Cutting through the noise

Visiting Scientist Shin Koyamatsu uses a statistical technique to home in on the most meaningful infrared sounder data

In preparation for processing data from Himawari-10, Japan’s first satellite that will carry a hyperspectral infrared sounder, Shin Koyamatsu is spending a year at EUMETSAT learning from its wealth of infrared sounder data and the techniques used to process them.

Last Updated

23 November 2023

Published on

22 November 2023

Shin Koyamatsu found his way toward investigating weather phenomena through an interest in puzzles. A high school physics class ignited his passion for piecing together various scientific laws, which he likened to putting together a puzzle. This interest only strengthened in college, where he studied astronomy out of a desire to combine various observations in order to clarify phenomena in the universe.

Now, as a visiting scientist from the Japan Meteorological Agency, Koyamatsu is turning his attention toward yet another complex puzzle – analysing data from hyperspectral infrared sounders. These satellite instruments, which measure features of the atmosphere such as temperature and humidity, collect an abundance of data points numbering in the thousands. And while the quantity of data is incredibly beneficial, unfortunately, amid the useful data lurks a profusion of meaningless information.

Since March, Koyamatsu has been hard at work at EUMETSAT sifting out the noise from the signal in infrared sounder data using a technique known as principal component analysis. This powerful statistical technique makes it possible to both compress the data, making the total volume smaller, and capture their large patterns.

“The volume of data from infrared sounders is enormous, so it is very important to know what information to extract from it. How we visualise the data is also important because what we see completely changes depending on which aspects of the data we cut out. I am learning more about it every day, as handling huge amounts of data requires a lot of ingenuity,” he said.

Koyamatsu is grateful for the opportunity to do this work at EUMETSAT, which has been operating the Infrared Atmospheric Sounding Instrument on polar-orbiting Metop satellites since 2006. Being able to draw on the extensive experience at EUMETSAT in processing this type of data as well as the expertise being currently built in preparation for the upcoming launch of the first Meteosat Third Generation - Sounder, which will carry an InfraRed Sounder, has been helpful.

“I think EUMETSAT is the best place for my work because it has great deal of knowledge about infrared sounder operation, data processing, and other related issues,” he said.

When he concludes his work at EUMETSAT this coming March, he will leave equipped with knowledge he can use as the Japan Meteorological Agency prepares to launch Himawari-10, its first satellite that will carry a hyperspectral infrared sounder.

As his time in Darmstadt is winding down, Koyamatsu took the opportunity to reflect on his experience in Germany.

“I find the people in Germany to be friendlier and kinder than in Japan,” he said “People in Japan, especially in Tokyo, tend to be less interested in their surroundings. What was interesting was that when I mentioned this to someone from EUMETSAT who had travelled to Japan on business, she said that Japanese people were kind.

“Maybe it is the same all over the world that people are kind to foreigners!”