Comparative analysis and enhancing rainfall prediction models for monthly rainfall prediction in the Eastern Thailand

IF 1.6 Q2 MULTIDISCIPLINARY SCIENCES

MethodsX Pub Date : 2024-12-10 DOI:10.1016/j.mex.2024.103094

Preeyanuch Chuasuk , Tachanat Bhatrasataponkul , Aniruj Akkarapongtrakul

{"title":"Comparative analysis and enhancing rainfall prediction models for monthly rainfall prediction in the Eastern Thailand","authors":"Preeyanuch Chuasuk , Tachanat Bhatrasataponkul , Aniruj Akkarapongtrakul","doi":"10.1016/j.mex.2024.103094","DOIUrl":null,"url":null,"abstract":"<div><div>Rainfall prediction is a crucial aspect of climate science, particularly in monsoon-influenced regions where accurate forecasts are essential. This study evaluates rainfall prediction models in the Eastern Thailand by examining an optimal lag time associated with the Oceanic Niño Index (ONI). Five deep learning models—RNN with ReLU, LSTM, GRU (single-layer), LSTM+LSTM, and LSTM+GRU (multi-layer)—were compared using mean absolute error (MAE) and root mean square error (RMSE). A novel hybrid deep learning model was developed with respect to different conditions of the El Niño and Southern Oscillation (ENSO).<ul><li><span>-</span><span><div>Our research compared the performance of five deep learning models in predicting monthly rainfall over five selected stations in the Eastern Thailand.</div></span></li><li><span>-</span><span><div>Different lag times were initially verified to optimize the time-interdependency between ONI and local meteorological parameters.</div></span></li><li><span>-</span><span><div>Our novel hybrid model demonstrated an improved accuracy across three distinct climate phases: El Niño, La Niña, and neutral events.</div></span></li></ul></div></div>","PeriodicalId":18446,"journal":{"name":"MethodsX","volume":"14 ","pages":"Article 103094"},"PeriodicalIF":1.6000,"publicationDate":"2024-12-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11718337/pdf/","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"MethodsX","FirstCategoryId":"1085","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S2215016124005454","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"MULTIDISCIPLINARY SCIENCES","Score":null,"Total":0}

引用次数: 0

Abstract

Rainfall prediction is a crucial aspect of climate science, particularly in monsoon-influenced regions where accurate forecasts are essential. This study evaluates rainfall prediction models in the Eastern Thailand by examining an optimal lag time associated with the Oceanic Niño Index (ONI). Five deep learning models—RNN with ReLU, LSTM, GRU (single-layer), LSTM+LSTM, and LSTM+GRU (multi-layer)—were compared using mean absolute error (MAE) and root mean square error (RMSE). A novel hybrid deep learning model was developed with respect to different conditions of the El Niño and Southern Oscillation (ENSO).

-
Our research compared the performance of five deep learning models in predicting monthly rainfall over five selected stations in the Eastern Thailand.
-
Different lag times were initially verified to optimize the time-interdependency between ONI and local meteorological parameters.
-
Our novel hybrid model demonstrated an improved accuracy across three distinct climate phases: El Niño, La Niña, and neutral events.

Abstract Image

查看原文本刊更多论文

泰国东部月降水预报的比较分析和改进模式。

降雨预测是气候科学的一个重要方面，特别是在受季风影响的地区，准确的预报是必不可少的。本研究通过检查与海洋Niño指数（ONI）相关的最佳滞后时间来评估泰国东部的降雨预测模型。使用平均绝对误差（MAE）和均方根误差（RMSE）对五种深度学习模型——rnn与ReLU、LSTM、GRU（单层）、LSTM+LSTM和LSTM+GRU（多层）进行了比较。针对厄尔尼诺Niño和南方涛动（ENSO）的不同条件，建立了一种新的混合深度学习模型。我们的研究比较了五个深度学习模型在预测泰国东部五个选定站点的月降雨量方面的表现。-最初验证了不同的滞后时间，以优化ONI与当地气象参数之间的时间相互依赖性。-我们的新型混合模型在三个不同的气候阶段（El Niño， La Niña和中性事件）中显示出更高的准确性。

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