Accumulated rainfall data is useful for evaluating the evidences of climatic and hydrological variation and for applying socio-economic impacts on natural disasters. Additionally, it is useful when monitoring flood and drought as well as making decisions regarding agricultural activities and water management; these decisions involve irrigation management and hydroelectric power production. We are interested in the role of the long-term periodic cycle of rainfall data. In order to find periodicity of heavy rainfall events, we employ power spectrum analysis and wavelet analysis analyzing rapidly changing transit signals.
According to National Institute of Korean Histories (2005a), the observation of precipitation by the government started in 1441 during the Joseon Dynasty (1392-1910). Seungjeongwon Ilgi (the Diaries) was steadily written. The Diaries are independently written by high ranked officers in the palace office without the king's personal gain. The Diaries recorded the public works with weather phenomena and amount of depth of precipitation. However, the Diaries before 1552 were burnt due to the invasion of Japan during 1592-1598, and the Diaries in 1552-1621 were burnt again due to the rebellion in 1624. Subsequently, the revised Diaries were burnt in 1744. The existing Diaries were revised in the National Archives of the Diaries (548 books) in 1746. In 1888, again 361 series of the Diaries in 1851-1888 were lost, and in 1890, a fire broke out. Although data was lost due to incidental accidents, fortunately, they were able to repair each time (National Institute of Korean Histories, 2005b).
The precipitation data shown in this paper is from the Diaries after 1778. Now we are starting to process quality check of precipitation data from the Diaries.
Among the Diaries, record from 1776 have been systematically recorded in a daily basis. This is the only actual precipitation data of the mid-latitude region from the 18th century.
In this study, we try to find out the statistical characteristics of rainfall based on time series data over 241 years through a rudimentary study. It will be fundamental and valuable information to understand the hydrological variability caused by precipitation variation in the mid-latitude environment, and it will provide clues to estimate other climate predictabilities.
2. Dataset
The data used in this study are divided into two periods: Period-1 observed in the 1777-1907 and the Period-2 recorded in the National Climate Data Center (NCDC) in KMA from 1908-2018. Seungjeongwon Ilgi was registered in UNESCO in 2001 as an official document written by the Kings’ secretariat in 1623-1907. This data is used to statistically analyze significant annual rainfall in Seoul during 1778-2016 (Lee and Kim, 2018). The data used in the study were analyzed from June to August, and further explained the reasons for taking summer data in our studies (Lee et. al. 2017).
Unfortunately, dataset of June 1831 August 1894 and June 1903 in Period-1 have missing values, and of June, July, August 1951 and August 1952 in Period-2. In this study, the missing data were replaced with monthly average values for each period. As a result predominant periods are not same to current research.
In order to rescue missing dataset for the future works, we will try to investigate historic records in the old official documents in Joseon Dynasty and analyze objective, estimate missing values using available adjacent precipitation observation. The reconstruction of missing data in Period-1 is available to recover and to take quality check through current electronic Seungjeongwon Ilgi. However, the missing data in the modern observation data in Period-2 is a result of the Korean War during 1950-1953. We will search for taking data restoration methodology of missing values treatments.
3. Analysis
The data are the only precipitation records in the world. The data of annual precipitations during 1778-2018 (241 years long) were analyzed by calculating monthly and seasonal precipitations. The analytical methods were power spectrum analysis and wavelet analysis to find statistically significant periods.
The power spectrum analysis was analyzed by two individual Periods and combining two Periods. Power spectrum analysis was chosen as an easy way to identify the statistically significant signal. {During the Period-1 of 130 years, the periodicity of 10 to 11 years appeared, and during the Period-2 of 111 years, a period of about 10 years together with a short cycle of three years appeared.} During the 241 years between 1778 and 2018, a period of 15 years in the short cycle of less than 50 years and a period of 11 years in more than 50 years were prominent, which proved similar to the sunspot cycle. The change of precipitation in climate change became an opportunity to explain strong correlation with the cycle of sunspot.
The wavelet analysis is shown three predominant periods in a given year as follows; periodicity of 48-56 years in the year of 1897, periodicity of 32-40 years in the year of 1920, and periodicity of 28-32 years in the year of 1963.
References
Lee, J. W., D. S .Kim, and H. D. Yoo, 2017: Statistical review of annual precipitation at Seoul in 1778-2015. 31st Conference on Hydrology. 97th Annual Meeting AMS in Seattle, WA USA.
Lee J. W. and D. S. Kim, 2018: Meteorological review on ten significant annual precipitations in Seoul during 1778-2016. 32nd Conference on Hydrology. 98th Annual Meeting AMS in Austin, TX USA.
National Institute of Korean Histories, 2005a: Veritable Records of the Joseon Dynasty (Joseon Wangjo Sillok): Records of King Se-Jonog Vol 93 (http://sillok.history.go.kr)
National Institute of Korean Histories, 2005b: The Daily Records of Royal Secretariat of Joseon Dynasty (http://sjw.history.go.kr)
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner