Extreme rainfall events variation during 1860–1909 in the Korean Peninsula: Investigation of the possible circulation mechanism by a method of analogue.

In the present paper, we made an investigation of extreme rainfall events for 1860―1909 in the Korean Peninsula, and the possible atmospheric circulation state by comparing recent rainfall variability with atmospheric circulation indices. In the present study, very significant historical rainfall records and flood damage records since the mid-19th century, modern rainfall records and global climate reanalysis data were applied. Results show: (1) Pentad extreme rainfall events had undergone a clear decrease, with an abrupt falling after shifting point of 1879―1880. (2) The monthly frequencies of extreme rainfall events revealed remarkable descent trends for June–October, which changed from 29 time/pentad to 7.3 time/pentad on average, especially for July–September. (3) Almost all climate zones except for northern land displayed significant decreasing trends, nevertheless no clear incidence of extreme rainfall event was found over the typical pluvial regions. (4) Through an analogue method by comparing the extreme precipitation weather and the atmospheric mechanism for recent decades (1961―2016), we suggested that West Pacific Subtropical High had experienced a strengthen change for the past period and the typhoon-related torrential rain processes had been suppressed for the study area, and the upper air jet stream might have been displaced poleward during 1860―1909. This study made a first investigation of the extreme rainfall events since the mid-19th century applying both instrumental and documental records in the study area, and some significant findings were obtained, which will help in understanding extreme climate variability and its mechanism of late 19th century and early 20th century in the Korean Peninsula. • Several main climate variables which are related to SR-NPKP were investigated. • 13 indices having a close connection with the variability of SR-NPKP were obtained. • Neural network model was proposed for 1-yr lag prediction of SR-NPKP. [ABSTRACT FROM AUTHOR]