cat articles/rainy-season

Was the end of Japan's 2022 rainy season unusually early?

Yesterday, the rainy season was considered to have ended in the Kanto-Koshin region. News reports said it was unusually early, 22 days earlier than normal, where "normal" means the statistics from the past 30 years. News reports usually avoid talking about standard deviation because they need to communicate clearly to the general public, but I wanted to check how unusual "unusually early" actually was.

For rainy-season start and end dates, I used data published by the Japan Meteorological Agency from 1951 onward. There is even a year, 1993, where the start and end could not be identified. The data was inconvenient to handle directly, so I pasted it into a spreadsheet and cleaned it up roughly.

Looking at about 70 years of data, the standard deviation of the rainy-season end date is 8.6 days. The 2022 date is 21 days earlier than the 70-year average. If we assume the data is normally distributed, it falls outside the 95% confidence interval, so it is fair to call it unusually early.

However, looking only at the past 10 years, the rainy season has already been ending earlier. The 70-year average end date is July 18, while the 10-year average is July 14. The average start date differs by only one day, so the end date has moved noticeably earlier. The standard deviation of the 10-year average end date is 12.59 days. Compared with the 10-year average, the 2022 end date differs by 17.2 days. If we look only at the narrow window of the past 10 years, it probably falls somewhere around a 75-85% confidence interval, so it feels like something that happens from time to time. The 2022 date is also not that different from 2018, only two days earlier.

In other words, viewed across the past 70 years, it is unusually early. Viewed across the past 10 years, it is only something that occasionally happens. The data makes me feel that changes in the global environment, including environmental issues, may already be affecting the timing of the rainy season's end.

cat related_articles/rainy-season.yaml

  1. Reading Basic Statistics by Kimio Miyakawa: statistics before machine learningAfter several months of studying machine learning, I realized I was missing the statistical foundations needed to understand data, experiments, estimation, testing, and model evaluation.
  2. Launching AI News and how I used OpenAI behind itI launched AI News, a site that collects AI, data science, and machine learning topics and summarizes them into three lines with AI. This article describes why I built it and how I used OpenAI APIs for classification and summarization.
  3. Releasing a Japanese StaticEmbedding Model for Practical 100x Faster Text EmbeddingsI released static-embedding-japanese, a fast non-Transformer embedding model for Japanese and English text, and evaluated it on JMTEB.