KoMet

KoMet数据集是由韩国科学技术院人工智能研究所开发，专注于韩国半岛的降水预测。该数据集包含来自全球数据同化和预测系统-韩国综合模型（GDAPS-KIM）的预测输出以及自动气象站（AWS）的观测数据，共计484个观测点。数据集覆盖了2020年和2021年7月和8月，这两个月是韩国半岛降水最多的时期。KoMet数据集旨在通过结合数值天气预测（NWP）和深度学习（DL）技术，提高降水预测的准确性和解释性，为社会提供更可靠的天气预测服务。

The KoMet dataset was developed by the Institute of Artificial Intelligence at the Korea Advanced Institute of Science and Technology (KAIST), focusing on precipitation prediction over the Korean Peninsula. This dataset includes forecast outputs from the Global Data Assimilation and Prediction System-Korea Integrated Model (GDAPS-KIM) and observational data from Automatic Weather Stations (AWS), with a total of 484 observation sites. The dataset spans July and August of 2020 and 2021, the two months with the peak precipitation across the Korean Peninsula. The KoMet dataset aims to improve the accuracy and interpretability of precipitation predictions by combining Numerical Weather Prediction (NWP) and Deep Learning (DL) technologies, providing more reliable weather forecast services for society.