Abstract:Introduction: When the forecasted typhoon track differs from the numerical model's prediction, the estimated precipitation and wind from the model may not be reliable. Typically, forecasters receive numerical model forecasts with a delay of 4 h or more in calculation time. However, a more timely reference of precipitation and wind forecasts is required in an emergency with an approaching typhoon. Analyses of the observational data of typhoon-related characteristics, such as heavy rainfall and strong winds, from 1997 to 2021 revealed that their distribution areas are considerably affected by typhoon tracks. In this study, we developed a precipitation and wind prediction system based on the observational data of the typhoons that affected the Korean Peninsula. Methods: Typhoon tracks were categorized into west-coast landfalls, southeast landfalls, and those passing the Korea Strait. Each category affects the Korean Peninsula differently in terms of rainfall and wind. We devised a system that predicts these patterns based on incoming typhoon tracks. We can make forecasts by comparing the approaching typhoons to previous instances and analyzing their center, movement direction, and size. Observations from these past typhoons were averaged to produce a forecast grid for each new typhoon. Results: Our system, validated from 2019 to 2022, showed a wind speed root-mean-square error of 3.37 m/s and a precipitation accuracy index of 0.72. For comparison, traditional numerical models yielded 5.04 m/s and 0.75, respectively. This indicates that our system is comparably efficient and computationally less demanding. Discussion: Our system's strength is its ability to offer real-time typhoon forecasts, often faster than numerical models. However, its dependence on historical data limits its predictive power for atypical weather scenarios. It is essential to consider integrating ensemble models with these observations for enhanced accuracy. Since 2022, this system has been operational at the Korea Meteorological Administration, showing consistent reliability in forecasting.

Regional Typhoon Track Prediction Using Ensemble k-Nearest Neighbor Machine Learning in the GIS Environment

Deep Learning Approaches in Predicting Tropical Cyclone Tracks: an Analysis Focused on the Northwest Pacific Region

Prediction of typhoon tracks using a generative adversarial network with observational and meteorological data

A Data-Driven Approach for Regional-Scale Fine-Resolution Disaster Impact Prediction under Tropical Cyclones

Typhoon cloud image prediction based on enhanced multi-scale deep neural network

Real-Time Flood Disaster Prediction System by Applying Machine Learning Technique

Typhoon track prediction using satellite images in a Generative Adversarial Network

Distributed Typhoon Track Prediction Based on Complex Features and Multitask Learning

A time series image prediction method combining a CNN and LSTM and its application in typhoon track prediction

A Neural Network Framework for Fine-Grained Tropical Cyclone Intensity Prediction

Western North Pacific tropical cyclone track forecasts by a machine learning model

A Novel Deep Learning Approach for Tropical Cyclone Track Prediction Based on Auto-Encoder and Gated Recurrent Unit Networks

Fast Prediction of Typhoon Tracks Based on A Similarity Method and Gis

Tropical Cyclone Track Prediction with an Encoding-to-Forecasting Deep Learning Model

Typhoon Trajectory Prediction by Three CNN+ Deep-Learning Approaches

Assessment of Ensemble Flood Forecasting with Numerical Weather Prediction by considering Spatial Shift of Rainfall Fields

GlobeNet: Convolutional Neural Networks for Typhoon Eye Tracking from Remote Sensing Imagery

Typhoon Tracks Prediction with ConvLSTM Fused Reanalysis Data

Development of a prediction system for precipitation- and wind-causing typhoons affecting the Korean peninsula using observational data

Improved Tropical Cyclone Track Simulation over the Western North Pacific using the WRF Model and a Machine Learning Method

Evaluation of a Regional Ensemble Data Assimilation System for Typhoon Prediction