Fig 2 - uploaded by AT Hudak
Content may be subject to copyright.
Inset map of Alaska showing location and extent of the Taylor Complex. The locations of the field sites and image collection areas are shown in the larger image.
Source publication
Wildfire is a major forest disturbance in interior Alaska that can both directly and indirectly alter ecological processes. We used a combination of pre-and post-fire forest floor depths and post-fire ground cover assessments measured in the field, and high-resolution airborne hyperspectral imagery, to map forest floor conditions after the 2004 Tay...
Contexts in source publication
Context 1
... sites were located on the Porcupine, Chicken and Wall Street Fires that burned in Alaska's interior boreal forest (Fig. 2). The Porcupine Fire began on 21 June 2004 and burned 115 170 ha. The Chicken and Wall Street Fires began on 15 June 2004 and burned together with a combined area of 220 150 ha. These wildfires eventually merged with other large wildfires to form the Taylor Complex (528 218 ha; 63843 0 28 00 N, 142850 0 36 00 W, centroid; elevation 424 ...
Context 2
... resolution of the images was 3 m. Image geolocation was achieved with an onboard GPS (3-10-m accuracy) and integrated inertial monitoring unit. Imagery was collected for 26 000 and 34 550 ha over the field sites on the Chicken-Wall Street and Porcupine Fires respectively. All field sites on both fires were within the area of image collection (Fig. ...
Similar publications
The spectral unmixing (SU) technique is an effective method of solving the mixed pixel problem in the hyperspectral remote sensed imagery (HSI). During the process, endmember extraction algorithm (EEA) is significant for the creation of material abundance maps. However, the traditional EEAs are not very reliable due to the low resolution of sensor...
Wildfi re effects on the ground surface are indicative of the potential for post-fi re watershed erosion response. Areas with remaining organic ground cover will likely experience less erosion than areas of complete ground cover combustion or exposed mineral soil. The Simi and Old fi res burned ~67,000 ha in southern California in 2003. Burn severi...
The N-FINDR, developed by Winter, is one of the most widely used algorithms for endmember extraction for hyperspectral images. N-FINDR usually needs an outer loop to control the stopping rule and two inner loops for pixel replacement, so it suffers from computational inefficiency, particularly when the size of the remote-sensing image is large. Rec...
In the last few decades, researchers have developed a plethora of hyperspectral Earth Observation (EO) remote sensing techniques, analysis and applications. While hyperspectral exploratory sensors are demonstrating their potential, Sentinel-2 multispectral satellite remote sensing is now providing free, open, global and systematic high resolution v...
Derivative analysis of spectral data was used as a technique to study the variation of optically active constituents (OACs) of water, using field data and hyperspectral imagery of EO-1 Hyperion orbital sensor. The Hyperion image used in this study was acquired on June 23, 2005, at the end of the high water period for the Amazon River. A field campa...
Citations
... While we may expect large C losses under future fire regimes, the effects of fire on soil can be highly variablei.e., individual patches of soil may have significantly different heating and combustion during a wildfire. For example, numerous studies have documented high spatial heterogeneity in the degree of O horizon combustion during a single fire with areas of unburned O horizon existing adjacent to bare mineral soil (Kasischke and Johnstone, 2005;Lewis et al. 2011). Thus, in order to understand the impact of fire on soil C cycling, we must consider variability in fire intensity. ...
Background
Boreal forests cover vast areas of land in the northern hemisphere and store large amounts of carbon (C) both aboveground and belowground. Wildfires, which are a primary ecosystem disturbance of boreal forests, affect soil C via combustion and transformation of organic matter during the fire itself and via changes in plant growth and microbial activity post-fire. Wildfire regimes in many areas of the boreal forests of North America are shifting towards more frequent and severe fires driven by changing climate. As wildfire regimes shift and the effects of fire on belowground microbial community composition are becoming clearer, there is a need to link fire-induced changes in soil properties to changes in microbial functions, such as respiration, in order to better predict the impact of future fires on C cycling.
Results
We used laboratory burns to simulate boreal crown fires on both organic-rich and sandy soil cores collected from Wood Buffalo National Park, Alberta, Canada, to measure the effects of burning on soil properties including pH, total C, and total nitrogen (N). We used 70-day soil incubations and two-pool exponential decay models to characterize the impacts of burning and its resulting changes in soil properties on soil respiration. Laboratory burns successfully captured a range of soil temperatures that were realistic for natural wildfire events. We found that burning increased pH and caused small decreases in C:N in organic soil. Overall, respiration per gram total (post-burn) C in burned soil cores was 16% lower than in corresponding unburned control cores, indicating that soil C lost during a burn may be partially offset by burn-induced decreases in respiration rates. Simultaneously, burning altered how remaining C cycled, causing an increase in the proportion of C represented in the modeled slow-cycling vs. fast-cycling C pool as well as an increase in fast-cycling C decomposition rates.
Conclusions
Together, our findings imply that C storage in boreal forests following wildfires will be driven by the combination of C losses during the fire itself as well as fire-induced changes to the soil C pool that modulate post-fire respiration rates. Moving forward, we will pair these results with soil microbial community data to understand how fire-induced changes in microbial community composition may influence respiration.
... Wildland fuel is defined as the vegetative material or biomass that burns The increasing usage of remote sensing tools to aid disaster response offers fire agencies an opportunity for additional surveillance. Optical imagery (French et al., 2008;Van Wagtendonk et al., 2004), as well as multi-and hyper-spectral imagery (Hislop et al., 2018;Lewis et al., 2011;Veraverbeke et al., 2010), have been shown to detect fire occurrence, burn severity, and burned area. However, smoke and weather conditions can impede current monitoring methods which rely on field survey and optical imagery. ...
Many communities coexist with wildfires that lead to loss of lives, property, and ecosystem services. Remote sensing tools can aid disaster response and post‐event assessment, offering fire agencies opportunities for additional surveillance with radar, an all‐weather instrument that can image day or night. The Station (2009) and Bobcat (2020) Fires are the two largest fires in Los Angeles County history, each burning over 100,000 acres. These areas are imaged with NASA's Uninhabited Aerial Vehicle Synthetic Aperture Radar L‐band instrument. We test whether polarimetric radar can detect fire scars, burn severity, and different fuel types through its sensitivity to different scattering mechanisms. Polarimetric SAR products are moved into geographic information system‐friendly formats, and in lieu of available field measurements are analyzed alongside agency data showing fire perimeters, burn progression outlines, and soil burn severity. We find that the HV polarization returns and the primary scattering mechanism, quantified through the Cloude‐Pottier decomposition, are the most sensitive parameters. Higher HV values pre‐fire correspond well to areas of moderate and high soil burn severity, and the pattern of fire progression follows higher HV to some extent. Using an HV difference threshold of 1.5 dB, the Bobcat burn scar is identified at 0.70 accuracy when compared with the published fire perimeter. Alpha 1 Angle can also demonstrate sensitivity to soil burn severity pre‐ and post‐fire, showing vegetation types with increased surface scattering post‐fire, which can be used to map burn scars and track recovery by vegetation type.
... Previous studies reported the potential of narrowband hyperspectral remote sensing data collected by airborne or the technologydemonstrator Hyperion spectrometers to produce meaningful, physically-based estimates of fire severity through the fractional cover retrieval of representative ground components in a post-fire landscape (Lewis et al., , 2011Veraverbeke et al., 2014;Fernández-Manso et al., 2019). Nevertheless, these findings are limited by the low ability to evaluate large surfaces and high costs of airborne acquisitions , as well as the low signal-to-noise ratio and narrow swath width of Hyperion scenes (Papeş et al., 2010). ...
... Second, the PRISMA retrieval performance of sub-pixel components directly related to fire severity was remarkably high in spite of the lower spatial resolution than airborne spectrometers, and in line with previous research involving airborne hyperspectral data (e.g. Lewis et al., 2011;Veraverbeke et al., 2014). ...
... This finding agrees with previous studies suggesting that the grouping of NPV and soil in a unique spectral library (NPVS) provided accurate results in the classification of burned landscapes Quintano et al., 2017;Quintano et al., 2019). The fewer potential MESMA models required by the PRISMA scene for subpixel fraction estimation than the Sentinel-2 scene may be associated with the increased spectral data volume of PRISMA, and thus the improved capture of distinctive spectral features within classes (Lewis et al., 2011;Degerickx et al., 2019). This may also explain both the high number of unmixed pixels that included a char endmember and the low char fraction noise from the PRISMA post-fire scene. ...
... Hyperspectral images (HSIs) capture hundreds of electromagnetic spectral bands, which extract spectral signatures for observing objects [19]. Thus, HSIs are widely used in the applications with huge data analysis, such as amount environmental studies, military surveillance, mineral detection and biomedical imaging, etc. [27,43,63]. However, due to the physical limitations of sensors, HSIs are often corrupted by several types of noise, i.e., random noise, stripe noise, and dead pixels [31,38,42]), in the acquisition process. ...
During the last decades, learning-based deep neural network (DNN) has shown its advantages on hyperspectral image (HSI) denoising. Compared to classical prior-based methods, DNN-based algorithms employ a larger scale of training samples for learning to simulate the complex image generation process with higher accuracy. However, most DNN-based HSI denoising methods are designed by a superposition convolution layer, which cannot fully use the frequency information in the image itself, especially the information containing a strong response to noise in high-frequency domain. Thus, we propose a high-frequency attention network (HFAN) assisted by both spectral and spatial high-frequency information to achieve accurate HSIs denoising in this paper. Our proposed HFAN comprises a high-frequency and denoising branch, and the auxiliary function of high-frequency information is realized by transmitting the characteristic information of the high-frequency component to the denoising branch. Specifically, the spatial-spectral attention (SSA) module is presented to recover more detail in space and spectra. Experiments on synthetic and real HSI data show that our proposed HFAN achieves better denoising results compare to the other advanced methods.
... In this study, the fire severity definition will be considered since the assessment of wildfire effects in the short term is essential for addressing on site stabilization/emergency actions Quintano et al., 2013) aimed at minimizing the most adverse ecological effects throughout the burned landscape. Previous research has established reliable methods for estimating fire severity in the field, based on the measurement of indicators such as percent change in canopy cover and basal area (Barden and Woods, 1976;Miller et al., 2009), height and depth of stem char (Hood et al., 2008), canopy scorch and consumption (Thompson and Spies, 2009), minimum tip diameter of remaining branches (Pérez and Moreno, 1998), forest-floor burn depth (Lewis et al., 2011), ash cover and depth (De Luis et al., 2003;Hudak et al., 2013), soil color and structure (Ketterings and Bingham, 2000;Vega et al., 2013) or soil chemical properties (Rodríguez-Alleres et al., 2012). Other common field-based assessments, such as the Composite Burn Index (CBI; Key and Benson, 2005), the Geometrically structured CBI (GeoCBI; De Santis and ) and their predecessors (Ryan and Noste, 1983) and variants (Fernandes et al., 2010), are based on measuring several fire severity attributes using a multi-strata approach, rather than a single indicator. ...
... The increasing usage of remote sensing tools to aid disaster response offers fire agencies an opportunity for additional surveillance. Optical imagery (French et al., 2008;Van Wagtendonk et al., 2004), as well as multi-and hyper-spectral imagery (Hislop et al., 2018;Lewis et al., 2011;Veraverbeke et al., 2010), have been shown to detect fire occurrence, burn severity, and burned area. However, smoke and weather conditions can impede current monitoring methods which rely on field survey and optical imagery. ...
Many communities coexist with wildfires that can lead to loss of lives, property, and ecosystem services. The increasing usage of remote sensing tools to aid disaster response and post-event assessment offers fire agencies an opportunity for additional surveillance. The adaptability of radar instruments in their ability to see through smoke, haze, and clouds during the day or night is especially relevant when cloud cover or lack of solar illumination inhibits traditional visual surveys of damage. The Station (2009) and Bobcat (2020) Fires are the two largest fires in Los Angeles County history, each burning over 100,000 acres. These areas are imaged with NASA’s UAVSAR (Uninhabited Aerial Vehicle Synthetic Aperture Radar) L-band synthetic aperture radar. For these neighboring fires, we investigate the usage of polarimetric radar products to detect fire scars, burn severity, and different fuel (vegetation) types. These fire characteristics are observed using individual HV (horizontally emitted, vertically collected) images and in eigenvector decomposition products derived from quad-polarimetric data. Traditionally unintuitive, yet powerful PolSAR (polarimetric SAR) products are moved into GIS-friendly (geographic information system) formats to be analyzed alongside agency data such as fire perimeters, burn progression outlines, and soil burn severity. We demonstrate the advantages of combining PolSAR with GIS datasets and methods to understand the fuel loads which contributed to the fires and to monitor post-fire vegetation recovery.
... By doing so, dNBR fire severity studies do not capitalize on the advantage of the wealth of available spectral information. Spectral mixture analysis (SMA) has been applied in multispectral postfire fire severity studies; however, some studies have capitalized upon hyperspectral data [27,28,[53][54][55][56][57][58][59][60]. The main advantage of SMA in fire severity assessments is that it does not require field data calibrations as it quantifies the abundance of ground cover classes. ...
Fire severity represents fire-induced environmental changes and is an important variable
for modeling fire emissions and planning post-fire rehabilitation. Remotely sensed fire severity is traditionally evaluated using the differenced normalized burn ratio (dNBR) derived from multispectral imagery. This spectral index is based on bi-temporal differenced reflectance changes caused by fires in the near-infrared (NIR) and short-wave infrared (SWIR) spectral regions. Our study aims to evaluate the spectral sensitivity of the dNBR using hyperspectral imagery by identifying the optimal bi-spectral NIR SWIR combination. This assessment made use of a rare opportunity arising from the pre- and post-fire airborne image acquisitions over the 2013 Rim and 2014 King fires in California
with the Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) sensor. The 224 contiguous bands of this sensor allow for 5760 unique combinations of the dNBR at a high spatial resolution of approximately 15 m. The performance of the hyperspectral dNBR was assessed by comparison against field data and the spectral optimality statistic. The field data is composed of 83 in situ measurements of fire severity using the Geometrically structured Composite Burn Index (GeoCBI) protocol. The optimality statistic ranges between zero and one, with one denoting an optimal measurement of the
fire-induced spectral change. We also combined the field and optimality assessments into a combined score. The hyperspectral dNBR combinations demonstrated strong relationships with GeoCBI field data. The best performance of the dNBR combination was derived from bands 63, centered at 0.962 µm, and 218, centered at 2.382 µm. This bi-spectral combination yielded a strong relationship with GeoCBI field data of R2 = 0.70 based on a saturated growth model and a median spectral index optimality statistic of 0.31. Our hyperspectral sensitivity analysis revealed optimal NIR and SWIR bands for the composition of the dNBR that are outside the ranges of the NIR and SWIR bands of the Landsat 8 and Sentinel-2 sensors. With the launch of the Precursore Iperspettrale Della Missione Applicativa (PRISMA) in 2019 and several planned spaceborne hyperspectral missions, such as the Environmental Mapping and Analysis Program (EnMAP) and Surface Biology and Geology (SBG), our study provides a timely assessment of the potential and sensitivity of hyperspectral data for assessing fire severity.
... proper timing of imagery) address some of these challenges (Key, 2005;Meddens et al., 2016;Picotte and Robertson, 2011a). Accuracy of prediction is predicated on ecological parameters used to define burn severity, which have varied widely (Jain, 2004;Keeley, 2009;Lentile et al., 2006;Lewis et al., 2011). Questions have also have been raised about whether NBR is as theoretically grounded as a remote-sensing index for differencing pre-burn versus post-burn conditions (Roy et al., 2006), particularly compared to Mid-Infrared Bispectral Index (MIRBI;McCarley et al., 2018;Trigg and Flasse, 2001). ...
Identifying meaningful measures of ecological change over large areas is dependent on the quantification of robust relationships between ecological metrics and remote sensing products. Over the past several decades, ground observations of wildfire and prescribed fire severity have been acquired across hundreds of wildland fires in the United States, primarily utilizing the Composite Burn Index (CBI) plot protocol. These observations have been coupled to spaceborne passive spectral reflectance indices (e.g. Landsat-derived variations of the Normalized Burn Ratio [NBR]) to produce regression models describing their relationship. Here we develop regression models by vegetation type for multiple vegetation classification systems representing a range of spatial scales, and a decision tree framework for evaluating these regression models. Our overall goals were to determine which scale of ecological classifications provided the best estimate of burn severity from Landsat data and how to choose the best regression model. We aggregated a total of 6280 CBI plots for 234 wildland fires that burned between 1994 and 2017 and produced Landsat-derived NBR and differenced NBR (dNBR) values for each plot. We then calculated best fit linear or higher order regression equations between CBI and NBR/dNBR for each landcover classification system from smallest to largest scale: LANDFIRE Biophysical Settings (BPS), National Vegetation Classification macrogroup (NVC) landcover classifications, Omernick III, II, and I ecoregions, LANDFIRE Fire Regime Groups (FRG), and the entire conterminous United States (CONUS) dataset. The CONUS regression model goodness of fit was moderate (R² = 0.55, P < 0.001) for dNBR and poor (R² = 0.30, P < 0.001) for NBR. Within landcover classifications, CBI was better fit by dNBR than NBR. Finer scale regional regression models including BPS (dNBR R2¯ = 0.56 and 0.00–0.83 R² range; NBR R2¯ = 0.43 and 0.00–0.82 R² range) and NVC (dNBR R2¯ = 0.55 and 0.15–0.78 R² range; NBR R2¯ = 0.41 and 0.00–0.79 R² range) were on average the same or better than the CONUS models for dNBR and NBR, with the strongest fit models exhibiting R² ≥ 0.70, whereas larger scale regional models R2¯ ranged from 0.28 to 0.5. However, variation in accuracy among landcover types indicate that dNBR and NBR regression models could be used to effectively estimate CBI for future fires in certain regions, while for other regions models may require additional field observations or alternative spectral transformations. Our decision tree schema can be used to help users determine which scale is likely to produce the most accurate results using our models. The CBI regression models developed here, paired with the decision tree, provide users with a simple method to estimate burn severity in units of CBI for any fire within CONUS with moderate to high levels of confidence and provide a template for further development of models with new data going forward.
... Forests present different environmental services to the world and play a significant role in sustaining the vital environmental processes including climate adjustment, ecological balance, water and soil conservation, and carbon storage (UNEP 2007). However, it has been proved that the total forest area in the world is decreasing because of many reasons, such as forest fires due to climate change and anthropogenic interferences in the natural areas (Lewis et al. 2011). Forest fires can cause many environmental problems, social issues, as well as economic losses (Herawati et al. 2006;Valdez et al. 2017). ...
The forest fire hazard mapping using the accurate models in the fire-prone areas has particular importance to predict the future fire occurrence and allocate the resources for preventing the fire ignition. This research aimed to compare the accuracy of some individual models including boosted regression tree (BRT), classification and regression trees (CART), functional discriminant analysis (FDA), generalized linear model (GLM), mixture discriminant analysis (MDA), random forest (RF) and two new hybrid models including FDA-GLM-MDA and RF-CART-BRT for predicting the fire hazard in a fire-prone area in the northeast Iran, Golestan Province. For this purpose, a comprehensive dataset from ten effective parameters including digital elevation model (DEM), slope angle (SA), plan curvature (PC), topographic wetness index (TWI), annual rainfall mean (ARM), annual temperature mean (ATM), wind effect (WE), distance to urban areas (DTU), distance to streams (DTS) and distance to roads (DTR) was created in GIS. Furthermore, 3705 historical fire locations in the Golestan Province from 2002 to 2017 were obtained from MODIS fire product dataset. Then, the variable importance was assessed using the XGBoost machine learning (ML) technique. Finally, the individual and hybrid models were evaluated using the ROC-AUC method. The results showed that the DTU was the most important factor in modeling and mapping the fire hazard in the Golestan Province. Also, the results demonstrated that the individual random forest (RF) (AUC = 0.855) and hybrid RF-CART-BRT algorithms (AUC = 0.854) were the most accurate predictive models for mapping the fire hazard in the Golestan Province, respectively. Considering the high significance of DTU in fire occurrence in this study, the area of fire hazard classes in fourteen different counties of the Golestan Province was calculated using the most accurate model (RF model). The final results indicated that Minudasht County had the most area (35.55%) of fire hazard in the very high fire hazard class. The results of this study are very useful for local forest managers to control the future fires using the best model in the natural areas of the Golestan Province, especially in Minudasht County. The protective management of the natural areas of the Golestan Province would be performed based on the fire hazard maps produced by RF and RF-CART-BRT algorithms. We recommend applying these models for fire danger mapping in fire-prone areas around the world which have semi-arid conditions. More comparative assessment of individual and ensemble models for fire danger mapping in semi-arid areas around the world could provide a baseline for monitoring fire danger in similar conditions.
... Some traditional fuel measurement methods can be colocated to provide direct measures of consumption, such as the wirelog method (Hedin and Turner 1977, Brown et al. 1991, Albini et al. 1995, to estimate consumption of logs, and the duff pin method (Beaufait et al. 1975, Lewis et al. 2011 for forest floor consumption. However, destructive sampling techniques that harvest, dry, and weigh fuels before the fire necessarily prevent postfire measurements at the same locations. ...
Methods to accurately estimate spatially explicit fuel consumption are needed because consumption relates directly to fire behavior, effects, and smoke emissions. Our objective was to quantify sparkleberry (Vaccinium arboretum Marshall) shrub fuels before and after six experimental prescribed fires at Fort Jackson in South Carolina. We used a novel approach to characterize shrubs non-destructively from three-dimensional (3D) point cloud data collected with a terrestrial laser scanner. The point cloud data were reduced to 0.001 m–3 voxels that were either occupied to indicate fuel presence or empty to indicate fuel absence. The density of occupied voxels was related significantly by a logarithmic function to 3D fuel bulk density samples that were destructively harvested (adjusted R2 = .32, P < .0001). Based on our findings, a survey-grade Global Navigation Satellite System may be necessary to accurately associate 3D point cloud data to 3D fuel bulk density measurements destructively collected in small (submeter) shrub plots. A recommendation for future research is to accurately geolocate and quantify the occupied volume of entire shrubs as 3D objects that can be used to train models to map shrub fuel bulk density from point cloud data binned to occupied 3D voxels.
