surface pressure and rainfall simulated by the model over the Indian region are examined for two convective schemes (a Kuo-type and a mass flux scheme). The monsoon circulation features simulated by RegCM3 are compared with those of the NCEP/NCAR reanalysis and the simulated rainfall is validated against observations from the Global Precipitation Climatology Centre (GPCC) and the India Meteorological Department (IMD). Validation of the wind and temperature fields shows that the use of the Grell convection scheme yields results close to the NCEP/NCAR reanalysis. Similarly, the Indian Summer Monsoon Rainfall (ISMR) simulated by the model with the Grell convection scheme is close to the corresponding observed values. In order to test the model response to land surface changes such as the Tibetan snow depth, a sensitivity study has also been conducted. For such sensitivity experiment, NIMBUS-7 SMMR snow depth data in spring are used as initial conditions in the RegCM3. Preliminary results indicate that RegCM3 is very much sensitive to Tibetan snow. The model simulated Indian summer monsoon circulation becomes weaker and the associated rainfall is reduced by about 30% with the introduction of 10 cm of snow over the Tibetan region in the month of April.     

RegCM3模式对中国东部夏季降水的模拟试验

利用最新发布的区域气候模式RegCM3对1998年5—8月中国东部降水进行了模拟试验，考察了模式对降水和大尺度环流系统的模拟能力。结果表明：不同对流性降水方案对不同月份、不同区域的降水模拟效果差别较大，采用Kuo方案和Grell方案时模拟的降水效果要好于BM方案；RegCM3能较成功地再现异常降水的月际尺度变化和空间分布等基本特征；模式还较好地模拟了西太平洋副高脊线的演变过程和两次向北传播的季节内振荡。该模式可应用于中国东部夏季风降水的研究。     

Simulation of heavy precipitation episode over eastern Peninsular Malaysia using MM5: sensitivity to cumulus parameterization schemes

A comparative study has been conducted to investigate the skill of four convection parameterization schemes, namely the Anthes–Kuo (AK), the Betts–Miller (BM), the Kain–Fritsch (KF), and the Grell (GR) schemes in the numerical simulation of an extreme precipitation episode over eastern Peninsular Malaysia using the Pennsylvania State University—National Center for Atmospheric Research Center (PSU-NCAR) Fifth Generation Mesoscale Model (MM5). The event is a commonly occurring westward propagating tropical depression weather system during a boreal winter resulting from an interaction between a cold surge and the quasi-stationary Borneo vortex. The model setup and other physical parameterizations are identical in all experiments and hence any difference in the simulation performance could be associated with the cumulus parameterization scheme used. From the predicted rainfall and structure of the storm, it is clear that the BM scheme has an edge over the other schemes. The rainfall intensity and spatial distribution were reasonably well simulated compared to observations. The BM scheme was also better in resolving the horizontal and vertical structures of the storm. Most of the rainfall simulated by the BM simulation was of the convective type. The failure of other schemes (AK, GR and KF) in simulating the event may be attributed to the trigger function, closure assumption, and precipitation scheme. On the other hand, the appropriateness of the BM scheme for this episode may not be generalized for other episodes or convective environments.     

Diurnal cycles of precipitation over subtropical China in IPCC AR5 AMIP simulations

Atmospheric Intercomparison Project simulations of the summertime diurnal cycle of precipitation and low-level winds over subtropical China by Intergovernmental Panel on Climate Change Fifth Assessment Report models were evaluated. By analyzing the diurnal variation of convective and stratiform components, results confirmed that major biases in rainfall diurnal cycles over subtropical China are due to convection parameterization and further pointed to the diurnal variation of convective rainfall being closely related to the closure of the convective scheme. All models captured the early-morning peak of total rainfall over the East China Sea, but most models had problems in simulating diurnal rainfall variations over land areas of subtropical China. When total rainfall was divided into stratiform and convective rainfall, all models successfully simulated the diurnal variation of stratiform rainfall with a maximum in the early morning. The models, overestimating noon-time （nocturnal） total rainfall over land, generally simulated too much convective rainfall, which peaked close to noon （midnight）, sharing some similarities in the closures of their deep convection schemes. The better performance of the Meteorological Research Institute atmospherer. ocean coupled global climate model version 3 （MRI-CGCM3） is attributed to the well captured ratio of the two kinds of rainfall, but not diurnal variations of the two components. Therefore, a proper ratio of convective and stratiform rainfall to total rainfall is also important to improve simulated diurnal rainfall variation.     

Sensitivity Study of the South China Sea Summer Monsoon in 1998 to Different Cumulus Parameterization Schemes

In this study, the improved high-resolution regional climate model of the China National Climate Center （RegCM_NCC） is used to examine the sensitivity of the simulated circulation and rainfall during the South China Sea summer monsoon （SCSSM） period during 1998 in an effort to compare to other cumulus param- eterization schemes. The investigation has indicated that the model is capable of simulating the seasonal march of the SCSSM and that the results were very sensitive to the choice of cumulus parameterization schemes. It seems that the Kuo cumulus parameterization scheme simulates the process of the SCSSM onset reasonably well, which can reproduce the onset timing and dramatic changes before and after the onset, especially the upper- and lower-level wind-fields. However, there are still some discrepancies between the simulations and observations. For example, the model can not completely simulate the intensity of the rainfall or the location of the western Pacific subtropical high as well as the feature of the rapid northward propagation of seasonal rain belt.     

两种积云对流参数化方案对1998年区域气候季节变化模拟的影响研究

利用区域气候模式RegCM3,选择Kuo-Anthes积云对流参数化方案和基于FC80假设的Grell积云对流参数化方案,对1998年东亚气候分别进行年尺度模拟,模拟结果对比分析表明:在春、夏季转换时期, 两者模拟的降水形势差别较大, 对江淮、中南和华南地区的夏季降水量模拟差别最为明显。对流层上层模式变量和模式大气质量对积云对流参数化方案的选择不敏感, 而对流层中、下层模式变量对积云对流参数化方案比较敏感。不同积云对流参数化方案对8天时间尺度的天气系统模拟差别比较大。在积云对流比较活跃的夏季,不同参数化方案会导致模式大气出现不同的系统性偏差。由于模式在陡峭地形处动力过程计算方案存在缺陷,在高原与盆地的交界处,模式误差会产生明显的突变。     

Climatic simulations of the eastern Andes low-level jet and its dependency on convective parameterizations

The South American low level jet (SALLJ) of the Eastern Andes is investigated with Regional Climate Model version 3 (RegCM3) simulations during the 2002–2003 austral summer using two convective parameterizations (Grell and Emanuel). The simulated SALLJ is compared with the special observations of SALLJEX (SALLJ Experiment). Both the Grell and Emanuel schemes adequately simulate the low level flow over South America. However, there are some intensity differences. Due to the larger (smaller) convective activity, the Emanuel (Grell) scheme simulates more intense (weaker) low level wind than analysis in the tropics and subtropics. The objectives criteria of Sugahara (SJ) and Bonner (BJ) were used for LLJ identification. When applied to the observations, both criteria suggest a larger frequency of the SALLJ in Santa Cruz, followed by Mariscal, Trinidad and Asunción. In Mariscal and Asunción, the diurnal cycle indicates that SJ occurs mainly at 12 UTCs (morning), while the BJ criterion presents the SALLJ as more homogenously distributed. The concentration into two of the four-times-a-day observations does not allow conclusions about the diurnal cycle in Santa Cruz and Trinidad. The simulated wind profiles result in a lower than observed frequency of SALLJ using both the SJ and BJ criteria, with fewer events obtained with the BJ. Due to the stronger simulated winds, the Emanuel scheme produces an equal or greater relative frequency of SALLJ than the Grell scheme. However, the Grell scheme using the SJ criterion simulates the SALLJ diurnal cycle closer to the observed one. Although some discrepancies between observed and simulated mean vertical profiles of the horizontal wind are noted, there is large agreement between the composites of the vertical structure of the SALLJ, especially when the SJ criterion is used with the Grell scheme. On an intraseasonal scale, a larger southward displacement of SALLJ in February and December when compared with January has been noted. The Grell and Emanuel schemes simulated this observed oscillation in the low-level flow. However, the spatial pattern and intensity of rainfall and circulation anomalies simulated by the Grell scheme are closer to the analyses than those obtained with the Emanuel scheme.     

东亚夏季风降水和环流数值模拟对不同对流参数化方案的敏感性

A 5-level spectral AGCM (ImPKU-5LAGCM) is used to examine the sensitivity of the simulated results of the summer monsoon rainfall and circulation in East Asia to different cumulus parameterization schemes in the climatological-mean case and in the cases of weak and strong Asian summer monsoons,respectively. The results simulated with the Arakawa-Schubert＇s(hereafter A-S＇s), Kuo＇s and Manabe＇s cumulus parameterization schemes show that these simulated distributions of the summer monsoon rainfall and circulation in East Asia depend strongly on the cumulus parameterization schemes either in the climatological-mean case or in the cases of weak and strong Asian summer monsoons. From the simulated results, it might be shown that the Kuo scheme appears to be more suitable for the simulation of the summer monsoon rainfall and circulation in East Asia than the A-S scheme or the Manabe scheme, although the A-S scheme is somewhat better in the simulations of the tropical rainfall. This might be due to that the Kuo＇s cumulus parameterization scheme is able to reflect well the characteristics of rainfall cloud system in the East Asian summer monsoon region, where the rainfall system used to be a mixing of cumulus and stratus.     

Role of cumulus parameterization schemes in simulating heavy rainfall episodes off the coast of Maharashtra state during 28 June–4 July 2007

Indian summer monsoon gives on an average 250 cm of rainfall due to mesoscale/synoptic scale systems over west coast of India; now-a-days, MM5 model plays a very crucial role in simulating such heavy rainfall episodes like Mumbai (India) on 26 July 2005, which caused devastation through flash floods. The main aim of this study is to simulate such heavy rainfall episodes using three different cumulus parameterization schemes (CPS) namely Kain–Fritsch-1, Anthes–Kuo and Grell and to compare their relative merits in identifying the characteristics of mesoscale systems over 14 stations in coastal Maharashtra state during 28 June–4 July 2007. MM5 control experiment results are analysed for the fields of mean sea level pressure, wind, geopotential height at 850 hPa and rainfall with the above schemes. It is interesting to note that Kain–Fritsch-1 scheme simulates heavy rainfall amount of 48 cm for an observed rainfall of 51 cm in 24 h. The Grell scheme underestimates heavy rainfall episodes, while the Anthes–Kuo scheme is found to over predict rainfall on both temporal and spatial scales. The reason for better performance of KF-1 scheme may be due to inclusion of updrafts and downdrafts. Later the simulated rainfall quantities at 14 stations over study region are validated with both 3B42RT and observed rain gauge data of India Meteorological Department (IMD) and the results are promising. Finally, for the heavy rainfall prediction cases, the best threat score is at 0.25 mm threshold for three CPSs. Thus, this study is a breakthrough in pointing out that the KF-1 experiment has the best skill in predicting heavy rainfall episodes.     

Customization of regional climate model (RegCM4) over Indian region

The regional climate model (RegCM4) is customized for 10-year climate simulation over Indian region through sensitivity studies on cumulus convection and land surface parameterization schemes. The model is configured over 30° E–120° E and 15° S–45° N at 30-km horizontal resolution with 23 vertical levels. Six 10-year (1991–2000) simulations are conducted with the combinations of two land surface schemes (BATS, CLM3.5) and three cumulus convection schemes (Kuo, Grell, MIT). The simulated annual and seasonal climatology of surface temperature and precipitation are compared with CRU observations. The interannual variability of these two parameters is also analyzed. The results indicate that the model simulated climatology is sensitive to the convection as well as land surface parameterization. The analysis of surface temperature (precipitation) climatology indicates that the model with CLM produces warmer (dryer) climatology, particularly over India. The warmer (dryer) climatology is due to the higher sensible heat flux (lower evapotranspiration) in CLM. The model with MIT convection scheme simulated wetter and warmer climatology (higher precipitation and temperature) with smaller Bowen ratio over southern India compared to that with the Grell and Kuo schemes. This indicates that a land surface scheme produces warmer but drier climatology with sensible heating contributing to warming where as a convection scheme warmer but wetter climatology with latent heat contributing to warming. The climatology of surface temperature over India is better simulated by the model with BATS land surface model in combination with MIT convection scheme while the precipitation climatology is better simulated with BATS land surface model in combination with Grell convection scheme. Overall, the modeling system with the combination of Grell convection and BATS land surface scheme provides better climate simulation over the Indian region.     

Simulation of China Summer Precipitation Using a Regional Air-Sea Coupled Model

A regional air-sea coupled model based on the regional climate model(RegCM3)and the regional oceanic model POM(Princeton Ocean Model)is developed and a series of experiments are performed to verify the ability of the coupled model in simulating the summer precipitation over China from 1963 to 2002.The results show that the space correlation coefficients between the GISST(Global Ice and Sea Surface Temperature)data and the simulated SST by RegCM3-POM exceed 0.9.Compared with the uncoupled experiments,the coupled model RegCM3-POM has a better performance in simulating the mean summer(June to August)precipitation over China,and the distribution of the rainband in the coupled model is more accurate.The improvement of the rainfall simulation is significant over the Yangtze River Valley and in South China.The rainbelt intraseasoaal evolution over eastern China in summer indicates that the simulation ability of RegCM3-POM is improved in comparison with the uncoupled model.The interannual summer rainfall variation over eastern China simulated by RegCM3-POM is in accordance with observation,while the spatial pattern of the interannual summer rainfall variation in the uncoupled model is inaccurate.The simulated correlation coefficient between the summer rainfall in the uncoupled model RegCM3 and observation is0.30 over the Yangtze River Valley and 0.29 in South China.The coefficient between the rainfall in the coupled RegCM3-POM and observation is 0.48 over the Yangtze River Valley and 0.61 in South China.The RegCM3-POM has successfully simulated the correlation coefficients between summer rainfall in the Yangtze River Valley and SST anomalies of the Bay of Bengal,South China Sea,and the Kuroshio area,whereas the uncoupled model RegCM3 fails to reproduce this relationship.The study further shows that the monsoon circulation and the path of the moisture transport flux simulated by RegCM3-POM are in good agreement with the NCEP/NCAR data.     

Sensitivity of the simulated African monsoon of summers 1993 and 1999 to convective parameterization schemes in RegCM3

In this study, the International Center for Theoretical Physics Regional Climate Model version 3 (RegCM3) was used to investigate the sensitivity of the simulation of the West African monsoon using four different cumulus and closures parameterization schemes of Anthes Kuo (AK), Grell and Fristish Chappell (GFC), Grell and Arakawa Schubert (GAS), and MIT-Emmanuel (EM) while maintaining other physical packages unchanged. The contrasting monsoon years of 1993 and 1999, which were dry and wet years, respectively, were simulated. The model was integrated from a period of 5 months, starting from May 1 to September 30 of each year using the European Centre for Medium-Range-Weather Forecast (ECMWF) Reanalysis data (ERA40) as input boundary conditions. The 6-hourly reanalysis data were used to provide the lateral boundary conditions, and the observed weekly Reynolds Sea Surface Temperature interpolated to 6 h was used as the lower boundary forcing. The results show that in West Africa, monsoon precipitations are sensitive to the choice of cumulus parameterization and closure schemes. None of the schemes is able to simulate the monsoon rainfall accurately, and furthermore, there is little difference in behavior among schemes between dry and wet years. The spatial features of precipitation are not identical among schemes, although they all show a northward shift of the rain bands, giving a very wet Sahel and dry Guinean Coast. The GFC and EM schemes are able to capture the diurnal cycle of precipitation and the zonal averages of stratiform rain fractions as observed in the Tropical Rainfall Measuring Mission (TRMM), although they overestimated rainfall amounts. The most important deficiencies, however, cannot be attributed to the schemes. In particular, the northward shift of both the rain band and the AEJ in RegCM3 is the result of unrealistic soil moisture resulting from the way albedo is parameterized, leading to an excessive northward penetration of monsoon flow. A sensitivity study showed that an adjustment of initial albedo values over the Sahel improved the simulation, and in particular the position of rain bands and of the AEJ.     

MODIFIED MASS FLUX CUMULUS CONVECTIVE PARAMETERIZATION SCHEME AND ITS SIMULATION EXPERIMENT—PART Ⅰ:MASS FLUX SCHEME AND ITS SIMULATION OF THE 1991 FLOOD EVENT*

Based on the existing cumulus convective parameterization schemes,a mass flux scheme (MFS) for cumulus convective parameterization has been successfully developed by reference to the work of Chen et al. (1996).The MFS is a comprehensive scheme.In MFS,not only the importance of the large-scale moisture convergence is taken into account,but also it includes the cumulus updrafts and downdrafts,cumulus-induced subsidence in the environmental air.entrainment,detrainment and evaporation.The interaction between the cumulus and the environment is described by using a one-dimensional bulk model.At the same time the scheme includes the penetrative and shallow convections.The MFS has been successfully incorporated into the regional climate model RegCM2 developed by NCAR.The new model has been applied to simulate summer monsoon characteristics and their variations of heavy rainfall process in the Changjiang-Huaihe River Basins for three months from May to July 1991.The results show that the new model can successfully simulate this rainfall prolonged process.By comparising the model outputs of RegCM2.using the Kuo scheme and the MFS.it is found that the MFS is better in simulating the surface temperature,rainfall position and amount,and rainfall duration.     

区域气候模拟中多种对流参数化方案的比较研究

利用区域气候模式RegCM2,选择Betts-Miller、Kuo-Anthes和Grell3种积云对流参数化方案,进行了1991年和1994年5～7月2个个例的东亚季风区区域气候模拟试验,重点分析比较了3种参数化方案对降水总量分布、主要降水时段、雨带南北变动和基本气象要素场的模拟能力,并相对实测降水和基本要素的观测分析进行了统计检验.通过对比,发现Betts方案的模拟能力明显优于其他两种方案.     

南海夏季风爆发的数值模拟

利用高分辨率的区域气候模式 (RegCM_NCC) 对南海夏季风爆发进行模拟研究。研究表明:该模式对积云对流参数化方案的选择十分敏感, 其中以Kuo积云参数化方案为最好, 可以比较成功地模拟出南海夏季风的爆发时间、爆发前后高、低层风场的剧烈变化以及季风与季风雨带的向北推进。然而该方案对于雨量和副热带高压位置的模拟, 与观测相比尚存在一定的偏差, 主要表现为副热带高压位置模拟偏北、偏东; 南海地区的降水量模拟偏少、降水范围偏小。此外, 采用4种参数化方案 (Kuo, Grell, MFS, Betts-Miller) 集成的结果在某种程度上要优于单个方案的结果, 这种改善主要体现在对南海地区季风爆发后降水的模拟上。     

MODIFIED MASS FLUX CUMULUS CONVECTIVE PARAMETERIZATION SCHEME AND ITS SIMULATION EXPERIMENT—PART Ⅰ:MASS FLUX SCHEME AND ITS SIMULATION OF THE 1991 FLOOD EVENT

Based on the existing cumulus convective parameterization schemes,a mass flux scheme(MFS)for cumulus convective parameterization has been successfully developed by reference to thework of Chen et al.(1996).The MFS is a comprehensive scheme.In MFS,not only theimportance of the large-scale moisture convergence is taken into account,but also it includes thecumulus updrafts and downdrafts,cumulus-induced subsidence in the environmental air.entrainment,detrainment and evaporation.The interaction between the cumulus and theenvironment is described by using a one-dimensional bulk model.At the same time the schemeincludes the penetrative and shallow convections.The MFS has been successfully incorporated into the regional climate model RegCM2developed by NCAR.The new model has been applied to simulate summer monsoon characteristicsand their variations of heavy rainfall process in the Changjiang-Huaihe River Basins for threemonths from May to July 1991.The results show that the new model can successfully simulate thisrainfall prolonged process.By comparising the model outputs of RegCM2.using the Kuo schemeand the MFS.it is found that the MFS is better in simulating the surface temperature,rainfallposition and amount,and rainfall duration.