Machine learning: 2
Life Sciences
Information and Communication
Nanotechnology / Materials
Manufacturing Technology
Human and Social Sciences
Energy
Environment
Tourism / Community development
Arctic Research
Social Infrastructure
Open Facilities
View All
View Detail-
Uncovering the relationships among air pollution (aerosols, PM2.5), wildfires, snow and ice, and climate change in the Arctic and cryosphere for a sustainable society in the future!
An atmospheric scientist working in the Arctic and cryosphere, tackling wildfires, air quality, snow and ice, and climate change
I am conducting a wide range of research on wildfires and their air quality (aerosols and PM2.5), including analysis and prediction of the factors that cause them and impact assessment (climate, health, economy, etc.), using various research methods from observation to data analysis and modeling.
-
Figure 1b from Yasunari et al. (2018, Sci. Rep.). Daily mean PM2.5 concentration on July 25, 2014, calculated using NASA's MERRA-2 reanalysis data. The white circle indicates the location of Sapporo City.
-
A commercial version of the PM2.5 measurement system for cold regions, updated from the prototype in Yasunari et al. (2022, J. Environ. Manage.). Anyone can purchase it from Tanaka Co., Ltd. (http://kktanaka.co.jp/products; the iron box and the low-cost PM2.5 sensor must be obtained separately)
-
A climate (atmospheric circulation) pattern that can likely cause co-occurrences of European heatwaves and wildfires in Siberia and subarctic North America (Alaska and Canada), as discovered in Yasunari et al. (2021, Environ. Res. Lett.): the pattern was named the circum-Arctic wave (CAW) pattern because it is a pattern in which anticyclonic circulation is arranged to surround the Arctic. The figure is from Figure 9 of the paper (created by the current “Science Manga Studio Co., Ltd.”: https://www.sciencemanga.jp/).
-
In Yasunari et al. (2024, Atmos. Sci. Lett.), the authors used the commercial PM2.5 measurement system for cold regions and, for the first time, performed the local ambient air quality observation (i.e., PM2.5 measurement) in Qaanaaq, northwest Greenland, in the summer of 2022. They also captured the worsened air quality during the local open waste burning (the figure is the Graphic Abstract of the paper).
Research
In recent years, we have been hearing more and more news about wildfires. Large-scale wildfires can transport air pollution (PM2.5) not only to the area where they occur but also to areas downwind, potentially affecting the people who live there. For this reason, it is necessary to identify the causes of wildfires and the atmospheric aerosols (air pollution) they produce and assess the diverse effects (such as climate, health, social and economic) that follow. In addition, it is extremely important to predict these effects based on the knowledge gained from the perspective of taking measures for people living in the downwind area from where the fires occur. To achieve the above objectives, we have developed a portable PM2.5 measurement system for cold regions, conducting multi-location observations of air quality such as PM2.5, analyzing large-scale global data (satellite, model, re-analysis data, etc.), and conducting research using various methods such as machine learning prediction (we are also conducting joint research with NASA and interdisciplinary research).
Teppei J. Yasunari Specially Appointed Associate ProfessorPh.D. in the field of Earth System Science -
-
Classification of Genetic Information Using Machine Learning
Predicting the binding of compounds to cell receptors
Various receptors on the surface of cells play important roles in maintaining homeostasis and environmental responses, but it is difficult to identify compounds that can bind to them. We propose a method for narrowing down the candidates for binding compounds by using machine learning.
Research
Although the human genome has been deciphered and many of the genes have been elucidated, the structure and function of receptors, which play an important role in homeostasis and environmental responses, have not been fully elucidated, because most of them are membrane proteins and their expression levels are low. Many receptors, however, are expected to be major targets for drug discovery in the future because of their functional aspects, and are thought to be the factors that cause individual differences. We are applying machine learning technology to efficiently narrow down compounds that can bind to receptors.
Toshinori Endo Professor
