Japan Climate Data Project

Project

Introduction

Paleoclimate reconstruction is essential to understanding past climate variability and its driving mechanisms within the global climate system. As Japan is a country which is blessed with an abundance of historical documents, unique climate reconstructions have been developed using qualitative meteorological information contained in historical documents. Additionally, documents with instrumental meteorological data collected before 1872, when official meteorological observations began, were recently discovered and prepared for scientific analysis.

Imaging and digitization of old papers with both qualitative and instrumental meteorological information must be carried out before these records are lost to decay. This kind of activity called “data rescue” is now taking place at many institutions all over the world. Our JCDP network has been working on data rescue for Southeast and East Asian instrumental meteorological records.

After rescuing the data, it is also important to share the data, so we are setting up this website called the “Japan-Asia Climate Data Program (JCDP)” to distribute the data. Besides the instrumental records, datasets of weather descriptions collected from Japanese historical documents will also be included and made available to the public.

Collecting historical meteorological information is significant not only for the extension of the meteorological record back in time, but also for the recovery of data in regions of the world that were hitherto poorly covered by climate data. Past climate variations reconstructed from historical meteorological data will enable us to discuss the “Little Ice Age” especially with regard to regional differences and the magnitude of the cold.

Qualitative meteorological information from historical documents

In Japan, official meteorological data collected by the Japan Meteorological Agency (JMA) are available only from 1872. There are, however, several kinds of historical documentary sources which enable reconstructions of climate variations before the 19th century in and around Japan. Historical documents including diaries of individuals, logs of clan offices, government documents and reports from temples and shrines are preserved in local libraries and museums. These often contain daily weather descriptions such as “cold”, “fine”, “rainy” and “windy”, and mention special climate-related natural phenomena such as “lake freezing” and “flower blooming”. Information about special climate-related natural phenomena such as “lake freezing” and “flower blooming” are also contained in historical documents.

Instrumental meteorological records before the foundation of official meteorological agency in Japan and other East/Southeast Asian countries.

Before the foundation of JMA, no instrumental meteorological records were believed to exist in Japan apart from qualitative weather descriptions in historical documents. However, documents containing instrumental meteorological data collected in several locations in the 19th century have been (re)discovered. Sub-daily meteorological records from the first half of 19th century were collected routinely by Dutch scientists living in Japan and by the so-called “Dutch Study” Japanese scholars who learned modern western science from the Dutch. Also, in the second half of the 19th century at the end of the period of national isolation of Japan, various Europeans (French, German, Russian) and Americans came to Japan and carried out meteorological observations. Besides Japan, for other countries such as China, Indonesia and Vietnam, the records cover the 19th and 20th centuries up to World War II.

Before the scientific analysis of the recovered data, the temperature and pressure data need to be converted to modern units and digitized into computer readable form. Secondly, the pressure data should be checked whether it was corrected for temperature, elevation, and gravity. The temperature data also need to be homogenized to compensate for changes in recording location. Then, both data sets were homogenized to account for varying observation schedules. Finally the corrected and homogenized data should go through quality check such as homogeneity tests and comparison with modern data.