Elijah Orland | Search USGS

USGS2025

Near real-time indicators of burn severity in the western U.S. from active fire tracking

Background Timely information on wildfire burn severity is critical to assess and mitigate potential post-fire impacts on soils, vegetation, and hillslope stability. Tracking individual fire spread and intensity using satellite active fire data provides a pathway to near real-time (NRT) information. Here, we generated a large database ( n = 2177) of wildfire events in the western United States (U.S.) between 2012 and 2021 using active fire detections from the Visible Infrared Imaging Radiometer Suite (VIIRS) sensor on the Suomi National Polar-orbiting Partnership (SNPP) satellite and the Fire Events Data Suite (FEDS) algorithm to track large fire growth every 12 h. We integrated fire tracking data with final fire perimeters and burn severity data from the Monitoring Trends in Burn Severity (MTBS) program to evaluate the relationship between burn severity and fire behavior metrics derived from the fire tracking approach, including the rate of fire spread and average fire radiative power (FRP) of fire detections for each 12-h growth increment. Results When stratified by vegetation type, FRP and rate of spread metrics were positively correlated with classified burn severity for each 12-h growth increment, highlighting the potential to rapidly identify areas of high and low severity burning. In forests, integrated measures of FRP over the fire lifetime captured persistent flaming and smoldering that compensated for initial differences between AM (01:30) and PM (13:30) fire detections. Predictive modeling of these relationships based on multiple fire behavior indicators and vegetation type from the LANDFIRE program yielded an accuracy of 78% for the separation of unburned/low and moderate/high burn severity classes. Conclusions These results demonstrate the ability to capture within-fire differences in burn severity using NRT indicators from fire tracking to assist with emergency management and disaster preparedness for post-fire hazards, such as landslides, debris flows, or changes in stream flow and water quality. As VIIRS data are available within minutes of each satellite overpass in the U.S., rapid estimates of burn severity based on fire tracking can be made days or weeks before a large wildfire is fully contained.

western United States↗

USGS2020

Deep Learning as a tool to forecast hydrologic response for landslide-prone hillslopes

Empirical thresholds for landslide warning systems have benefitted from the incorporation of soil‐hydrologic monitoring data, but the mechanistic basis for their predictive capabilities is limited. Although physically based hydrologic models can accurately simulate changes in soil moisture and pore pressure that promote landslides, their utility is restricted by high computational costs and nonunique parameterization issues. We construct a deep learning model using soil moisture, pore pressure, and rainfall monitoring data acquired from landslide‐prone hillslopes in Oregon, USA, to predict the timing and magnitude of hydrologic response at multiple soil depths for 36‐hr intervals. We find that observation records as short as 6 months are sufficient for accurate predictions, and our model captures hydrologic response for high‐intensity rainfall events even when those storm types are excluded from model training. We conclude that machine learning can provide an accurate and computationally efficient alternative to empirical methods or physical modeling for landslide hazard warning.

Oregon↗