Machine Learning and Deep Learning for Wildfire Spread Prediction: A Review DOI Creative Commons

Henintsoa S. Andrianarivony,

Moulay A. Akhloufi

Fire, Journal Year: 2024, Volume and Issue: 7(12), P. 482 - 482

Published: Dec. 18, 2024

The increasing frequency and intensity of wildfires highlight the need to develop more efficient tools for firefighting management, particularly in field wildfire spread prediction. Classical models have relied on mathematical empirical approaches, which trouble capturing complexity fire dynamics suffer from poor flexibility static assumptions. emergence machine learning (ML) and, specifically, deep (DL) has introduced new techniques that significantly enhance prediction accuracy. ML models, such as support vector machines ensemble use tabular data points identify patterns predict behavior. However, these often struggle with dynamic nature wildfires. In contrast, DL convolutional neural networks (CNNs) recurrent (CRNs), excel at handling spatiotemporal complexities data. CNNs are effective analyzing spatial satellite imagery, while CRNs suited both sequential data, making them highly performant predicting This paper presents a systematic review recent developed prediction, detailing commonly used datasets, improvements achieved, limitations current methods. It also outlines future research directions address challenges, emphasizing potential play an important role management mitigation strategies.

Language: Английский

Assessing Pan-Canada wildfire susceptibility by integrating satellite data with novel hybrid deep learning and black widow optimizer algorithms DOI Creative Commons
Khabat Khosravi,

Ashkan Mosallanejad,

Sayed M. Bateni

et al.

The Science of The Total Environment, Journal Year: 2025, Volume and Issue: 977, P. 179369 - 179369

Published: April 15, 2025

In light of the rising frequency severe wildfires and their widespread socio-ecological impacts, it is essential to develop cost-effective reliable methods for accurately predicting mapping wildfire occurrences. This study aimed several novel deep-learning models determine probability occurrence on a national scale in Canada by integrating remote sensing data, deep learning, metaheuristic algorithms. present study, standalone long short-term memory (LSTM), recurrent neural network (RNN), bidirectional LSTM (BiLSTM), RNN (BiRNN) were developed, these hybridized with black widow optimizer (BWO). To train test models, 4240 historical (2014-2023) large locations collected across Canada. Fourteen wildfire-related predictors used map susceptibility, Gini coefficient determining each predictor's importance occurrence. Finally, developed evaluated tested using area under receiver operating characteristic curve (AUC), other statistical error metrics. During testing stage, hybrid BiLSTM-BWO model outperformed (AUC = 0.9686), followed RNN-BWO 0.9683), LSTM-BWO 0.9672), BiRNN-BWO 0.9643), BiLSTM 0.9420), 0.9367), BiRNN 0.9247) 0.8737). Based model, 19.7 %, 42.6 13.4 14.5 9.8 % was classified as having very low, moderate, high, high susceptibility future wildfires, respectively. Saskatchewan, Manitoba, British Columbia Alberta among provinces areas while Prince Edward Island Newfoundland Labrador from Atlantic had lowest According coefficient, windspeed, land use cover, precipitation, specific humidity maximum temperature strongest impact highlights effectiveness prediction potential improve management, prevention, mitigation strategies Canada's future.

Language: Английский

Citations

0

Machine Learning and Deep Learning for Wildfire Spread Prediction: A Review DOI Creative Commons

Henintsoa S. Andrianarivony,

Moulay A. Akhloufi

Fire, Journal Year: 2024, Volume and Issue: 7(12), P. 482 - 482

Published: Dec. 18, 2024

The increasing frequency and intensity of wildfires highlight the need to develop more efficient tools for firefighting management, particularly in field wildfire spread prediction. Classical models have relied on mathematical empirical approaches, which trouble capturing complexity fire dynamics suffer from poor flexibility static assumptions. emergence machine learning (ML) and, specifically, deep (DL) has introduced new techniques that significantly enhance prediction accuracy. ML models, such as support vector machines ensemble use tabular data points identify patterns predict behavior. However, these often struggle with dynamic nature wildfires. In contrast, DL convolutional neural networks (CNNs) recurrent (CRNs), excel at handling spatiotemporal complexities data. CNNs are effective analyzing spatial satellite imagery, while CRNs suited both sequential data, making them highly performant predicting This paper presents a systematic review recent developed prediction, detailing commonly used datasets, improvements achieved, limitations current methods. It also outlines future research directions address challenges, emphasizing potential play an important role management mitigation strategies.

Language: Английский

Citations

1