Font Size: a A A

Projection And Evaluation Of The Precipitation Extremes Indices Over China By CMIP5Models

Posted on:2014-03-30Degree:MasterType:Thesis
Country:ChinaCandidate:Y H HangFull Text:PDF
GTID:2250330401970262Subject:Development and utilization of climate resources
Abstract/Summary:PDF Full Text Request
As an important factor in the impact of climate change and climate resource utilization, change characteristics of extreme precipitation is more worthy of attention in context of global warming. Based on the interpolated-grid daily precipitation dataset of China for1960-2005(Chen et al.2007), performance of the20global climate models participating Coupled Model Intercomparison Project (CMIP5) are evaluated in terms of8extreme precipitation indices, then the rankings of models on simulations capabilities in spatial distribution and time characteristic of precipitation extremes over China are got, with the purpose to obtain the projection of climate change in precipitation extremes in21st century. The main results are as follows:(1) Using the methods such as average deviation, Taylor Figure and space techniques rating (SS), time sequence related coefficient and M2index, etc. assessment on the space-time-comprehensive simulation ability of8extreme precipitation indices indicates that each models exists with significant difference. According to the evaluation for the Space-temporal simulation capabilities of precipitation indices, the rankings of models are got, which show that the best four models are MRI-CGCM3MIROC4h, HadCM3and FGOALS2-s. The correlation coefficient between the simulation capabilities rank of models in space and time is0.42, indicating that the space-time simulation capability of the models has a higher consistency. The preferred spatial field simulation capabilities of best4models with20models ensemble mean comparison shows that the preferred models can decrease the average error percentage more than10%, which raise the most obvious part of eastern northwest a reduction of20%-30%.(2) Compared with the CMIP3models (Jiang et Al,2011), the CMIP5models generally have been greatly improved in space-time simulation capability of precipitation in Chinese region. There is a significantly improvement in space simulation capacity, especially extreme precipitation contribution rate (pfl95) and precipitation intensity (pint) were increased by6%and26.5%in average percentage error than CMIP3. In terms of temporal, especially for the simulation of daily precipitation intensity index (pint), correlation coefficient of GFDL-ESM2M increased from0.32to a0.38, and CNRM-CM5increased from0.14to0.36.(3) The ability in simulation between indexes are different. Space simulation capability index isrextreme precipitation contribution rate (pf195), average daily precipitation intensity (pint),90percentile-day extreme precipitation threshold (pq90) and maximum precipitation of5consecutive days (px5d), which the mean square error (standardized) between modeling and observation field is less than1.5, and these four index analog highly relevant, with correlation coefficients greater than0.4generally. Time simulation capability index:annual average precipitation (pav), extreme precipitation contribution rate (pfl95) and maximum precipitation of5consecutive days (px5d), for analog M2values are less than4in most of the country other than the Northwest Territories. Space-temporal simulation capabilities are poor for most models in maximum consecutive dry days (pxcdd) and maximum consecutive wet days (pxcwd), that model’s simulation capability still needs to be improved. In general, the models’simulation capacity of the extreme precipitation for the climate field is better than time change. Additionally, space and time simulation capability of models are better in summer than in winter.(4) Finally, combing the overall rankings of models on the simulations capabilities of temporal and spatial variability and using the method of preferred collection projection and probability projection with different weights, the precipitation extremes in the21th century under the RCP45scenario over China is projected. The results show that there are significant increases in rainfall in the early21st century,(2016-2035), an increase precipitation with the probability of0.81and0.98in summer and throughout the year will be happen in China, respectively, increased by more than10%with probability of0.58in winter. The main increasing area is located in the Northwest Territories and Yangtze-Huaihe River region, precipitation increases by more than10%with the probability of greater than0.5. The extreme precipitation (pq90and px5d) will be increasing nationally, but less than5%, which the probability is more than0.5. The persistent rainfall will increase in north and decrease in South, the maximum continuous dry days will reduce in north of Yangtze River; maximum consecutive dry days is increased by5%and10%with probability of0.6and0.3in the area south of the Yangtze River, respectively. There are more significant increasing of precipitation in the late century (2070-2099) than the early years, both the average precipitation and extreme precipitation is increasing greater than the previous. A10%increasing with the probability of0.73and0.67of summer and annual precipitation will be happen in China, respectively, winter precipitation will increase by more than25%with probability of0.49. The precipitation intensity will substantial increase, the more than10%increase has a possibility greater than0.5in Northwest Territories and Yangtze-Huaihe River region. Extreme precipitation increased significantly:in these areas, an increasing in pq90of more than10%has a probability of greater than0.5. The persistent rainfall is consistent with the previous, but the maximum continuous dry days’ increase of more than5%has less probability than0.5, and the pre-drought eased in the area south of the Yangtze River.
Keywords/Search Tags:CMIP5, extreme precipitation indices, Rank-based weighting method, Probability Projection
PDF Full Text Request
Related items