Decoding Agricultural Drought Resilience: A Triple-Validated Random Forest Framework Integrating Multi-Source Remote Sensing for High-Resolution Monitoring in the North China Plain

Agricultural drought poses a severe threat to food security in the North China Plain, necessitating accurate and timely monitoring approaches. This study presents a novel drought assessment framework that innovatively integrates multiple remote sensing indices through an optimized random forest algo...

Full description

Saved in:
Bibliographic Details
Main Authors: Xianyong Meng, Song Zhang, Guoqing Wang, Jianli Ding, Chengbin Chu, Jianyun Zhang, Hao Wang
Format: Article
Language:English
Published: MDPI AG 2025-04-01
Series:Remote Sensing
Subjects:
Online Access:https://www.mdpi.com/2072-4292/17/8/1404
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850144284077981696
author Xianyong Meng
Song Zhang
Guoqing Wang
Jianli Ding
Chengbin Chu
Jianyun Zhang
Hao Wang
author_facet Xianyong Meng
Song Zhang
Guoqing Wang
Jianli Ding
Chengbin Chu
Jianyun Zhang
Hao Wang
author_sort Xianyong Meng
collection DOAJ
description Agricultural drought poses a severe threat to food security in the North China Plain, necessitating accurate and timely monitoring approaches. This study presents a novel drought assessment framework that innovatively integrates multiple remote sensing indices through an optimized random forest algorithm, achieving unprecedented accuracy in regional drought monitoring. The framework introduces three key innovations: (1) a systematic integration of six drought-related factors including vegetation condition index (VCI), temperature condition index (TCI), precipitation condition index (PCI), land cover type (LC), aspect (ASPECT), and available water capacity (AWC); (2) an optimized random forest algorithm configuration with 100 decision trees and enhanced feature extraction capability; and (3) a robust triple-validation strategy combining standardized precipitation evapotranspiration index (SPEI), comprehensive meteorological drought index (CI), and soil moisture verification. The framework demonstrates exceptional performance with R<sup>2</sup> values consistently above 0.80 for monthly assessments, reaching 0.86 during autumn and 0.73 during summer seasons. Particularly, it achieves 87% accuracy in mild drought (−1.0 < SPEI ≤ −0.5) and 85% in moderate drought (−1.5 < SPEI ≤ −1.0) detection. The 20-year (2000–2019) spatiotemporal analysis reveals that moderate drought events dominated the region (23.7% of total occurrences), with significant intensification during the 2010–2012 and 2014–2016 periods. Summer drought frequency peaked at 12–15 months in south-central Shandong (37°N, 117°E) and eastern Henan (34°N, 114°E). The framework’s high spatial resolution (1 km) and comprehensive validation protocol establish a reliable foundation for agricultural drought monitoring and water resource management, offering a transferable methodology for regional drought assessment worldwide.
format Article
id doaj-art-5de6e898139044eca55c31dd2bca0e20
institution OA Journals
issn 2072-4292
language English
publishDate 2025-04-01
publisher MDPI AG
record_format Article
series Remote Sensing
spelling doaj-art-5de6e898139044eca55c31dd2bca0e202025-08-20T02:28:25ZengMDPI AGRemote Sensing2072-42922025-04-01178140410.3390/rs17081404Decoding Agricultural Drought Resilience: A Triple-Validated Random Forest Framework Integrating Multi-Source Remote Sensing for High-Resolution Monitoring in the North China PlainXianyong Meng0Song Zhang1Guoqing Wang2Jianli Ding3Chengbin Chu4Jianyun Zhang5Hao Wang6School of Smart Water Conservancy Engineering, Xinjiang Institute of Technology, Aksu 843000, ChinaCollege of Resources and Environmental Sciences, China Agricultural University, Beijing 100080, ChinaResearch Center for Climate Change, Ministry of Water Resources, Nanjing 210029, ChinaSchool of Smart Water Conservancy Engineering, Xinjiang Institute of Technology, Aksu 843000, ChinaESIEE Paris, Gustave Eiffel University, F-77454 Marne-la-Vallée, FranceResearch Center for Climate Change, Ministry of Water Resources, Nanjing 210029, ChinaChinese Academy of Engineering, No. 2 Bingjiaokou Hutong, Beijing 100088, ChinaAgricultural drought poses a severe threat to food security in the North China Plain, necessitating accurate and timely monitoring approaches. This study presents a novel drought assessment framework that innovatively integrates multiple remote sensing indices through an optimized random forest algorithm, achieving unprecedented accuracy in regional drought monitoring. The framework introduces three key innovations: (1) a systematic integration of six drought-related factors including vegetation condition index (VCI), temperature condition index (TCI), precipitation condition index (PCI), land cover type (LC), aspect (ASPECT), and available water capacity (AWC); (2) an optimized random forest algorithm configuration with 100 decision trees and enhanced feature extraction capability; and (3) a robust triple-validation strategy combining standardized precipitation evapotranspiration index (SPEI), comprehensive meteorological drought index (CI), and soil moisture verification. The framework demonstrates exceptional performance with R<sup>2</sup> values consistently above 0.80 for monthly assessments, reaching 0.86 during autumn and 0.73 during summer seasons. Particularly, it achieves 87% accuracy in mild drought (−1.0 < SPEI ≤ −0.5) and 85% in moderate drought (−1.5 < SPEI ≤ −1.0) detection. The 20-year (2000–2019) spatiotemporal analysis reveals that moderate drought events dominated the region (23.7% of total occurrences), with significant intensification during the 2010–2012 and 2014–2016 periods. Summer drought frequency peaked at 12–15 months in south-central Shandong (37°N, 117°E) and eastern Henan (34°N, 114°E). The framework’s high spatial resolution (1 km) and comprehensive validation protocol establish a reliable foundation for agricultural drought monitoring and water resource management, offering a transferable methodology for regional drought assessment worldwide.https://www.mdpi.com/2072-4292/17/8/1404triple-validation strategymulti-source remote sensingremote sensing fusionagricultural water managementclimate change adaptationrandom forest optimization
spellingShingle Xianyong Meng
Song Zhang
Guoqing Wang
Jianli Ding
Chengbin Chu
Jianyun Zhang
Hao Wang
Decoding Agricultural Drought Resilience: A Triple-Validated Random Forest Framework Integrating Multi-Source Remote Sensing for High-Resolution Monitoring in the North China Plain
Remote Sensing
triple-validation strategy
multi-source remote sensing
remote sensing fusion
agricultural water management
climate change adaptation
random forest optimization
title Decoding Agricultural Drought Resilience: A Triple-Validated Random Forest Framework Integrating Multi-Source Remote Sensing for High-Resolution Monitoring in the North China Plain
title_full Decoding Agricultural Drought Resilience: A Triple-Validated Random Forest Framework Integrating Multi-Source Remote Sensing for High-Resolution Monitoring in the North China Plain
title_fullStr Decoding Agricultural Drought Resilience: A Triple-Validated Random Forest Framework Integrating Multi-Source Remote Sensing for High-Resolution Monitoring in the North China Plain
title_full_unstemmed Decoding Agricultural Drought Resilience: A Triple-Validated Random Forest Framework Integrating Multi-Source Remote Sensing for High-Resolution Monitoring in the North China Plain
title_short Decoding Agricultural Drought Resilience: A Triple-Validated Random Forest Framework Integrating Multi-Source Remote Sensing for High-Resolution Monitoring in the North China Plain
title_sort decoding agricultural drought resilience a triple validated random forest framework integrating multi source remote sensing for high resolution monitoring in the north china plain
topic triple-validation strategy
multi-source remote sensing
remote sensing fusion
agricultural water management
climate change adaptation
random forest optimization
url https://www.mdpi.com/2072-4292/17/8/1404
work_keys_str_mv AT xianyongmeng decodingagriculturaldroughtresilienceatriplevalidatedrandomforestframeworkintegratingmultisourceremotesensingforhighresolutionmonitoringinthenorthchinaplain
AT songzhang decodingagriculturaldroughtresilienceatriplevalidatedrandomforestframeworkintegratingmultisourceremotesensingforhighresolutionmonitoringinthenorthchinaplain
AT guoqingwang decodingagriculturaldroughtresilienceatriplevalidatedrandomforestframeworkintegratingmultisourceremotesensingforhighresolutionmonitoringinthenorthchinaplain
AT jianliding decodingagriculturaldroughtresilienceatriplevalidatedrandomforestframeworkintegratingmultisourceremotesensingforhighresolutionmonitoringinthenorthchinaplain
AT chengbinchu decodingagriculturaldroughtresilienceatriplevalidatedrandomforestframeworkintegratingmultisourceremotesensingforhighresolutionmonitoringinthenorthchinaplain
AT jianyunzhang decodingagriculturaldroughtresilienceatriplevalidatedrandomforestframeworkintegratingmultisourceremotesensingforhighresolutionmonitoringinthenorthchinaplain
AT haowang decodingagriculturaldroughtresilienceatriplevalidatedrandomforestframeworkintegratingmultisourceremotesensingforhighresolutionmonitoringinthenorthchinaplain