Integrating Physical-Based Models and Structure-from-Motion Photogrammetry to Retrieve Fire Severity by Ecosystem Strata from Very High Resolution UAV Imagery DOI Creative Commons
José Manuel Fernández‐Guisuraga, Leonor Calvo,

Luis A. Pérez-Rodríguez

et al.

Fire, Journal Year: 2024, Volume and Issue: 7(9), P. 304 - 304

Published: Aug. 27, 2024

We propose a novel mono-temporal framework with physical basis and ecological consistency to retrieve fire severity at very high spatial resolution. First, we sampled the Composite Burn Index (CBI) in 108 field plots that were subsequently surveyed through unmanned aerial vehicle (UAV) flights. Then, mimicked methodology for CBI assessment remote sensing framework. strata identified individual tree segmentation geographic object-based image analysis (GEOBIA). In each stratum, wildfire effects estimated following methods: (i) vertical structural complexity of vegetation legacies was computed from 3D-point clouds, as proxy biomass consumption; (ii) biophysical variables retrieved multispectral data by inversion PROSAIL radiative transfer model, direct link remaining after canopy scorch torch. The scores predicted UAV ecologically related metrics level featured fit respect field-measured (R2 > 0.81 RMSE < 0.26). Conversely, conventional retrieval using battery spectral predictors (point height distribution indices) plot provided much worse performance = 0.677 0.349).

Language: Английский

Impact of fire severity on forest structure and biomass stocks using NASA GEDI data. Insights from the 2020 and 2021 wildfire season in Spain and Portugal DOI Creative Commons
Juan Guerra-Hernández, José M. C. Pereira, Atticus Stovall

et al.

Science of Remote Sensing, Journal Year: 2024, Volume and Issue: 9, P. 100134 - 100134

Published: May 16, 2024

Wildfires have been progressively shrinking the C sink capacity of forest accelerating climate change effects on biodiversity, especially where megafires are recurrent and increased frequency such as in Mediterranean. Data from The Global Ecosystem Dynamics Investigation (GEDI) mission can inform changes structure to fire impacts vegetation. In this study, we assessed performance GEDI at measuring biomass structural wildfires using 2020/21 summer seasons Spain Portugal. hybrid-inference method was used calculate mean total pre- post-fire stages, while footprint data further explain severity classes derived optical data. Our results showed increasing impact stocks ecological metrics by severity. More than over stocks, severe fires substantially altered trends plant area volume density. integration had an accuracy 52% considering five 69% when three main classes: unburned, moderate high. Structural be improve optical-based estimates globally evaluate potential based time-series tracks showcased but also measure recovery between seasons. extension is a major support for wildfire mapping efforts, integrated approaches capture biodiversity monitoring carbon stocks.

Language: Английский

Citations

5
Optimising fire severity mapping using pixel-based image compositing DOI Creative Commons
Néstor Quintero, Olga Viedma, Sander Veraverbeke

et al.

Remote Sensing of Environment, Journal Year: 2025, Volume and Issue: 321, P. 114687 - 114687

Published: March 6, 2025

Language: Английский

Citations

0
Biophysical drivers of Pinus nigra subsp. salzmannii post-fire regeneration: role of fire refugia DOI Creative Commons
Ana Lucía Méndez‐Cartín, Lluís Coll, Meg A. Krawchuk

et al.

Landscape Ecology, Journal Year: 2025, Volume and Issue: 40(4)

Published: April 9, 2025

Language: Английский

Citations

0
Improved fire severity prediction using pre-fire remote sensing and meteorological time series: Application to the French Mediterranean area DOI Creative Commons
Victor Pénot, Thomas Opitz, François Pimont

et al.

Agricultural and Forest Meteorology, Journal Year: 2025, Volume and Issue: 371, P. 110588 - 110588

Published: May 9, 2025

Language: Английский

Citations

0
Are fire intensity and burn severity associated? Advancing our understanding of FRP and NBR metrics from Himawari-8/9 and Sentinel-2 DOI Creative Commons
Konstantinos Chatzopoulos-Vouzoglanis, Karin Reinke, Mariela Soto‐Berelov

et al.

International Journal of Applied Earth Observation and Geoinformation, Journal Year: 2024, Volume and Issue: 127, P. 103673 - 103673

Published: Jan. 23, 2024

Burn severity has been widely studied. Typical approaches use spectral differencing indices from remotely sensed data to extrapolate in-situ assessments. Next generation geostationary offer near-continuous fire behaviour information, which used for detection and monitoring but remains underutilized impact estimation. Here, we explore the association between intensity metrics understand whether where they describe similar wildfire effects. The commonly Differenced Normalised Ratio (dNBR) index was calculated Advanced Himawari Imager (AHI − 2 km) Sentinel-2 (20 m) compared different Fire Radiative Power (FRP) derived hotspot detections AHI across Australia. comparison implemented through stratifications based on biogeographical region, land cover, type, percentage of pixel burned (fire fractional cover). results indicate that FRP dNBR do not correlate in most scenarios, noting correlations being marginally stronger hotter fires. However, become significantly when are grouped using type information with peaking (R = 0.75) large fires 41–60 % an pixel. In conclusion, proxies capture aspects impact, only each other after auxiliary data. Spectral have extensively during past decades, however high-frequency estimations potential augment existing reveal new ways characterizing over areas.

Language: Английский

Citations

2
Linking crown fire likelihood with post-fire spectral variability in Mediterranean fire-prone ecosystems DOI Creative Commons
José Manuel Fernández‐Guisuraga, Leonor Calvo, Carmen Quintano

et al.

International Journal of Wildland Fire, Journal Year: 2024, Volume and Issue: 33(4)

Published: April 11, 2024

Background Fire behaviour assessments of past wildfire events have major implications for anticipating post-fire ecosystem responses and fuel treatments to mitigate extreme fire subsequent wildfires. Aims This study evaluates the first time potential remote sensing techniques provide explicit estimates type (surface fire, intermittent crown continuous fire) in Mediterranean ecosystems. Methods Random Forest classification was used assess capability spectral indices multiple endmember mixture analysis (MESMA) image fractions (char, photosynthetic vegetation, non-photosynthetic vegetation) retrieved from Sentinel-2 data predict across four large wildfires Key results MESMA fraction images procured more accurate broadleaf conifer forests than indices, without remarkable confusion among types. High likelihood linked a char fractional cover about 0.8, providing direct physical interpretation. Conclusions Intrinsic biophysical characteristics such as sub-pixel with basis are given Implications may be leveraged by land managers determine areas, but further validation field is advised.

Language: Английский

Citations

1
Digital soil mapping of soil burn severity DOI
Stewart G. Wilson,

Samuel E. Prentice

Soil Science Society of America Journal, Journal Year: 2024, Volume and Issue: 88(4), P. 1045 - 1067

Published: June 14, 2024

Abstract Fire alters soil hydrologic properties leading to increased risk of catastrophic debris flows and post‐fire flooding. As a result, US federal agencies map burn severity (SBS) via direct observation adjustment rasters burned area reflectance. We developed unique application digital mapping (DSM) SBS in the Creek which 154,000 ha Sierra Nevada. utilized 169 ground‐based observations combination with raster proxies forming factors, pre‐fire fuel conditions, fire effects vegetation build model (DSMSBS) using random forest algorithm compared DSMSBS established map. The had cross‐validation accuracy 48%. technique 46% agreement between field pixels. However, since is manual, it could not be cross‐validation. produced class uncertainty maps, showed high prediction probabilities around observations, low away from observations. aid assessment teams sample prioritization. report 107 km 2 more classified as moderate technique. conclude that blending factors based can improve mapping. This represents shift validating remotely sensed reflectance imagery toward quantitative landscape model, incorporates both soils information directly predict SBS.

Language: Английский

Citations

1
Comparative Study of Random Forest and Support Vector Machine for Land Cover Classification and Post-Wildfire Change Detection DOI Creative Commons

Yong How Jonathan Tan,

Lia Duarte, Ana Cláudia Teodoro

et al.

Land, Journal Year: 2024, Volume and Issue: 13(11), P. 1878 - 1878

Published: Nov. 10, 2024

The land use cover (LULC) map is extensively employed for different purposes. Machine learning (ML) algorithms applied in remote sensing (RS) data have been proven effective image classification, object detection, and semantic segmentation. Previous studies shown that random forest (RF) support vector machine (SVM) consistently achieve high accuracy classification. Considering the important role of Portugal’s Serra da Estrela Natural Park (PNSE) biodiversity nature conversation at an international scale, availability timely on PNSE emergency evaluation periodic assessment crucial. In this study, application RF SVM classifiers, object-based (OBIA) pixel-based (PBIA) approaches, with Sentinel-2A imagery was evaluated using Google Earth Engine (GEE) platform classification a burnt area PNSE. This aimed to detect change closely observe vegetation recovery after 2022 wildfire. combination OBIA achieved highest all metrics. At same time, comparison Normalized Difference Vegetation Index (NDVI) Conjunctural Land Occupation Map (COSc) 2023 year indicated PBIA resembled maps better.

Language: Английский

Citations

1
The Effects of Fire Severity on Vegetation Structural Complexity Assessed Using SAR Data Are Modulated by Plant Community Types in Mediterranean Fire-Prone Ecosystems DOI Creative Commons

Laura Jimeno-Llorente,

Elena Marcos, José Manuel Fernández‐Guisuraga

et al.

Fire, Journal Year: 2023, Volume and Issue: 6(12), P. 450 - 450

Published: Nov. 24, 2023

Vegetation structural complexity (VSC) plays an essential role in the functioning and stability of fire-prone Mediterranean ecosystems. However, we currently lack knowledge about effects increasing fire severity on VSC spatial variability, as modulated by plant community type complex post-fire landscapes. Accordingly, this study explored, for first time, effect different communities one year after leveraging field inventory Sentinel-1 C-band synthetic aperture radar (SAR) data. The field-evaluated retrieved scenarios from γ0 VV VH backscatter data featured high fit (R2 = 0.878) low predictive error (RMSE 0.112). Wall-to-wall estimates showed that types strongly response to severity, with linked regenerative strategies dominant species community. Moderate severities had a strong impact, fire, broom shrublands Scots pine forests, dominated facultative obligate seeder species, respectively. In contrast, fire-induced impacts were not significantly between moderate fire-severity resprouter i.e., heathlands Pyrenean oak forests.

Language: Английский

Citations

2
Improved Fire Severity Prediction Using Pre-Fire Remote Sensing and Meteorological Time Series: Application to the French Mediterranean Area DOI
Victor Pénot, Thomas Opitz, François Pimont

et al.

Published: Jan. 1, 2024

Fire severity, or how an environment is affected by fire, can be estimated over large areas using remotely sensed fire severity indices, such as the Relative Burnt Ratio (RBR). RBR predictions are generally based on data collected a single date immediately before aggregated time to scalar value. However, accurate temporal and spatial prediction of remains challenging. To improve predictability RBR, we build new predictive models series spanning several months fuel proxies, derived from optical remote sensing meteorological data. The approach applied fires French Mediterranean area during summers 2016-2021 period. Lagged Generalized Additive Model (LGAM) Functional Linear (FLM) used estimate influence explanatory variables up prior while (GAM), which relies immediate pre-fire predictors at date, benchmark. Training carried out fire–land-cover scale with training dataset composed independent those in test datasets. FLM achieves best accuracy (R=0.68, RMSE=0.057) compared LGAM (R=0.60, RMSE=0.063) benchmark (R=0.52, RMSE=0.069) also less sensible overfitting. selected correctly predicts even highest values when Normalized Difference Vegetation Index decreases faster than average fire-weather Duff Moisture Code increases 65 days fire. 17% decrease RMSE GAM shows that knowledge dynamics two provides valuable information for ranking according severity.

Language: Английский

Citations

0