The skillful prediction of high-impact weathers remains one of the greatest scientific and societal challenges of the 21st century. Following is the goals of Forecast Research Laboratory (FRL) :

1) to enhance the understanding of atmospheric phenomena, particularly the structure of severe
    weather phenomena and their formation and development mechanisms
2) to extend a leading time for early warning of severe weathers
3) to build up short-range quantitative precipitation forecast capabilities
4) to strengthen the now-casting skills.
Development of atmospheric numerical weather prediction models
Study of synoptic and dynamical structure of the atmosphere
Short and medium-range weather prediction
SStudy of meso-scale weather phenomena
The research area of FRL is focused on the performance of research and development integrating three components such as the intensive observation, understanding of mechanism, and modeling on high-impact weather. Main researches at FRL are mainly divided into three parts: 1) Korean Enhanced Observing Program (KEOP) as a atmospheric field-based experiment, 2) development and operation of Short-range Analysis and Prediction System (SRAPS), and 3) development of nowcasting system for very short-range prediction.
Warm season quantitative precipitation forecast has been a challenging issue due to its high societal impact and the low level operational forecast skill. To provide more accurate precipitation forecast guidance , FRL has been developed the Short-Range Analysis and Prediction System (SRAPS).

The SRAPS consists of 6-hour assimilation cycle for 15km large domain and 3-hour hot start cycle for 5 km inner domain. Radar reflectivity, lightning, satellite, and METAR observations are essential components for diabatic initialization applied to short-range forecasts.

In the meantime, the assimilation system is also capable of providing comprehensive three-dimensional analysis fields every hour in about 10 min process time. Various kinds of local observations are ingested into the analysis system. The final three-dimensional analysis fields are displayed through Forecaster’s Analysis System originated from NOAA NWS/GSD (FAS) system.
One of the most practical ways to improve quantitative precipitation forecasts using numerical model is ensemble approach that samples uncertainties in models and then gives probabilistic guidance.
The Short-range Ensemble prediction System (SENS) seeks solution using a number of different short-range precipitation forecasts which come from SRAPS prediction with slightly different configuration.
A super ensemble approach based on Singular Vector Decomposition (SVD) analysis and a Probability Density Function (PDF) matching over 9th training period yield to additional improvements of precipitation forecast skill score.
High-impact weather made its effect on society, economy and environment. The World Meteorological Organization (WMO) reported that global costs of extreme weather events have increased during the last decade. Improved weather forecasts will help to reduce the social costs of high-impact weather events.
Prediction accuracy has continuously increased due to the improvement of NWP and computer performance. However, most operational models have low ability to high-impact weather forecasting.
High-impact weather systems are typically related to the rapid developing cyclones. These include high precipitation events and areas of very strong surface wind speeds. Major failures in forecasting high-impact weather arise from the combinations of inaccurate initial conditions and model errors.
Goals of this project :
1) Examine predictability and? observing systems,
2) Establish the potential
to produce significant and
statistically verifiable
improvements in forecasting high-impact weather.?
The National center for High-Impact weather Intensive
Observation has been
operated by? nimr.
Cloud streaks (black circles in the figure) are inclined from right-bottom to left-up. The movement and the velocity of convective system can be calculated from the incline of a cloud streak.
As a result, convective
systems propagate eastward with a persistent pattern in the longitude-time space even though it is forced by synoptic systems. These characteristics are helpful to determine its predictability.
Explicit initialization of microphysical species is a promising approach to alleviate the spin-up problem in the short-range (0-12 hr) precipitation forecasts. FRL has focused on applying diabatic initialization technique based on the LAPS which was originally developed by GSD/NOAA.

Initialization of microphysical species is based on three-dimensional cloud analysis using satellite, radar reflectivity, and surface cloud reports. Another crucial step is to make balance between adjusted moisture field and other dynamic variables.
A series of sensitivity studies shows that forecasts after the diabatic initialization yield to dramatic improvement of precipitation score up to 6-9 hour forecast time. This result indicates the possibility of a reasonable guidance to short-time precipitation forecast.
SCAN is an integrated suite of multi-sensor applications which detects, analyzes, and monitors convection. It also generates short-term probabilistic forecast and warning guidance for severe weather automatically within FAS.

SCAN is an integrated suite of multi-sensor applications which detects, analyzes, and monitors convection. It also generates short-term probabilistic forecast and warning guidance for severe weather automatically within FAS.

The intended benefits are
1) longer lead times on warned events
2) fewer missed events
3) increased forecaster situational awareness, and
4) reduced forecaster fatigue during warning situations.