Development of Real-Time Ultrafine Dust Forecast System... "World's Best Performance"

GIST Research Team Achieves 24% Higher Accuracy
Korean-Style Atmospheric Chemistry Modeling System

Domestic researchers have developed the world's highest-performance forecasting system capable of accurately predicting ultrafine dust concentrations in real time.

The Gwangju Institute of Science and Technology (GIST) announced on the 28th that a joint research team led by Professor Song Cheol-han of the Department of Earth and Environmental Engineering developed the 'Korean Air Chemistry Modeling System (K_ACheMS)' to improve the accuracy of ultrafine dust forecasts in South Korea.

This system applied the Korean-type model 'CMAQ-GIST,' which reflects the characteristics of East Asia, to the US Environmental Protection Agency's widely used model for ultrafine dust and air quality modeling (CMAQ, Community Multiscale Air Quality). In this study, the team significantly improved short-term forecast accuracy for 1 to 3 days by enhancing the 'initial conditions of the atmospheric chemistry model' using data assimilation techniques.

The research team developed a system that generates optimal initial conditions by combining ultrafine dust concentration information obtained from the Korean geostationary satellite, ground air quality observation networks, and the Korean model 'CMAQ-GIST.' Generally, the initial conditions of atmospheric chemistry models are set based on climate data, which involves considerable uncertainty. However, this system applies real-time observational data through data assimilation techniques, allowing the initial conditions to reflect the current air quality information.

Real-time forecasts conducted under optimal initial conditions showed an accuracy rate 24% higher than that of the European Centre for Medium-Range Weather Forecasts, known for its world-class performance. Since January last year, the research team has been performing real-time air quality forecasts using this system and has made them publicly available on the GIST website. This system accurately predicted ultrafine dust concentrations much better than the European Centre for Medium-Range Weather Forecasts during three high-concentration ultrafine dust events that occurred since last year.

Professor Song emphasized, "We are focusing our research efforts on combining the Korean air quality model with an artificial intelligence system currently under development to achieve synergistic effects," adding, "It will evolve into the best air quality modeling system leading the global ultrafine dust and air quality forecasting field."

The research results were published on the 23rd of last month in 'npj Climate and Atmospheric Science,' a sister journal of the international journal Nature.

This content was produced with the assistance of AI translation services.