Detection Of Anthropogenic Signal And Projection In Extreme Precipitation Changes Over China

Based on the observation dateset over China,we gave the results of detection of anthropogenic signal in extreme precipitation changes using a robust statistical method.The perfomenace of CMIP5 models in simulating the spatial pattern and internnual variability of extreme precipitation changes is evaluated when determin the influene of anthropogenic.According to the perfomenace of models and the method of ranking-wighted,probability projection of extreme precipitation during 21st century is then studied.The changes in intensity and probability of extreme precipitation are further investigated when global warming reaches to 1.5? and 2?,we focus on the extra half a degree effect.The main conclusions of these analyses are drawn as follows:(1)According to the field significance test,significant influence of anthropogenic forcing on extreme precipitation in China may have not emerged yet.The signal of change in extreme precipitation due to anthropogenic change can be significantly detectable by around 2035 under RCP8.5 sceario.Changes in extreme precipitation would have been already quite when they are detectable.Regional events expected once every 20,50,and 100 years in current climate is expected to occur about every 15,34,and 63 years when anthropogenic signal has clearly emerged.(2)CMIP5 models still have wet bias in the regions with an overestimation of PRCPTOT and an underestimation of CDD,especially in western China.Generally speaking,the CMIP5 models can well reproduce the basic characteristics of extreme precipitation indices including both the spatial variation and interannual variability in eastern China.However,in western China,most models can hardly reproduce the basic characteristics of extreme precipitation indices.Consistency of models'performance between spatial pattern and inteannual variability is higher in eastern China than in western China.MRI-CGCM3,CMCC-CM,MIROC4h,HadGEM2-ES and HadGEM2-CC are top five models in eastern China.However,MRI-CGCM3,EC-EARTH,HadCM3,IPSL-CM5A-LR are the leading models in western China.The wet bias in BMME is significantly reduced over western and northern China.BMME can relatively well reproduce the cross-equatorial flows,which contribute to the reduction of wet bias in western and northern China.(3)Compared to equal weighted scheme,the projection from the weighted scheme is more credible,as its Signal-to-noise ratio(SNR)is generally larger.For almost the whole domain in western China and the Yellow-Huai River Basin in eastern China,there is a probability greater than 50%for an increase of PRCPTOT,SDII and R95T in excess of 10%during the end of 21st century.We can thus confidently conclude that the precipitation in those regions is likely to be more abundant and of more extremes.There is a negative change for CDD to the north of 3 0°N and a positive change to the south of 3 0°N during the entire 21st century.This indicates that drought-favorable conditions in northern China would reduce in the future and they are more likely to increase in southern China.However,the SNR for the projection of CDD is less than 1.0 almost everywhere else.(4)Magnitude of extreme precipitation is projected to increase by about 6%for 1.5? warming and 11%for 2? warming relative to reference period(1986-2005).The additional half a degree warming makes the magnitude to increase by 4%.Probabilities of a given change depend on the return periods.The rarest events(the longest return period)have largest increase in the risk for a given warming level.For historical 100-year events,their occurrence probability increases by 1.6 times at 1.5?warming,whereas at a 2? warming,the increase is about 2.4 times.This implies that regional average event expected once every 100 years in historical period is expected to occur about every 62 and 42 years under 1.5? and 2? warming climate,respectively.The projection of regional averaged changes in extreme precipitation is independent of the RCPs scenarios.Uncertainties among models for magnitude and probability of change increase with global warming and higher return period events exhibits larger uncertainties.