How Old Journals Reveal the Amount of Sunshine in Tokyo in the 18th and 19th Centuries

Before 1838, no reliable instrumental measurement of sunshine existed. Yet, understanding the amount of solar radiation received at a given time is crucial to grasp climate trends and their impacts on agriculture and local weather. To compensate for the lack of pyrheliometer data, scientists innovated by exploiting an unexpected material: personal journals kept in Tokyo during the 18th and 19th centuries.

Personal journals to reconstruct historical sunshine

Researchers analyzed several hundred pages of daily accounts in which residents described the weather, notably brightness and the appearance of the sky. These qualitative descriptions were converted into quantitative indices estimating the amount of solar radiation received. This method allows establishing a time series of sunshine before the era of instrumental measurements, thus offering an unprecedented insight into Tokyo's past climate.

How to extract light measurements from human accounts

Scientists used machine learning techniques to analyze the language employed in these journals. By classifying mentions of sunlight, sky clarity, or cloud presence, they could assign a light intensity score to each day. These textual data were then calibrated based on the first available instrumental measurements from 1838 onward, ensuring consistency between indirect observations and physical data.

A new tool to understand the climate before modern instruments

This approach paves the way for reconstructing historical climate datasets in other regions where human archives abound but scientific measurements are lacking. It allows refining climate models by integrating longer and more varied data, thus reducing uncertainty in past reconstructions. For Tokyo, this means better understanding sunshine fluctuations that influence crop growth and local weather cycles.

Why this research is crucial facing current climate challenges

In a context of rapid climate change, understanding natural variations of solar radiation over long periods helps distinguish anthropogenic trends from natural fluctuations. By enriching climate archives with historical data from human sources, climatologists can more precisely calibrate their models and anticipate future impacts on ecosystems and societies. This work thus demonstrates the importance of integrating interdisciplinary approaches combining history, climatology, and artificial intelligence.

Historical context and importance of local archives in Tokyo

In the 18th and 19th centuries, Tokyo, then called Edo, was the capital of a feudal but rapidly changing Japan. Residents regularly kept personal journals, a cultural practice widespread among the literate classes and merchants. These daily accounts were not limited to personal events but often included weather observations, reflecting attention to natural conditions influencing urban and agricultural life. This documentary richness constitutes a true treasure for today’s climatologists, as it offers a unique window into past climate before the advent of modern instruments.

The stakes of this research thus go beyond mere data reconstruction: they allow better understanding of how populations of the time perceived and adapted to climate variations, notably in a society where agriculture remained an essential economic and social pillar. In this, these human archives complement geological or dendrochronological sources often used, bringing a valuable socio-historical dimension to climate studies.

Scientific challenges and tactics for analyzing textual data

The main challenge lay in converting human language, subjective and sometimes poetic, into exploitable quantitative data. For this, the use of artificial intelligence and machine learning algorithms was decisive. These tools allowed identifying keywords related to brightness, such as "clear sky," "bright sun," or "threatening clouds," then classifying them according to a light intensity scale. This process required meticulous calibration by cross-referencing results with the first available instrumental measurements to ensure reliability of estimates over several decades.

This innovative tactic well illustrates the interdisciplinarity needed in modern climatology: it combines linguistics, history, computer science, and physics to overcome the limits of direct data. Moreover, by processing several hundred accounts, researchers were able to smooth individual biases linked to authors’ subjectivity, thus strengthening the robustness of results. This approach could be extended to other historical archives in different world regions, opening a new field of investigation for climate sciences.

Impact on climate understanding and future perspectives

Thanks to this reconstruction, climatologists now have a time series of sunshine for Tokyo covering a period well before the first instruments. This data is essential for modeling natural climate variations over several centuries, notably solar cycles and fluctuations linked to long-term atmospheric phenomena. Better knowledge of these cycles allows refining current climate projections, especially in a context where anthropogenic climate change adds to natural variations.

In the future, this methodology could be applied to other types of historical observations, such as descriptions of precipitation or temperatures, to build even richer climate databases. It also offers a valuable opportunity for researchers wishing to study the impact of climate on ancient societies by cross-referencing environmental data and historical archives. Finally, in a global context, this type of study contributes to understanding climate evolution in urban areas, which today are particularly vulnerable to weather variations and extreme events.

In summary

This innovative research demonstrates that personal journals from the 18th and 19th centuries in Tokyo constitute a precious source for estimating sunshine before the advent of modern instruments. By transforming human descriptions into quantifiable data thanks to artificial intelligence, scientists have succeeded in filling a significant gap in climate archives. This advance not only allows better understanding of natural fluctuations of solar radiation but also improves current and future climate models. It thus highlights the importance of addressing environmental issues with interdisciplinary methods combining history, technology, and climatology.