Back To Top

The past and the present of weather prediction

Yoo Hee-dong, Administrator of the Korea Meteorological Administration (KMA)
Yoo Hee-dong, Administrator of the Korea Meteorological Administration (KMA)

By Yoo Hee-dong, Administrator of the Korea Meteorological Administration

A series of unexpected extreme weather events have been reported this summer in the United States and Europe. In the eastern US, more than 1,400 records for high temperatures were broken in July, and the high temperatures continued for quite some time. Coningsby in England, which has an average July temperature of 20-25 degrees, saw the UK’s hottest-ever temperature of over 40 degrees.

In addition, the southern part of the US has recently been hit by record-setting torrential rain, recording 2,300 mm rainfall in a single day. Such extreme weather events are not just a summer problem. Last winter, Texas made headlines as it was hit by the most intense cold wave in 90 years. Climate change has indeed placed the world in crisis and chaos. In an era of climate crisis, the future we face might be harsher than we ever imagined.

We are also experiencing changes in daily life driven by climate change. The Korea Meteorological Administration (KMA) walks on thin ice, especially in the summer, as our weather forecasts are challenged and tested every day. On August 8 and 9, Seoul was hit by its heaviest rainfall in 115 years. The KMA forecasted heavy rainfall of 50-100 mm per hour, but the actual rainfall was 141.5 mm per hour. Even if we had accurately forecasted a moving path or speed of a group of clouds bearing heavy rain, the rain exceeded the forecasting range based on decades of observation, making what happened not a matter of difference in the forecast amount, but an event outside decades of forecast experience.

Weather forecasts are made on the basis of the close link between forecasters and infrastructure, tangible and intangible, supporting forecasters. Forecasts can pay off when four things operate harmoniously like an orchestra -- massive amount of observations (data) collected from around the world; supercomputers (hardware) that can quickly calculate the data; numerical weather prediction models (software) that predict future weather; and forecasters’ capability.

Among others, numerical weather prediction (NWP) models are one of the most scientific tools we can use to obtain future weather information such as temperature, humidity, wind, and precipitation by inputting the current atmospheric conditions. The accuracy of NWP data forecasted by NWP models is one of the factors directly related to weather forecasts. Thus, major advanced countries in meteorology have put a lot of effort and investment over the last 20 to 30 years to have technical skills in developing NWP models.

To improve the accuracy of forecasts, the KMA has developed forecasters and gradually expanded investment in infrastructure such as weather satellites, rainfall radars, and meteorological supercomputers in order to put together the puzzle pieces of meteorological skills. However, there was one last piece of puzzle that has not been found -- software known as NWP models.

The KMA begun to develop its own NWP model in 2011 based on the leading countries’ level of infrastructure. As a result, in 2019 we finally found the last piece of the puzzle: we succeeded in developing Korea’s operational NWP model, the Korean Integrated Model (KIM). With the development of the KIM, South Korea became the 9th country in the world to develop its own NWP model, joining the ranks of global meteorological leaders. The KIM has been used in forecasting since April 2020, and its forecast performance is equivalent to those found in countries like the US and Japan.

Even though we realized the dream of developing our own model, that does not mean the KMA has stopped dreaming. As it is natural that software without upgrades is gradually destined to be less useful, continued updates and improvements in software are a must. To support generating more accurate, useful weather and climate information based on our own first NWP model, the KMA started to develop the next-generation numerical weather prediction model in the second half of 2020.

In 2026, when this project is completed, the grid size for forecasting over the Korean peninsula will be scaled down to as little as 1 km, and the model’s forecasting range will be extended from 10 to 30 days. Also, on top of the atmospheric information that can be produced by the KIM, the new model will be able to make more precise forecasts in areas such as marine weather, crops, and soil.

The KMA is again moving forward toward a new dream. The journey to turn a dream into reality will be full of failures and challenges. At the forefront of the climate crisis, the KMA will continue to make its best efforts until we realize the dream.

By Korea Herald (
Korea Herald daum