Analysing Fire Propagation Models: A Case Study on FARSITE for Prolonged Wildfires

Fire, Journal Year: 2025, Volume and Issue: 8(5), P. 166 - 166

Published: April 23, 2025

With increasing wildfire severity and duration driven by climate change, accurately predicting fire behavior over extended time frames is critical for effective management mitigation of such wildfires. Fire propagation models play a pivotal role in these efforts, providing simulations that can be used to strategize respond active fires. This study examines the area simulator (FARSITE) model’s performance simulating recent events persisted 24 h with limited firefighting intervention mostly remote access areas across diverse ecosystems. Our findings reveal key insights into prolonged scenarios potentially informing improvements operational long-term predictive accuracy, as comparisons indexes showed reasonable results between detected fires from information resource systems (FIRMSs) first following days. A case Madeira Island highlights integration real-time weather predictions post-event data analysis. analysis underscores potential combining accurate forecasts retrospective validation improve capabilities dynamic environments, which guided development software platform designed analyse ongoing real-time, leveraging image satellite predictions.

Language: Английский

---

Wildfire response of forest species from multispectral LiDAR data. A deep learning approach with synthetic data

Ecological Informatics, Journal Year: 2024, Volume and Issue: 81, P. 102612 - 102612

Published: April 22, 2024

Forests play a crucial role as the lungs and life-support system of our planet, harbouring 80% Earth's biodiversity. However, we are witnessing an average loss 480 ha forest every hour because destructive wildfires spreading across globe. To effectively mitigate threat wildfires, it is to devise precise dependable approaches for forecasting fire dynamics formulating efficient management strategies, such utilisation fuel models. The objective this study was enhance classification that considers only structural information, Prometheus model, by integrating data on responses various tree species other vegetation elements, ground litter shrubs. This distinction can be achieved using multispectral (MS) Light Detection Ranging (LiDAR) in mixed forests. methodology involves novel approach semantic classifications forests generating synthetic with labels regarding reflectance information at different spectral bands, real MS scanner device would detect. Forests, which highly intricate environments, present challenges accurately classifying point clouds. address complexity, deep learning (DL) model trained clouds formats achieve best performance when leveraging data. Forest plots region were scanned Terrestrial Laser Scanning sensors wavelengths 905 1550 nm. Subsequently, interpolation process applied generate each plot, DL classify them. These surpassed thresholds 90% 75% accuracy intersection over union, respectively, resulting more categorisation models based distinct elements fire. results reveal potential LiDAR improving retrieval ecosystems enhancing wildfire efforts.

Language: Английский

---

A Deep Learning Framework: Predicting Fire Radiative Power From the Combination of Polar-Orbiting and Geostationary Satellite Data During Wildfire Spread

IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, Journal Year: 2024, Volume and Issue: 17, P. 10827 - 10841

Published: Jan. 1, 2024

Fire Radiative Power (FRP) is a key indicator for evaluating the intensity of wildfires, unlike traditional real-time fire lines or combustion areas that only provide binary information, and its accurate prediction more important firefighting actions environmental pollution assessment. To this end, we used combination data from geostationary satellites polar orbit to correct FRP data. Incorporating various factors affect wildfire spread, such as meteorological conditions, topography, vegetation indexes, population density, constructed comprehensive California spread dataset, covering information since 2017. Then, established deep learning framework integrates modules analyze multimodal imagery. We investigated impact input sequence length loss function design on model predictive performance, leading subsequent optimization. Furthermore, our has demonstrated acceptable performance in transfer multi-step prediction, emphasizing application value management. It can detailed about showcasing powerful capability process potential emerging field prediction.

Language: Английский

---

PRISMethaNet: A novel deep learning model for landfill methane detection using PRISMA satellite data

ISPRS Journal of Photogrammetry and Remote Sensing, Journal Year: 2024, Volume and Issue: 218, P. 802 - 818

Published: Oct. 20, 2024

---

Trajectory-based fish event classification through pre-training with diffusion models

Ecological Informatics, Journal Year: 2024, Volume and Issue: 82, P. 102733 - 102733

Published: July 28, 2024

This study contributes to advancing the field of automatic fish event recognition in natural underwater videos, addressing current gap studying interaction and competition, including predator-prey relationships mating behaviors. We used corkwing wrasse (Symphodus melops) as a model, marine species commercial importance that reproduces sea-weed nests built cared for by single male. These attract wide range visitors are focal point behavior such spawning, chasing, maintenance. propose deep learning methodology analyze movement trajectories nesting male classify associated events observed their habitat. Our approach leverages unsupervised pre-training based on diffusion models, leading improved feature learning. Additionally, we introduce dataset comprising 16,937 across 12 classes, making it largest terms class diversity. results demonstrate superior performance our method compared several architectures. The code proposed can be found at https://github.com/NoeCanovi/Fish_Behaviors_Generative_Models.

Language: Английский

---

Capturing and interpreting wildfire spread dynamics: attention-based spatiotemporal models using ConvLSTM networks

Ecological Informatics, Journal Year: 2024, Volume and Issue: 82, P. 102760 - 102760

Published: Aug. 7, 2024

Predicting the trajectory of geographical events, such as wildfire spread, presents a formidable task due to dynamic associations among influential biophysical factors. Geo-events like wildfires frequently display short and long-range spatial temporal correlations. Short-range effects are direct contact near-contact spread fire front. Long-range represented by processes spotting, where firebrands carried wind ignite fires distant from flaming front, altering shape speed an advancing This study addresses these modeling challenges clearly defining analyzing scale-dependent spatiotemporal dynamics that influence focusing on interplay between factors behavior. We propose two unique attention-based models using Convolutional Long Short-Term Memory (ConvLSTM) networks. These designed learn capture range local global The proposed were tested datasets: high-resolution dataset produced with semi-empirical percolation model satellite observed data in California 2012–2021. Results indicate accurately predict front movements consistent known spread-biophysical dynamics. Our research suggests there is considerable potential for attention mechanisms behavior transferability, can guide rapid deployment management operations. also highlight directions future studies focus how self-attention mechanism could enhance performance geospatial applications.

Language: Английский

---

A Novel Spatio-Temporal Vision Transformer Model for Improving Wetland Mapping Using Multi-Seasonal Sentinel Data

Remote Sensing Applications Society and Environment, Journal Year: 2024, Volume and Issue: 37, P. 101401 - 101401

Published: Nov. 17, 2024

Language: Английский

---

Wildfire Spread Prediction in North America Using Satellite Imagery and Vision Transformer

Published: June 25, 2024

Language: Английский

---

The compound effect of topography, weather, and fuel type on the spread and severity of the largest wildfire in NW of Turkey

Natural Hazards, Journal Year: 2024, Volume and Issue: unknown

Published: Sept. 25, 2024

Language: Английский

---

FusionFireNet: A CNN-LSTM Model for Short-Term Wildfire Hotspot Prediction Utilizing Spatio-Temporal Datasets

Remote Sensing Applications Society and Environment, Journal Year: 2024, Volume and Issue: 37, P. 101436 - 101436

Published: Dec. 16, 2024

Language: Английский

---