Reconstruction And Variation Of Tropical Cyclone Frequency Over East Asia Since The Early 20^th Century

Tropical cyclones,as intense weather systems generated over tropical or subtropical oceans,have a serious impact on the whole world.Before the satellite era,especially before 1950,the observation of tropical cyclones basically relied only on meteorological stations,lighthouses and ship observations on the ocean.Different organizations and institutions have different observation standards and methods for tropical cyclones,which resulting in great heterogeneity of historical time series.With the increasing concern about extreme climate change and its future impacts prediction and forecast,the need for the reconstruction and analysis of tropical cyclone time series is increasing.In this paper,we collect early(before 1950)instrumental hourly pressure data,integrate the data.Interpolation and homogenization of pressure series with good continuity and high quality are carried out.Finally,obtain a set of 24 stations of dayby-day pressure data series in East Asia.Using modern high-quality pressure and tropical cyclone observations,a set of technical methods for identifying tropical cyclone processes using instrumental data is developed.Based on the early daily pressure data series,the tropical cyclone processes affecting East Asia are identified.Climatology and climate change analysis of the identified tropical cyclone sequences have been studied.The main results are as follows:(1)A 120-year homogeneity daily mean surface pressure data set in East Asia was established.Four sets of(ISD,ISPD,ACRE-China,SURF)early hourly pressure data for East Asia were collected and integrated.The missing information in ACRE-China data was completed,and the units,format and time zones of SURF data were unified.The integrated dataset has been quality controlled.The 24 pre-processed stations with high observational quality,good data integrity and a certain degree of independence between 1900 and 1950 were selected as the main stations.Spatial-temporal interpolation technique were used to interpolated the main site by using nearby station(nearby stations within 150 km of the main station).And the modern(1951-2020)ISD and National Information Centre of the China Meteorological Administration pressure data were used for splicing main station series and perform intraday natural spline interpolation.A set of 24 daily barometric data from 1900-2020 was obtained,of which four daily observations(3,9,15,21)were homogenized and a daily average was obtained to get a set of 24 homogenized daily pressure data series for the main stations in East Asia.(2)A technique based on percentile threshold of geostrophic wind speed is developed to identify tropical cyclones over East Asia.Using modern high-quality pressure observation data,a pressure triangle constructed from any three stations,capable of inverting the geostrophic wind sequence in the area of the pressure triangle.Six pressure triangles that are formed from the 5 sites(Beihai,Shenzhen,Fuzhou,Sanya and Xujiahui).For each triangle,the geostrophic wind speeds are calculated from pressure values for the same time at the three sites that form the triangle.Fixing the 99%quantile threshold as the tropical cyclone impact threshold.All tropical cyclone processes that affected this region in 2016 and 2017 can be identified by this threshold.(3)The time series of tropical cyclones affecting East Asia from 1900 to 2020 are reconstructed.Combining the pressure data from 24 main stations in East Asia and the identifying tropical cyclone technique were using to construct three pressure triangles at different latitudes in East Asia.Identification of tropical cyclone processes affecting three different latitudes by setting thresholds for each of the three triangles.The results show that the identified tropical cyclone sequences are in good agreement with the measured tropical cyclone sequences,with the best results identified in the southern and central regions.Tropical cyclones affecting East Asia throughout the period 1900-2020 were identified,with the highest number of tropical cyclones identified in the southern region and decreasing in the central and northern regions in that order.The series of historical tropical cyclones in the three regions all show fluctuating characteristics,with decreasing and increasing trends in the southern and northern regions respectively,and insignificant changes in the central region.The analysis also shows that the PDO index is more significantly related and correlated with tropical cyclones identified in the northern region.