Tempo-spatial Characteristics Of Precipitation Change And The Uncertainties In Asia During The Past 100 Years

Atmospheric precipitation is one of the key components of terrestrial water cycle.In the context of global warming,long-term changes in precipitation have a serious impact on the natural world.Complex terrain and sea-land distribution characteristics in Asia make a remarkable regional difference in precipitation and extreme precipitation events.In order to detect and understand the response of regional terrestrial water cycle to global warming,it is important to analyze the temporal and spatial variations of precipitation in continental or sub-continental scales in the past century.The application of regional average methods and different precipitation indicators in different regions will be important sources of uncertainty in the study of regional precipitation changes.In this paper,various precipitation observation data from different sources were used to seek the optimal research methods in China with the least data uncertainty,in order to solve the uncertainty caused by the research methods,and explore the optimal regional precipitation research method.The long-term changes in the rainy season in East Asia with good data integrity and veracity were analyzed,and uncertainties caused from the use of data;and study period were discussed.The changes in precipitation and extreme precipitation indices of the Hindu Kush Himalaya(HKH)region during recent decades were analyzed.The data and methods used to study the characteristics of precipitation over recent 100 years in various regions of Asia were finally determined,and then the residual uncertainties were discueesd.The main conclusions are summarized as follows:(1)Based on the "China ground and CMORPH fusion daily precipitation product(V1.00)" developed by the National Meteorological Information Center as the reference value data,the methods to calculating the regional average precipitation time series in China were compared by using the precipitation data of 2425 stations in China.The results showed that the linear trends and the standard deviations of the annual precipitation got by the 2.5 degree grid and the 5 degree grid area average methods were the closest to those got by the reference value from 1998 to 2012.They were the optimal and suboptimal methods respectively.If the grid was too sparse or too dense,the regional average result would be far away from the reference value.The error of the direct average method was also too large to be used to calculate the regional average precipitation in China.(2)A comparative analysis of the five-degree longitude and latitude grid average sequence of different precipitation indicators in China,western China,and eastern China from 1951 to 2016.The selection of different precipitation indicators could bring great differences in the estimation of regional average precipitation.From 1951 to 1957,because of the lack of data from stations,the regional average raw values of precipitation,precipitation days,and precipitation intensities in China had spuriously high biases,which caused large deviations in the linear trend estimates from 1951 to 2016..In the arid regions of western China,the time series of the regional average precipitation anomaly percentages will fluctuate significantly.Anomalies and normalized anomalies could better suit for different data missing conditions and different climate values.The regional average original values and anomaly values of precipitation amount and days were basically the reflections of the precipitation changes in the humid regions in eastern China.The changes of the regional average precipitation anomaly percentagew were mainly composed of the precipitation changes in the arid regions.Normalized anomaly could comprehensively reflect the precipitation changes in humid regions and arid regions.(3)This paper analyzed the temporal variation of the start and end dates of rainy season over East Asia using two different daily precipitation datasets(GHCN-D and CGP-D)for two different time periods(1951-2009 and 1951-2012).The same method was employed.GHCN-D results showed that region-averaged start and end dates of rainy season in the study region insignificantly advanced,and the duration of rainy season insignificantly increased from 1951 to 2009.However,the CGP results showed that region-averaged start dates delayed and the duration of rainy season decreased.The main reason for the disparity in regional average results was the data problem,such as the integrity issue of GHCN-D data after 2000,which was the main factor of the uncertainties in the research conclusions.Changes in the study period also had an impact on the size of the linear trend and were secondary factors of uncertainties.However,the application of data in different datasets and at different periods had less impact on the spatial distribution of site change trends.(4)2.5-degree latitude and longitude grid area-weighted regional average method was used to analyze the precipitation and extreme precipitation index changes in the Hindu Kush Himalayan region(HKH region)in recent decades based on the daily precipitation data of CGP-D.The average annual precipitation in the Hindu Kush Himalaya(HKH)region increased significantly during 1951-2013,and the number of annual precipitation days did not decrease significantly.Precipitation intensity increased significantly.In the Qinghai-Tibet plateau region,the precipitation,daily,and intensity all increased significantly.In most of India,the precipitation days increased significantly while the precipitation intensity decreased significantly.In the southwestern China and the central high mountains of southern Central Asia,the precipitation days decreased significantly and the precipitation intensity increased significantly.There was a significant increase in the amount and days of light and intense precipitation over various parts of India and northern Tibetan Plateau during 1961-2012,and the regional average series were also significantly increased but the intensity of light precipitation decreased.The amount and frequency of intense precipitation events mostly increased significantly on the Tibetan Plateau,and regional average annual extreme intense precipitation events amount and frequency significantly increased over the HKH during 1961-2012.There were significant increased tendencies in the regional average series of continuous heavy-rain and consecutive wet-days during 1961-2012.Prior to 2000,the variation of the percentages of the anomalies(PA)and standardized anomalies(S A)of the extreme precipitation indices were similar,but after 2000,the PA indices showed abnormally significant increases,but the increase trends of the SA indices were not so significant.(5)Based on the Russian datasets with corrected for precipitation measurement deviations,the Russian data in CGP was revised.Normalized precipitation anomaly(NPA)indicator was selected,and the 5-degree latitude and longitude grid area-weighted regional average method was applied to study the annual and seasonal precipitation variations during 1901-2016 in Asian regions.Regional average annual precipitation in Asian continent increased significantly during 1901-2016,and the upward trend was mainly related to a shift in cold areas such as North Asia during the mid-20th century;2)regional average spring and autumn precipitation amounts in Asian continent had similar characteristics to those of the long-term change in annual precipitation,and winter precipitation increased slightly for the entire period of 1901-2016,with a highly significant increase occurring from 1945 to 1955,while it decreased significantly during 1930s;3)North Asia and Central Asia had experienced a significantly increasing trend of precipitation for every season except for summer during 1901-2016,while South Asia had witnessed a significant decreasing trend in winter during the same period;4)during 1901-2016,the proportion of summer precipitation declined uniformly in most parts of the continent,with a significant decline found in cold areas including North Asia.In summary,this paper has basically completed the analysis of precipitation changes in Asia over the past 100 years.However,due to the residual problem of completeness and correctness of precipitation data in Asia,such as the potential inhomogeneities of precipitation data in Central Asia,some uncertainties still exict in the current conclusions,which need to be further studied and solved.