Evaluation of Three Algorithms and Forest Fire Risk Prediction in Zhejiang Province of China DOI Open Access

Richard Bian,

Keji Chen,

Guoqiang Li

et al.

Forests, Journal Year: 2024, Volume and Issue: 15(12), P. 2146 - 2146

Published: Dec. 5, 2024

Forest fires represent a paramount natural disaster of global concern. Zhejiang Province has the highest forest coverage rate in China, and are one main disasters impacting management region. In this study, we comprehensively analyzed spatiotemporal distribution based on MODIS data from 2013 to 2023. The results showed that annual incidence shown an overall downward trend 2023, with occurring more frequently winter spring. By utilizing eight contributing factors fire occurrence as variables, three models were constructed: Logistic Regression (LR), Random (RF), eXtreme Gradient Boosting (XGBoost). RF XGBoost demonstrated high predictive ability, achieving accuracy rates 0.85 0.92, f1-score 0.84 AUC values 0.892 0.919, respectively. Further analysis using revealed elevation precipitation had most significant effects fires. Additionally, predictions risk generated by indicated is southern part Province, particularly Wenzhou Lishui areas, well southwest Hangzhou area north Quzhou area. future, can be predicted site models, providing scientific reference for aiding prevention mitigation impacts

Language: Английский

Assessing the Potential of UAV for Large-Scale Fractional Vegetation Cover Mapping with Satellite Data and Machine Learning DOI Creative Commons
Xunlong Chen, Yiming Sun, Xinyue Qin

et al.

Remote Sensing, Journal Year: 2024, Volume and Issue: 16(19), P. 3587 - 3587

Published: Sept. 26, 2024

Fractional vegetation cover (FVC) is an essential metric forvaluating ecosystem health and soil erosion. Traditional ground-measuring methods are inadequate for large-scale FVC monitoring, while remote sensing-based estimation approaches face issues such as spatial scale discrepancies between ground truth data image pixels, well limited sample representativeness. This study proposes a method integrating uncrewed aerial vehicle (UAV) satellite imagery using machine learning (ML) models. First, we assess the extraction performance of three classification (OBIA-RF, threshold, K-means) under UAV imagery. The optimal then selected binary aggregated to generate high-accuracy reference matching resolutions different images. Subsequently, construct models four ML algorithms (KNN, MLP, RF, XGBoost) utilize SHapley Additive exPlanation (SHAP) impact spectral features indices (VIs) on model predictions. Finally, best used map in region. Our results indicate that OBIA-RF effectively extract information from images, achieving average precision recall 0.906 0.929, respectively. generates data. With improvement resolution variability decreases continuity increases. RF outperforms others at 10 m 20 resolutions, with R2 values 0.827 Conversely, XGBoost achieves highest accuracy 30 resolution, 0.847. also found was significantly related number VIs (including red edge near-infrared bands), this correlation enhanced coarser proposed addresses shortcomings conventional methods, improves monitoring erosion areas, serves ecological environment technology.

Language: Английский

Citations

3
Estimation of unrealized forest carbon potential in China using time-varying Boruta-SHAP-random forest model and climate vegetation productivity index DOI
Tao Li, Yi Wu,

Fang Ren

et al.

Journal of Environmental Management, Journal Year: 2025, Volume and Issue: 377, P. 124649 - 124649

Published: Feb. 22, 2025

Language: Английский

Citations

0
Machine Learning-Based Wildfire Susceptibility Mapping: A Gis-Integrated Predictive Framework DOI

Yehya Bouzeraa,

Nardjes Bouchemal,

Salim Djaaboub

et al.

Published: Jan. 1, 2025

Language: Английский

Citations

0
Predicting forest above-ground biomass using SAR imagery and GEDI data through machine learning in GEE cloud DOI Creative Commons
Chiranjit Singha, Kishore Chandra Swain, Satiprasad Sahoo

et al.

Forest Science and Technology, Journal Year: 2025, Volume and Issue: unknown, P. 1 - 20

Published: April 7, 2025

Language: Английский

Citations

0
Modeling climate change impacts and predicting future vulnerability in the Mount Kenya forest ecosystem using remote sensing and machine learning DOI Creative Commons

Terry Amolo Otieno,

L. H. Otieno,

Brian Rotich

et al.

Environmental Monitoring and Assessment, Journal Year: 2025, Volume and Issue: 197(6)

Published: May 6, 2025

Abstract The Mount Kenya forest ecosystem (MKFE), a crucial biodiversity hotspot and one of Kenya’s key water towers, is increasingly threatened by climate change, putting its ecological integrity vital services at risk. Understanding the interactions between extremes dynamics essential for conservation planning, especially in Forest Ecosystem where rising temperatures erratic rainfall are altering vegetation patterns, reducing resilience, threatening both security. This study integrates remote sensing machine learning to assess historical changes predict areas risk future. Landsat imagery from 2000 2020 was used derive indices comprising Normalized Difference Vegetation Index (NDVI), Enhanced (EVI), Soil-Adjusted (SAVI), Bare Soil (BSI). Climate variables, including extreme precipitation temperature indices, were extracted CHIRPS ERA5 datasets. Machine models, Random (RF), XGBoost, Support Vector Machines (SVM), trained climate-vegetation relationships future under SSP245 scenario using Coupled Model Intercomparison Project Phase 6 (CMIP6) downscaled projections. RF model achieved high accuracy ( R 2 = 0.82, RMSE 0.15) predicting conditions. projections show 49–55% decline EVI across 2040, with most pronounced losses likely lower montane zones, which more sensitive climate-induced stress. Results emphasize critical role sustaining health highlight urgent need adaptive management strategies, afforestation, sustainable land-use policy-driven efforts. provides scalable framework modelling impacts on ecosystems globally offers actionable insights policymakers.

Language: Английский

Citations

0
Multi-Hazard Susceptibility Mapping Using Machine Learning Approaches: A Case Study of South Korea DOI Creative Commons
Changju Kim,

Soonchan Park,

Heechan Han

et al.

Remote Sensing, Journal Year: 2025, Volume and Issue: 17(10), P. 1660 - 1660

Published: May 8, 2025

The frequency and magnitude of natural hazards have been steadily increasing, largely due to extreme weather events driven by climate change. These pose significant global challenges, underscoring the need for accurate prediction models systematic preparedness. This study aimed predict multiple in South Korea using various machine learning algorithms. area, (100,210 km2), was divided into a grid system with 0.01° resolution. Meteorological, climatic, topographical, remotely sensed data were interpolated each cell analysis. focused on three major hazards: drought, flood, wildfire. Predictive developed two algorithms: Random Forest (RF) Extreme Gradient Boosting (XGB). analysis showed that XGB performed exceptionally well predicting droughts floods, achieving ROC scores 0.9998 0.9999, respectively. For wildfire prediction, RF achieved high score 0.9583. results integrated generate multi-hazard susceptibility map. provides foundational development hazard management response strategies context Furthermore, it offers basis future research exploring interaction effects multi-hazards.

Language: Английский

Citations

0
Occurrence, probable causes, and management of forest wildfires in the Northern Highlands of Pakistan DOI Creative Commons
Mohammad Nafees, Wajid Rashid, Hameeda Sultan

et al.

Environmental Challenges, Journal Year: 2024, Volume and Issue: 15, P. 100930 - 100930

Published: April 1, 2024

Pakistan is forest-deficient and cannot afford forest losses associated with large-scale wildfire destruction. The Northern Mountainous Range (NMR) in Khyber Pakhtunkhwa province has substantial cover, yet complex human-environment interactions have created vast fire-prone areas. This study examines the historical environmental drivers of wildfires NMR prospective management strategies to reduce their devastating impacts. We reviewed news articles surveyed local fire offices obtain records occurrences develop a spatial model. used maximum entropy (MaxEnt) model evaluate probability based on climatic influences found district Swat neighboring areas region be hotspot due mainly factors precipitation. Furthermore, an analysis was conducted regional governance, encompassing legislative that impede safety deficient budgets. In addition, we detailed recognized causes fires area, revealing significant human contribution. There rising consensus governance regionally, locally, within communities must comprehensively handle compounding complicated concerns surrounding wildfires. extensive seeks impact support protection through actions update policy, encourage participatory planning at level, prepare for future by allocating enough budget emergency disasters, raise public awareness, educate about dangers.

Language: Английский

Citations

2
Integrating ensemble machine learning and explainable AI for enhanced forest fire susceptibility analysis and risk assessment in Türkiye’s Mediterranean region DOI
Hasan Tonbul

Earth Science Informatics, Journal Year: 2024, Volume and Issue: unknown

Published: Sept. 6, 2024

Language: Английский

Citations

2
Dynamic optimization can effectively improve the accuracy of reference evapotranspiration in southern China DOI
Xiang Xiao, Ziniu Xiao, Xiaogang Liu

et al.

Computers and Electronics in Agriculture, Journal Year: 2024, Volume and Issue: 230, P. 109881 - 109881

Published: Dec. 31, 2024

Language: Английский

Citations

2
Integrating Long term Satellite Data and Machine Learning to Spatiotemporal Fire Analysis in Hour al Azim International Wetland DOI

Seyed Fazel Hashemi,

Hossein Mohammad Asgari

Water Air & Soil Pollution, Journal Year: 2024, Volume and Issue: 235(7)

Published: June 13, 2024

Language: Английский

Citations

1