인공신경망 앙상블 모델 기반 동중국해 북부해역의 3차원 수온장 추정

• Published in: Ocean and polar research Vol. 46; no. 4; pp. 183 - 195
• Main Authors: 이재욱, 이은주, 박재훈, Jae-Wook Lee, Eun-Joo Lee, Jae-Hun Park
• Format: Journal Article
• Language: Korean
• Published: 한국해양과학기술원 01.12.2024
• Subjects: 해양학; Convolutional Long Short-Term Memory; ensemble model; 3-D sea temperature; ConvLSTM
• ISSN: 1598-141X; 2234-7313
• DOI: 10.4217/OPR.2024020

This study used an artificial neural network model to predict the three-dimensional temperature field by capturing the non-linear relationships between input and target data. Unlike most previous studies, which focused on open ocean regions, this research develops a model specifically designed to address the unique characteristics of the coastal areas in the Northern East China Sea. Vertical temperature profiles observed from 2000 to 2022 were used as target data, while input data representing temperature structures or forcing temperature variations were utilized for training the artificial neural network model. The optimized artificial neural network model achieved a root mean square error of 1.1℃ on the test dataset. Therefore, the artificial neural network model established in this study is expected to be effectively applied to predict the three-dimensional temperature fields in the coastal seas.