Comparisons Of The Different Convection Parameterization Schemes In Regional Climate Simulations

Climate and climate change can have very important influences on the economic development and social life. Studies on them have become a hot topic in scientific fields. Meanwhile, regional climate and its change draw more and more attention from the governments and scientists. Research work on this topic is usually performed by regional climate models, which are sensitive to the selections of horizontal resolution, simulating area, buffer zone, vertical layer, convection parameterization scheme, etc. Being one of the most important physical processes of non-adiabatic heating, cumulus parameterization plays a dominating role in the model's ability in precipitation simulation. In order to obtain the best simulation results, it is very important to compare the performances among different convection parameterization schemes in the regional climate models. Many works have been carried out on this topic. Though much has learned from these works, further research is still needed since the duration of the simulations completed in those previous studies was relatively short, and simulating area was limited (usually covered only east China). There is a lack of tests on other regions in China by using different convection parameterization schemes, especially for the new version of the model RegCM3.In this study, we compared the simulation results by using two convection parameterization schemes (Grell scheme using the Fritsch and Chappell closure and Emanuel scheme) in the latest version of a regional climate model (RegCM3), which was developed in The Abdus Salam International Center for Theoretical Physics (ICTP) and was driven by the re-analysed data of ERA40 from ECMWF. Two 15-year length (1987-2001) simulations were carried out. Firstly we analysed and compared the capabilities between the two parameterization schemes in simulating the character of the atmospheric circulation over East Asia, and the distribution pattern of surface air temperature and precipitation in China. Then, we ensembled the two schemes as the third scheme, and divided the whole China area into nine different climate regions. The simulated spatial distribution of temperature and precipitation by using the three schemes in China and its nine sub-regions were compared, respectively. The abilitiesin reproducing the inter-annual varibilities of temperature and precipitation over each sub-reigion were also examined and compared.The comparisons show that the simulation abilities of different convection parameterization schemes are different in a certain extent. The differences among them are more remarkable in summer than in winter, especially in the precipitation simulation.Our main conclusions are as follows:RegCM3 can reproduce well the annual and seasonal atmospheric circulation over East Asia both by using Grell and Emanuel schemes. The western jets in the upper level of the atmosphere are well described in winter, while the location and strength of the Western Pacific Subtropical High and South Asia High are well simulated in summer. The shift from winter monsoon to summer monsoon and the thermal difference between summer and winter in the troposphere are also well simulated by the two schemes. The comparisons indicate that Grell scheme is better in reproducing the characteristics of pressure and wind in the upper level of the atmosphere than Emanuel scheme, while Emanuel scheme performs better in describing the temperature and moisture in the lower troposphere.The distribution patterns of annual and seasonal mean surface air temperature in China and in its sub-regions are well simulated by both Grell and Emanuel scheme, as well as their ensembled scheme. However, a general cold bias is found in each of the sub-regions. The differences among three schemes are generally small in a specific sub-region. The simulation is better in summer than in winter, when comparing to the observation except in NW2 sub-region. It is indicated that Grell scheme is better in winter, while Emanuel scheme is better in summer.The spatial distribution patterns of annual and seasonal mean precipitation over China and its sub-regions are well simulated by the three schemes. The simulated northward shift of the monsoon rain belt in China is also quite realistic. As for temperature, a better performance in summer is found than that in winter. The results show that Emanuel scheme overestimates the precipitation in most area of China, while Grell scheme underestimates it in CC, EA and SC sub-regions but overestimates it in other regions. Significant differences among the three schemes can be found, especially in summer. The comparisons among them indicate that Grell scheme is better than the other two schemes in all the sub-regions except in CC, EA and SC. Thecapability of simulation for precipitation is improved by Ensemble scheme in those sub-regions mentioned above. The three schemes all can reproduce well the features of spatial distribution of precipitation in NW1, NE and CC, while they all perform poorly in SC.The simulated inter-annual variabilities of temperature by the three schemes over China and its sub-regions are also rational, but with a smaller value when comparing to the observation. The simulated abilities of the precipitation in many regions remain to be improved.