Sea surface temperature prediction by stacked generalization ensemble of deep learning

IF 2.3 3区 地球科学 Q2 OCEANOGRAPHY
Hao Dai , Famei Lei , Guomei Wei , Xining Zhang , Rui Lin , Weijie Zhang , Shaoping Shang
{"title":"Sea surface temperature prediction by stacked generalization ensemble of deep learning","authors":"Hao Dai ,&nbsp;Famei Lei ,&nbsp;Guomei Wei ,&nbsp;Xining Zhang ,&nbsp;Rui Lin ,&nbsp;Weijie Zhang ,&nbsp;Shaoping Shang","doi":"10.1016/j.dsr.2024.104343","DOIUrl":null,"url":null,"abstract":"<div><p>Accurate sea surface temperature (SST) prediction is of great significance for fishery farming, marine ecological protection, and planning of maritime activities. In this paper, the stacked generalization ensemble is demonstrated for SST prediction to improve every single deep-learning model. Long-term high-resolution satellite-derived SST is used with the sub-regions of the Taiwan Strait and East China Sea taken as the study area. We select the Multilayer Perceptron, Long short-term memory (LSTM), Convolutional Neural Networks (CNN), CNN-LSTM as individual learners, and Convolutional LSTM as the meta-learner. The individual learners are trained and validated on the retained data subset I, while the meta-learner is trained and validated by constructing the samples with the predictions of validated individual learners on the retained data subset II. The two types of models are evaluated on the same test dataset with root-mean-square error and coefficient of efficiency as the scoring criteria. We find that the meta-model outperforms any individual model and other baselines for the one-day-/three-day-ahead forecasts in the Taiwan Strait and one-day-/three-day-/five-day-ahead predictions in the East China Sea. Furthermore, when the lead time is 1 day and 3 days, the meta-model has a better spatial distribution of prediction metrics across all grid points in the Taiwan Strait sub-area. For the East China Sea sub-region, the meta-model advantage is extended to the lead time of 5 days. Probably due to the higher quality of offshore satellite data, the prediction ability enhancement of the stacked ensemble applied in the East China Sea is better than that in the Taiwan Strait. The better-performing meta-model prediction suggests that the stacked generalization ensemble is encouraging and promising for improving the short-term prediction of the daily SST field.</p></div>","PeriodicalId":51009,"journal":{"name":"Deep-Sea Research Part I-Oceanographic Research Papers","volume":null,"pages":null},"PeriodicalIF":2.3000,"publicationDate":"2024-06-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Deep-Sea Research Part I-Oceanographic Research Papers","FirstCategoryId":"89","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S0967063724001134","RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"OCEANOGRAPHY","Score":null,"Total":0}
引用次数: 0

Abstract

Accurate sea surface temperature (SST) prediction is of great significance for fishery farming, marine ecological protection, and planning of maritime activities. In this paper, the stacked generalization ensemble is demonstrated for SST prediction to improve every single deep-learning model. Long-term high-resolution satellite-derived SST is used with the sub-regions of the Taiwan Strait and East China Sea taken as the study area. We select the Multilayer Perceptron, Long short-term memory (LSTM), Convolutional Neural Networks (CNN), CNN-LSTM as individual learners, and Convolutional LSTM as the meta-learner. The individual learners are trained and validated on the retained data subset I, while the meta-learner is trained and validated by constructing the samples with the predictions of validated individual learners on the retained data subset II. The two types of models are evaluated on the same test dataset with root-mean-square error and coefficient of efficiency as the scoring criteria. We find that the meta-model outperforms any individual model and other baselines for the one-day-/three-day-ahead forecasts in the Taiwan Strait and one-day-/three-day-/five-day-ahead predictions in the East China Sea. Furthermore, when the lead time is 1 day and 3 days, the meta-model has a better spatial distribution of prediction metrics across all grid points in the Taiwan Strait sub-area. For the East China Sea sub-region, the meta-model advantage is extended to the lead time of 5 days. Probably due to the higher quality of offshore satellite data, the prediction ability enhancement of the stacked ensemble applied in the East China Sea is better than that in the Taiwan Strait. The better-performing meta-model prediction suggests that the stacked generalization ensemble is encouraging and promising for improving the short-term prediction of the daily SST field.

通过深度学习的堆叠泛化集合预测海面温度
准确的海面温度(SST)预测对渔业养殖、海洋生态保护和海洋活动规划具有重要意义。本文展示了用于 SST 预测的堆叠泛化集合,以改进每个单一的深度学习模型。本文以台湾海峡和东海为研究区域,使用长期高分辨率卫星衍生 SST。我们选择多层感知器、长短期记忆(LSTM)、卷积神经网络(CNN)、CNN-LSTM 作为单个学习器,卷积 LSTM 作为元学习器。单个学习器在保留数据子集 I 上进行训练和验证,而元学习器则通过在保留数据子集 II 上用已验证的单个学习器的预测构建样本来进行训练和验证。以均方根误差和效率系数作为评分标准,在同一测试数据集上对这两种模型进行评估。我们发现,在台湾海峡提前一天/三天的预测和东海提前一天/三天/五天的预测中,元模型优于任何单个模型和其他基线。此外,当提前期为 1 天和 3 天时,元模式在台湾海峡次区域所有网格点的预测指标空间分布更好。在东海子区域,元模型的优势扩大到了 5 天前。可能是由于近海卫星数据质量较高,在东海应用的叠加集合的预报能力增强效果优于台湾海峡。更好的元模式预测结果表明,叠加广义集合在改善日海温场的短期预测方面是令人鼓舞和充满希望的。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 求助全文
来源期刊
CiteScore
4.60
自引率
4.20%
发文量
144
审稿时长
18.3 weeks
期刊介绍: Deep-Sea Research Part I: Oceanographic Research Papers is devoted to the publication of the results of original scientific research, including theoretical work of evident oceanographic applicability; and the solution of instrumental or methodological problems with evidence of successful use. The journal is distinguished by its interdisciplinary nature and its breadth, covering the geological, physical, chemical and biological aspects of the ocean and its boundaries with the sea floor and the atmosphere. In addition to regular "Research Papers" and "Instruments and Methods" papers, briefer communications may be published as "Notes". Supplemental matter, such as extensive data tables or graphs and multimedia content, may be published as electronic appendices.
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信