[M-IS22] History X Earth and Planetary Science

[MIS22-08] Objective evaluation of degree of correspondence between weather record of diaries and a meteorological parameter

*Kenjiro Sho¹, Yoshimi Ikeda², Mika Ichino³, Junpei Hirano⁴, Kooiti Masuda⁵, Asanobu Kitamoto^3,6 (1.Nagoya Institute of Technology, 2.Ikeda Gakuen Ikeda Senior High School, 3.Center for Open Data in the Humanities, 4.Teikyo University, 5.Rissho University, 6.National Institute of Informatics)

Keywords:Weather record of diaries, Paleoclimate proxy data, Quantification, Overlapping rate

Daily weather record of old diaries is very useful for climate reconstruction in historical times because of its exceptionally high temporal resolution as paleoclimate proxy data. However, there is difficulty in estimating quantitative meteorological variables such as precipitation from their qualitative and subjective weather description. In this study, we attempt to develop a method to quantify weather record of diaries by using old diaries written in Meiji and Taisho period (late 19th to early 20th centuries) that are comparable to instrumental meteorological data.
Here we introduce "overlapping rate" as an index of degree of correspondence between weather record of diaries and a meteorological parameter selected. This can be defined as the overlapping area of relative frequency distributions of the meteorological parameter corresponding to two weather categories such as “precipitated” and “non-precipitated” or “fine” and “cloudy.” Smaller (nearer to 0) value of this index means better correspondence between selected weather categories and the meteorological parameter while larger (nearer to 1) value means worse correspondence. For example, the value of overlapping rate of daily precipitation corresponding to weather categories “precipitated” and “non-precipitated” of “Takuma diary” (Kyoto, 1897-1911) was calculated as 0.28 and the value of overlapping rate of daily mean cloud cover corresponding to weather categories “fine” and “cloudy” was calculated as 0.52, which means the boundary of weather categories “precipitated” and “non-precipitated” is clearer than “fine” and “cloudy”. Also, we calculated the values of overlapping rate for these cases using sub-daily (four-hourly) meteorological data and found the value is lower for daytime data than for night-time data for both cases, which means weather description of diaries reflect situation more in daytime than in night-time. Likewise, this index can be useful to search the optimum method to categorize weather record and the best corresponding meteorological parameter as well as to evaluate the quality of daily weather record as paleoclimate proxy data.