skip to main content
US FlagAn official website of the United States government
dot gov icon
Official websites use .gov
A .gov website belongs to an official government organization in the United States.
https lock icon
Secure .gov websites use HTTPS
A lock ( lock ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites.

Explore Research Products in the PAR
It may take a few hours for recently added research products to appear in PAR search results.


Title: Modeling wildland fire burn severity in California using a spatial Super Learner approach
Given the increasing prevalence of wildland fires in the Western US, there is a crit- ical need to develop tools to understand and accurately predict burn severity. We develop a novel machine learning model to predict post-fire burn severity using pre- fire remotely sensed data. Hydrological, ecological, and topographical variables col- lected from four regions of California — the site of the Kincade fire (2019), the CZU Lightning Complex fire (2020), the Windy fire (2021), and the KNP Fire (2021) — are used as predictors of the differenced normalized burn ratio. We hypothesize that a Super Learner (SL) algorithm that accounts for spatial autocorrelation using Vec- chia’s Gaussian approximation will accurately model burn severity. We use a cross- validation study to show that the spatial SL model can predict burn severity with reasonable classification accuracy, including high burn severity events. After fitting and verifying the performance of the SL model, we use interpretable machine learn- ing tools to determine the main drivers of severe burn damage, including greenness, elevation, and fire weather variables. These findings provide actionable insights that enable communities to strategize interventions, such as early fire detection systems, pre-fire season vegetation clearing activities, and resource allocation during emer- gency responses. When implemented, this model has the potential to minimize the loss of human life, property, resources, and ecosystems in California.  more » « less
Award ID(s):
2051010
PAR ID:
10500352
Author(s) / Creator(s):
; ; ; ; ; ; ;
Publisher / Repository:
Springer
Date Published:
Journal Name:
Environmental and Ecological Statistics
Edition / Version:
Published version
ISSN:
1352-8505
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. ; ; ; (, Earth Surface Processes and Landforms)
    Abstract Post‐fire debris flows represent one of the most erosive consequences associated with increasing wildfire severity and investigations into their downstream impacts have been limited. Recent advances have linked existing hydrogeomorphic models to predict potential impacts of post‐fire erosion at watershed scales on downstream water resources. Here we address two key limitations in current models: (1) accurate predictions of post‐fire debris flow volumes in the absence of triggering storm rainfall intensities and (2) understanding controls on grain sizes produced by post‐fire debris flows. We compiled and analysed a novel dataset of depositional volumes and grain size distributions (GSDs) for 59 post‐fire debris flows across the Intermountain West (IMW) collected via fieldwork and from the literature. We first evaluated the utility of existing models for post‐fire debris flow volume prediction, which were largely developed for Southern California. We then constructed a new post‐fire debris flow volume prediction model for the IMW using a combination of Random Forest modelling and regression analysis. We found topography and burn severity to be important variables, and that the percentage of pre‐fire soil organic matter was an essential predictor variable. Our model was also capable of predicting debris flow volumes without data for the triggering storm, suggesting that rainfall may be more important as a presence/absence predictor, rather than a scaling variable. We also constructed the first models that predict the median, 16th percentile, and 84th percentile grain sizes, as well as boulder size, produced by post‐fire debris flows. These models demonstrate consistent landscape controls on debris flow GSDs that are related to land cover, physical and chemical weathering, and hillslope sediment transport processes. This work advances our ability to predict how post‐fire sediment pulses are transported through watersheds. Our models allow for improved pre‐ and post‐fire risk assessments across diverse ranges of watersheds in the IMW. 
    more » « less
  2. ; ; (, Earth System Science Data)
    Abstract. Forest fires, while destructive and dangerous, are important to the functioning and renewal of ecosystems. Over the past 2 decades, large-scale, severe forest fires have become more frequent globally, and the risk is expected to increase as fire weather and drought conditions intensify. To improve quantification of the intensity and extent of forest fire damage, we have developed a 30 m resolution global forest burn severity (GFBS) dataset of the degree of biomass consumed by fires from 2003 to 2016. To develop this dataset, we used the Global Fire Atlas product to determine when and where forest fires occurred during that period and then we overlaid the available Landsat surface reflectance products to obtain pre-fire and post-fire normalized burn ratios (NBRs) for each burned pixel, designating the difference between them as dNBR and the relative difference as RdNBR. We compared the GFBS dataset against the Canada Landsat Burned Severity (CanLaBS) product, showing better agreement than the existing Moderate Resolution Imaging Spectrometer (MODIS)-based global burn severity dataset (MOdis burn SEVerity, MOSEV) in representing the distribution of forest burn severity over Canada. Using the in situ burn severity category data available for the 2013 wildfires in southeastern Australia, we demonstrated that GFBS could provide burn severity estimation with clearer differentiation between the high-severity and moderate-/low-severity classes, while such differentiation among the in situ burn severity classes is not captured in the MOSEV product. Using the CONUS-wide composite burn index (CBI) as a ground truth, we showed that dNBR from GFBS was more strongly correlated with CBI (r=0.63) than dNBR from MOSEV (r=0.28). RdNBR from GFBS also exhibited better agreement with CBI (r=0.56) than RdNBR from MOSEV (r=0.20). On a global scale, while the dNBR and RdNBR spatial patterns extracted by GFBS are similar to those of MOSEV, MOSEV tends to provide higher burn severity levels than GFBS. We attribute this difference to variations in reflectance values and the different spatial resolutions of the two satellites. The GFBS dataset provides a more precise and reliable assessment of burn severity than existing available datasets. These enhancements are crucial for understanding the ecological impacts of forest fires and for informing management and recovery efforts in affected regions worldwide. The GFBS dataset is freely accessible at https://doi.org/10.5281/zenodo.10037629 (He et al., 2023). 
    more » « less
  3. ; ; (, Land)
    The 1988 Yellowstone fire altered the structure of the local forest ecosystem and left large non-recovery areas. This study assessed the pre-fire drivers and post-fire characteristics of the recovery and non-recovery areas and examined possible reasons driving non-recovery of the areas post-fire disturbance. Non-recovery and recovery areas were sampled with 44,629 points and 77,501 points, from which attribute values related to topography, climate, and subsequent soil conditions were extracted. We calculated the 1988 Yellowstone fire burn thresholds using the differenced Normalized Burn Ratio (dNBR) and official fire maps. We used a burn severity map from the US Forest Service to calculate the burn severity values. Spatial regressions and Chi-Square tests were applied to determine the statistically significant characteristics of a lack of recovery. The non-recovery areas were found to cover 1005.25 km2. Among 11 variables considered as potential factors driving recovery areas and 13 variables driving non-recovery areas, elevation and maximum temperature were found to have high Variance Inflation Factors (4.73 and 4.72). The results showed that non-recovery areas all experienced severe burns and were located at areas with steeper slopes (13.99°), more precipitation (871.73 mm), higher pre-fire vegetation density (NDVI = 0.38), higher bulk density (750.03 kg/m3), lower soil organic matter (165.61 g/kg), and lower total nitrogen (60.97 mg/L). Chi-square analyses revealed statistically different pre-fire forest species (p < 0.01) and soil order (p < 0.01) in the recovery and non-recovery areas. Although Inceptisols dominated in both recovery and non-recovery areas, however, the composition of Mollisols was higher in the non-recovery areas (14%) compared to the recovery areas (11%). This indicated the ecological memory of the non-recovery site reverting to grassland post-disturbance. Unlike conventional studies only focusing on recovery areas, this study analyzed the non-recovery areas and found the key characteristics that make a landscape not resilient to the 1988 Yellowstone fire. The significant effects of elevation, precipitation, and soil pH on recovery may be significant to the forest management and forest resilience in the post-fire period. 
    more » « less
  4. ; ; ; ; ; (, International Journal of Wildland Fire)
    As wildfires become larger and more severe across western North America, it grows increasingly important to understand how they will affect the biogeochemical processes influencing ecosystem recovery. Soil nitrogen (N) cycling is a key process constraining recovery rates. In addition to its direct responses to fire, N cycling can also respond to other post-fire transformations, including increases or decreases in microbial biomass, soil moisture, and pH. To examine the short-term effects of wildfire on belowground processes in the northern Sierra Nevada, we collected soil samples along a gradient from unburned to high fire severity over 10  months following a wildfire. This included immediate pre- and post-fire sampling for many variables at most sites. While season and soil moisture did not substantially alter pH, microbial biomass, net N mineralisation, and nitrification in unburned locations, they interacted with burn severity in complex ways to constrain N cycling after fire. In areas that burned, pH increased (at least initially) after fire, and there were non-monotonic changes in microbial biomass. Net N mineralisation also had variable responses to wetting in burned locations. These changes suggest burn severity and precipitation patterns can interact to alter N cycling rates following fire. 
    more » « less
  5. ; ; ; ; ; (, Remote Sensing)
    null (Ed.)
    In recent years, there have been rapid improvements in both remote sensing methods and satellite image availability that have the potential to massively improve burn severity assessments of the Alaskan boreal forest. In this study, we utilized recent pre- and post-fire Sentinel-2 satellite imagery of the 2019 Nugget Creek and Shovel Creek burn scars located in Interior Alaska to both assess burn severity across the burn scars and test the effectiveness of several remote sensing methods for generating accurate map products: Normalized Difference Vegetation Index (NDVI), Normalized Burn Ratio (NBR), and Random Forest (RF) and Support Vector Machine (SVM) supervised classification. We used 52 Composite Burn Index (CBI) plots from the Shovel Creek burn scar and 28 from the Nugget Creek burn scar for training classifiers and product validation. For the Shovel Creek burn scar, the RF and SVM machine learning (ML) classification methods outperformed the traditional spectral indices that use linear regression to separate burn severity classes (RF and SVM accuracy, 83.33%, versus NBR accuracy, 73.08%). However, for the Nugget Creek burn scar, the NDVI product (accuracy: 96%) outperformed the other indices and ML classifiers. In this study, we demonstrated that when sufficient ground truth data is available, the ML classifiers can be very effective for reliable mapping of burn severity in the Alaskan boreal forest. Since the performance of ML classifiers are dependent on the quantity of ground truth data, when sufficient ground truth data is available, the ML classification methods would be better at assessing burn severity, whereas with limited ground truth data the traditional spectral indices would be better suited. We also looked at the relationship between burn severity, fuel type, and topography (aspect and slope) and found that the relationship is site-dependent. 
    more » « less
  • Citation Formats
  • MLA
  • APA
  • Chicago
  • BibTeX
  • Save / Share this Record
  • Send to Email