Predictive modeling of river blockage severity from debris flows: Integrating statistical and machine learning approaches with insights from Sichuan Province, China

IF 5.4 1区 农林科学 Q1 GEOSCIENCES, MULTIDISCIPLINARY
Wei Zhou, Yaping Zhou, Renwen Liu, Huaqiang Yin, Haowen Nie
{"title":"Predictive modeling of river blockage severity from debris flows: Integrating statistical and machine learning approaches with insights from Sichuan Province, China","authors":"Wei Zhou,&nbsp;Yaping Zhou,&nbsp;Renwen Liu,&nbsp;Huaqiang Yin,&nbsp;Haowen Nie","doi":"10.1016/j.catena.2024.108576","DOIUrl":null,"url":null,"abstract":"<div><div>River blockages caused by debris flows pose serious threats to the environment and infrastructure. This study introduces the river blockage index (<em>RBI</em>) as a key measure of river blockage severity. We collected data from 60 debris flow events to build a comprehensive dataset. To enhance model robustness and accuracy, we optimized variable selection using multicollinearity analysis and the Akaike information criterion (<em>AIC</em>). Four statistical models were developed, including multiple linear, logarithmic, power, and exponential regressions. We also constructed models based on machine learning algorithms, including random forests and gradient boosting decision trees, and tested them using 5-fold cross-validation. After confirming that the training dataset met linear statistical assumptions, we built robust regression models. We tested the significance of the regression equations and coefficients using <em>F</em>-tests and <em>t</em>-tests. Hyperparameters of the machine learning algorithms were optimized through Bayesian methods. Model performance was evaluated using metrics such as <em>R</em><sup>2</sup>, adjusted <em>R</em><sup>2</sup>, mean absolute error (<em>MAE</em>), mean relative error (<em>MRE</em>), and root mean square error (<em>RMSE</em>). Results show that the most important factors influencing <em>RBI</em> are catchment area (<em>A</em>) and the discharge ratio between the debris flow and the main river (<em>Q</em>). Among the statistical models, the logarithmic and power models performed best due to their simplicity and efficiency. The random forest model demonstrated the highest predictive accuracy and stability overall. By combining statistical methods with machine learning, we improved prediction accuracy and provided practical guidance for disaster prevention strategies. This approach overcomes the limitations of numerical simulations and experimental studies, offering a more flexible and efficient method for <em>RBI</em> prediction. Future work will extend these findings to other geological settings to further enhance model adaptability and performance.</div></div>","PeriodicalId":9801,"journal":{"name":"Catena","volume":"248 ","pages":"Article 108576"},"PeriodicalIF":5.4000,"publicationDate":"2024-11-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Catena","FirstCategoryId":"97","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S0341816224007732","RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"GEOSCIENCES, MULTIDISCIPLINARY","Score":null,"Total":0}
引用次数: 0

Abstract

River blockages caused by debris flows pose serious threats to the environment and infrastructure. This study introduces the river blockage index (RBI) as a key measure of river blockage severity. We collected data from 60 debris flow events to build a comprehensive dataset. To enhance model robustness and accuracy, we optimized variable selection using multicollinearity analysis and the Akaike information criterion (AIC). Four statistical models were developed, including multiple linear, logarithmic, power, and exponential regressions. We also constructed models based on machine learning algorithms, including random forests and gradient boosting decision trees, and tested them using 5-fold cross-validation. After confirming that the training dataset met linear statistical assumptions, we built robust regression models. We tested the significance of the regression equations and coefficients using F-tests and t-tests. Hyperparameters of the machine learning algorithms were optimized through Bayesian methods. Model performance was evaluated using metrics such as R2, adjusted R2, mean absolute error (MAE), mean relative error (MRE), and root mean square error (RMSE). Results show that the most important factors influencing RBI are catchment area (A) and the discharge ratio between the debris flow and the main river (Q). Among the statistical models, the logarithmic and power models performed best due to their simplicity and efficiency. The random forest model demonstrated the highest predictive accuracy and stability overall. By combining statistical methods with machine learning, we improved prediction accuracy and provided practical guidance for disaster prevention strategies. This approach overcomes the limitations of numerical simulations and experimental studies, offering a more flexible and efficient method for RBI prediction. Future work will extend these findings to other geological settings to further enhance model adaptability and performance.
基于泥石流的河流堵塞严重程度预测建模:将统计和机器学习方法与中国四川省的见解相结合
泥石流造成的河流堵塞对环境和基础设施构成严重威胁。本文引入河流堵塞指数(RBI)作为衡量河流堵塞严重程度的关键指标。我们收集了60起泥石流事件的数据,建立了一个全面的数据集。为了提高模型的鲁棒性和准确性,我们利用多重共线性分析和赤池信息准则(Akaike information criterion, AIC)对变量选择进行了优化。开发了四种统计模型,包括多元线性、对数、幂和指数回归。我们还构建了基于机器学习算法的模型,包括随机森林和梯度增强决策树,并使用5倍交叉验证对它们进行了测试。在确认训练数据集满足线性统计假设后,我们建立了稳健的回归模型。我们使用f检验和t检验检验回归方程和系数的显著性。通过贝叶斯方法对机器学习算法的超参数进行优化。使用R2、调整后的R2、平均绝对误差(MAE)、平均相对误差(MRE)和均方根误差(RMSE)等指标评估模型的性能。结果表明,影响RBI最重要的因素是集水区面积(A)和泥石流与干流流量比(Q),其中对数模型和幂模型以其简单、高效的特点表现最好。随机森林模型总体上显示出最高的预测精度和稳定性。将统计方法与机器学习相结合,提高预测精度,为防灾策略提供实用指导。该方法克服了数值模拟和实验研究的局限性,为RBI预测提供了一种更加灵活高效的方法。未来的工作将把这些发现扩展到其他地质环境,以进一步提高模型的适应性和性能。
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
Catena
Catena 环境科学-地球科学综合
CiteScore
10.50
自引率
9.70%
发文量
816
审稿时长
54 days
期刊介绍: Catena publishes papers describing original field and laboratory investigations and reviews on geoecology and landscape evolution with emphasis on interdisciplinary aspects of soil science, hydrology and geomorphology. It aims to disseminate new knowledge and foster better understanding of the physical environment, of evolutionary sequences that have resulted in past and current landscapes, and of the natural processes that are likely to determine the fate of our terrestrial environment. Papers within any one of the above topics are welcome provided they are of sufficiently wide interest and relevance.
×
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