Post-fire Assessment Research Articles

A Comparison of Spectral Angle Mapper and Artificial Neural Network Classifiers Combined with Landsat TM Imagery Analysis for Obtaining Burnt Area Mapping

Satellite remote sensing, with its unique synoptic coverage capabilities, can provide accurate and immediately valuable information on fire analysis and post-fire assessment, including estimation of burnt areas. In this study the potential for burnt area mapping of the combined use of Artificial Neural Network (ANN) and Spectral Angle Mapper (SAM) classifiers with Landsat TM satellite imagery was evaluated in a Mediterranean setting. As a case study one of the most catastrophic forest fires, which occurred near the capital of Greece during the summer of 2007, was used. The accuracy of the two algorithms in delineating the burnt area from the Landsat TM imagery, acquired shortly after the fire suppression, was determined by the classification accuracy results of the produced thematic maps. In addition, the derived burnt area estimates from the two classifiers were compared with independent estimates available for the study region, obtained from the analysis of higher spatial resolution satellite data. In terms of the overall classification accuracy, ANN outperformed (overall accuracy 90.29%, Kappa coefficient 0.878) the SAM classifier (overall accuracy 83.82%, Kappa coefficient 0.795). Total burnt area estimates from the two classifiers were found also to be in close agreement with the other available estimates for the study region, with a mean absolute percentage difference of ∼1% for ANN and ∼6.5% for SAM. The study demonstrates the potential of the examined here algorithms in detecting burnt areas in a typical Mediterranean setting.

Combining ASTER multispectral imagery analysis and support vector machines for rapid and cost-effective post-fire assessment: a case study from the Greek wildland fires of 2007

Abstract. Remote sensing is increasingly being used as a cost-effective and practical solution for the rapid evaluation of impacts from wildland fires. The present study investigates the use of the support vector machine (SVM) classification method with multispectral data from the Advanced Spectral Emission and Reflection Radiometer (ASTER) for obtaining a rapid and cost effective post-fire assessment in a Mediterranean setting. A further objective is to perform a detailed intercomparison of available burnt area datasets for one of the most catastrophic forest fire events that occurred near the Greek capital during the summer of 2007. For this purpose, two ASTER scenes were acquired, one before and one closely after the fire episode. Cartography of the burnt area was obtained by classifying each multi-band ASTER image into a number of discrete classes using the SVM classifier supported by land use/cover information from the CORINE 2000 land nomenclature. Overall verification of the derived thematic maps based on the classification statistics yielded results with a mean overall accuracy of 94.6% and a mean Kappa coefficient of 0.93. In addition, the burnt area estimate derived from the post-fire ASTER image was found to have an average difference of 9.63% from those reported by other operationally-offered burnt area datasets available for the test region.

Using fuzzy C-means and local autocorrelation to cluster satellite-inferred burn severity classes

Burn severity classifications derived from multitemporal Landsat Thematic Mapper images and the Normalised Burn Ratio (NBR) are commonly used to assess the post-fire ecological effects of wildfires. Ongoing efforts to retrospectively map historical burn severity require defensible, objective methods of classifying continuous differenced Normalised Burn Ratio (dNBR) data where field data are often unavailable. For three fires, we compare three methods of classifying pre- and post-fire Landsat data: (1) dNBR classification using Composite Burn Index (CBI) field data to assign severity classes; (2) fuzzy C-means classification of a dNBR image; (3) local Getis–Ord statistic (Gi*) output applied to a dNBR image, classified using fuzzy C-means clustering. We then use a Kappa statistic to evaluate the agreement of severity classes assigned to a pixel with its corresponding CBI plot. For two of the three fires, the C-means clustering of the dNBR and the Gi* output performed as well or better than dNBR images classified using CBI data, with strong agreement for moderate- and high-severity classes. These results suggest that clustering of dNBR data may be a suitable approach for classifying burn severity data without field data. This method may also be useful as a tool for rapid post-fire assessments (e.g. Burned Area Emergency Response and Burned Area Reflectance Classification maps), where images must often be classified quickly and subjectively. Further analysis using additional field data and across different vegetation types will be necessary to better understand the importance of localised spatial variability in classifying burn severity data or other remote sensing change-detection analyses.

Burn severity mapping using simulation modelling and satellite imagery

Although burn severity maps derived from satellite imagery provide a landscape view of fire impacts, fire effects simulation models can provide spatial fire severity estimates and add a biotic context in which to interpret severity. In this project, we evaluated two methods of mapping burn severity in the context of rapid post-fire assessment for four wildfires in western Montana using 64 plots as field reference: (1) an image-based burn severity mapping approach using the Differenced Normalised Burn Ratio, and (2) a fire effects simulation approach using the FIREHARM model. The image-based approach was moderately correlated with percentage tree mortality but had no relationship with percentage fuel consumption, whereas the simulation approach was moderately correlated with percentage fuel consumption and weakly correlated with percentage tree mortality. Burn severity maps produced by the two approaches had mixed results among the four sampled wildfires. Both approaches had the same overall map agreement when compared with a sampled composite burn index but the approaches generated different severity maps. Though there are limitations to both approaches and more research is needed to refine methodologies, these techniques have the potential to be used synergistically to improve burn severity mapping capabilities of land managers, enabling them to quickly and effectively meet rehabilitation objectives.

Collapse rates of hollow-bearing trees following low intensity prescription burns in the Pilliga forests, New South Wales

Hollows in trees are recognized as a critical and threatened resource for a wide range of fauna in Australian forests and woodlands, yet little data are available on the impact of fire on hollow-bearing trees. We report an opportunistic, post-fire assessment of the proportion of burnt, hollow-bearing trees that collapsed in stands near roads following low intensity prescription burns in three areas of mixed eucalypt forest in the Pilliga forests. Mean collapse rates on 29 plots (40 by 50m), separated by burn Area, ranged from 14?26% for a total of 329 burnt hollow-bearing trees. Collapse rates on individual plots ranged from 0?50%. Collapsed, hollow-bearing trees were predominantly older, with 40% of senescent trees and 44% of live stags collapsing. The best predictor in models of tree collapse was the presence of a basal fire entry point. We cannot determine the extent to which collapse rates on our plots are representative of burnt areas away from containment roads due to sampling limitations, but they appear to be higher than those reported from wildfire and more intense prescription burns in southern Australia. Our results point to an urgent need for comprehensively designed studies to address the impacts of prescribed burns on hollow-bearing trees.

Assessment of Post-fire Soil Erosion Risk in Fire-Affected Watersheds Using Remote Sensing and GIS

Soil erosion is a prominent cause of land degradation and desertification in Mediterranean countries. The detrimental effects of soil erosion are exemplified in climate (in particular climate change), topography, human activities, and natural disasters. Forest fires, which are an integral part of Mediterranean ecosystems, are responsible for the destruction of above-and below-ground vegetation that protects against soil erosion. Under this perspective, the estimation of potential soil erosion, especially after fire events, is critical for identifying watersheds that require management to prevent sediment loss, flooding, and increased ecosystem degradation. The objective of this study was to model the potential post-fire soil erosion risk following a large and intensive wildland fire, in order to prioritize protection and management actions at the watershed level in a Mediterranean landscape. Burn severity and preand post-fire land cover/uses were mapped using an ASTER image acquired two years before the fire, air photos acquired shortly after the fire, and a Landsat TM image acquired within one month after-fire. We estimated pre-and post-fire sediment loss using an integrated GIS-based approach, and additionally we analyzed landscape erosion patterns. The overall accuracy of the severity map reached 83%. Severe and heavy potential erosion classes covered approximately 90% of the total area following the fire, compared to 55% before. The fire had a profound effect on the spatial erosion pattern by altering the distribution of the potential erosion classes in 21 out of 24 watersheds, and seven watersheds were identified as being the most vulnerable to post-fire soil erosion. The spatial pattern of the erosion process is important because landscape cover heterogeneity induced especially by fire is a dominant factor controlling runoff generation and erosion rate, and should be considered in post-fire erosion risk assessment.

Burned area mapping and post-fire impact assessment in the Kassandra peninsula (Greece) using Landsat TM and Quickbird data

The purpose of this study was to assess the environmental impacts of forest fires on part of the Mediterranean basin. The study area is on the Kassandra peninsula, prefecture of Halkidiki, Greece. A maximum likelihood supervised classification was applied to a post-fire Landsat TM image for mapping the exact burned area. Land-cover types that had been affected by fire were identified with the aid of a CORINE land-cover type layer. Results showed an overall classification accuracy of 95%, and 83% of the total burned area was ‘forest areas’. A normalized difference vegetation index threshold technique was applied to a post-fire Quickbird image which had been recorded six years after the fire event to assess the vegetation recovery and to identify the vegetation species that were dominant in burned areas. Four classes were identified: ‘bare soil’, ‘sparse shrubs’, ‘dense shrubs’ and ‘tree and shrub communities’. Results showed that ‘shrublands’ is the main vegetation type which has prevailed (65%) and that vegetation recovery is homogeneous in burned areas.

Operational active fire mapping and burnt area identification applicable to Mexican Nature Protection Areas using MODIS and NOAA-AVHRR direct readout data

Since 1999, the National Commission for the Knowledge and Use of the Biodiversity (CONABIO) in Mexico has been developing and managing the “Operational program for the detection of hot-spots using remote sensing techniques”. This program uses images from the MODerate resolution Imaging Spectroradiometer (MODIS) onboard the Terra and Aqua satellites and from the Advanced Very High Resolution Radiometer of the National Oceanic and Atmospheric Administration (NOAA-AVHRR), which are operationally received through the Direct Readout station (DR) at CONABIO. This allows the near-real time monitoring of fire events in Mexico and Central America. In addition to the detection of active fires, the location of hot spots are classified with respect to vegetation types, accessibility, and risk to Nature Protection Areas (NPA). Besides the fast detection of fires, further analysis is necessary due to the considerable effects of forest fires on biodiversity and human life. This fire impact assessment is crucial to support the needs of resource managers and policy makers for adequate fire recovery and restoration actions. CONABIO attempts to meet these requirements, providing post-fire assessment products as part of the management system in particular for satellite-based burnt area mapping. This paper provides an overview of the main components of the operational system and will present an outlook to future activities and system improvements, especially the development of a burnt area product. A special focus will also be placed on the fire occurrence within NPAs of Mexico.

Factors affecting post-fire crown regeneration in cork oak (Quercus suber L.) trees

Cork oak (Quercus suber) forests are acknowledged for their biodiversity and economic (mainly cork production) values. Wildfires are one of the main threats contributing to cork oak decline in the Mediterranean Basin, and one major question that managers face after fire in cork oak stands is whether the burned trees should be coppiced or not. This decision can be based on the degree of expected crown regeneration assessed immediately after fire. In this study we carried out a post-fire assessment of the degree of crown recovery in 858 trees being exploited for cork production in southern Portugal, 1.5 years after a wildfire. Using logistic regression, we modelled good or poor crown recovery probability as a function of tree and stand variables. The main variables influencing the likelihood of good or poor crown regeneration were bark thickness, charring height, aspect and tree diameter. We also developed management models, including simpler but easier to measure variables, which had a lower predictive power but can be used to help managers to identify, immediately after fire, trees that will likely show good crown regeneration, and trees that will likely die or show poor regeneration (and thus, potential candidates for trunk coppicing).

Burn severity estimation from remotely sensed data: Performance of simulation versus empirical models

Burn severity is a key factor in post-fire assessment and its estimation is traditionally restricted to field work and empirical fitting from remotely sensed data. However, the first method is limited in terms of spatial coverage and cost effectiveness and the second is site- and data-specific. Since alternative approaches based on radiative transfer models (RTM) have been usefully applied in retrieving several biophysical plant parameters (leaf area index, water and dry matter content, chlorophyll), this paper has applied the inversion of a simulation model to estimate burn severity in terms of the Composite Burn Index (CBI). The performance of the model inversion method was compared to standard empirical techniques. The study area chosen was a large forest fire in central Spain which occurred in July 2005. The model inversion showed the most accurate estimation for high severity levels (for CBI > 2.7, RMSE = 0.30) and for unburned areas (CBI < 0.5, RMSE = 0). In both methodologies, the error associated to CBI from 0.5 to 2.7 was not acceptable (RMSE > 0.7), because it is higher than 25% of the total range of the index. Finally, burn severity maps from both methods were compared.

Remote sensing information for fire management and fire effects assessment

Over the past decade, much research has been carried out on the utilization of advanced geospatial technologies (remote sensing and geographic information systems) in the fire science and fire management disciplines. Recent advances in these technologies were the focus of a workshop sponsored by the EARSEL special interest group (SIG) on forest fires (FF‐SIG) and the Global Observation of Forest and Land Cover Dynamics (GOFC‐GOLD) fire implementation team. Here we summarize the framework and the key findings of papers submitted from this meeting and presented in this special section. These papers focus on the latest advances for near real‐time monitoring of active fires, prediction of fire hazards and danger, monitoring of fuel moisture, mapping of fuel types, and postfire assessment of the impacts from fires.

Combining AVHRR imagery with CORINE Land Cover data to observe forest fires and to assess their consequences

This paper describes an operational application of AVHRR satellite imagery in combination with the satellite-based land cover database CORINE Land Cover (CLC) for the comprehensive observation and follow-up of 10000 fire outbreaks and of their consequences in Greece during summer 2000. In the first stage, we acquired and processed satellite images on a daily basis with the aim of smoke-plume tracking and fire-core detection at the national level. Imagery was acquired eight times per day and derived from all AVHRR spectral channels. In the second stage, we assessed the consequences of fire on vegetation by producing a burnt-area map on the basis of multi-annual normalised vegetation indices using AVHRR data in combination with CLC. In the third stage we used again CLC to assess the land cover of burnt areas in the entire country. The results compared successfully to available inventories for that year. Burnt area was estimated with an accuracy ranging from 88% to 95%, depending on the predominant land cover type. These results, along with the low cost and high temporal resolution of AVHRR satellite imagery, suggest that the combination of low-resolution satellite data with harmonised CLC data can be applied operationally for forest fire and post-fire assessments at the national and at a pan-European level.