| Research on extreme climate is an important part of IPCC assessment report.For now most research on extreme climat is based on single point or single station,in this paper, an objective identification technique for regional extreme events(OITREE) was used,this method take regional events which have influence scope and duration as research object, and it does well in revealing the spatial durative variation characteristics of regional extreme events. Climate system model is only a certain approximation of actual climate system,when simulate and project the climate change,the rliability degree of models must be verified,then we can know how much confidence the results of projection are.Observation data of daily minimum temperature was used to learn about the basic situation of China’s regional low tempetature events during1961-2005. Combined with NCEP reanalysis data, the feasibility of identifying regional low temperature events with grid data that was downscaled with OPR method is verified. By multi-model ensemble with14CMIP5models, the simulation ability of multi-model ensemble was assessed.Furthermore, China’s regional low temperature events during2006-2009under different Emission Scenarios was projected with multi-model ensembles,and the uncertainty was analysised.The main contents and conclusions are as the follow:(1) The OITREE method did well in identifying China’s regional low tempetature events,159regional low tempetature events were identified during1961-2005.The7indices of regional low tempetature events including frequency,duration, extreme intensity, accumulated intensity, maximum impacted area, accumulated impacted area and integrated index all present decrease tendency during the45years.The decadal variation of regional low temperature events also present decrease tendency.The occurrence of regional low temperature events showed obvious seasonal characteristics which only happens during November to March of next year. The stongest area of accumulated intensity is in middle-west part of Inner Mongolia and south of Northeast China,and the stongest area of frequency is east China.Regional low temperature events to some extent reveal characteristics of the route of strong cold air masses in our country.This also reveals that a regional low temperature event is influnenced by one or several cold air.(2) NCEP reanalysis data was downscaled to station data with the method of the Optimal Points Regression (OPR), results show that NCEP reanalysis data shows very good consistency with the observational daily minimum temperatures, which illustrates that OPR method is feasible in downscaling daily minimum temperature. For China’s regional low temperature events, NCEP reanalysis data shows very good consistency with the observations in the long term trend, annual variation, and distributions of accumulated intensity and frequency during1961-2005.This verifys the feasibility of identifying regional low temperatue events with grid data after downscaled,which provides the premise for identifying regional low temperatue events with grid data of CMIP5models.(3) The total frequency of all grades of regional low temperature events during1961-2005after multi-model ensemble is less than observational result, but the ratio of all grades of regional low temperature events is about1:1:4:3, which is consistent with observation. All indices of China’s regional low temperature events of multi-model ensembles present decrease tendency which is consistent with the observations,their magnitudes are same,but the linear trend of multi-model ensembles result is tenderer than observation.The restult of multi-model ensembles can remove the inconsistent of an index of single model,which is convinient to judge the linear trend of regional low temperature events in the mass.The spatial correlation coefficient of accumulated intensity and frequency between multi-model ensemble and observation is respectively0.721and0.808,which is a little less than very few models,but the magnitude of accumulated intensity and frequency is close to observation.The reason of accumulated intensity in northern xinjiang being stronger than observation with some models is that the western route cold air in these models is strong.The result of accumulated intensity and frequency of multi-model ensembles is not the best in comparision with14models,but still in the top three.Simulation of indices of regional low temperature events with a single model is unstable,multi-model ensemble can overcome this shortcoming,and the the magnitude is closer to observation.Not only temporal variation,but also spatial distribution of China’s regional low tempetature events.not all of the indices of multi-model ensemble is not the best in compare with14models,but the result of multmodel ensemble is the best by comprehensive evaluation,and the magnitude is closer to observation.(4) The total frequency of China’s regional low temperature events under RCP2.6emission scenarios is the most,followed by RCP4.5,and RCP8.5. The increase in greenhouse gas emissions doesn’t trigger severe changes in climate.Both the linear trend and the magnitude of spatial distribution revels that regional low temperature events decrease at first and then increase during2006-2099under the emission scenarios of RCP2.6,will be decreasing under RCP4.5,the decreasing tendency is more obvious under RCP8.5,which is closely related to greenhouse gases emissions.The spatial distribution of accumulate intensity and frequency of regional low temperature events do not change obviously in the three period under three emission scenarios,which means that the increase of greenhouse gas has no influence on the spatial distribution of China’s regional low temperature events,but has influence on the magnitude.The high-value of accumulate intensity is northern part of north china and northern xinjiang,and the value decreases gradually from north to south. The high-value of frequency is east-central China, and the value decreases gradually from east to west.The biggest uncertainty area of accumulate intensity of regional low temperature events of projection is northern part of north China and northern Xinjiang. The biggest uncertainty area of frequency of regional low temperature events is northern part of north China and south of northeast China.Both of the ratio of absolute value of accumulate intensity and frequency of multi-model ensemble to corresponding deviation is bigger than1,which means thar the projecion result of multi-model ensemble is authentic. |
PDF Full Text Request