Gust fronts, the outflow boundaries of thunderstorms, are often associated with an abrupt change of wind speed and direction that can significantly change the behavior and spread of wildfires. Anticipating such changes is essential to ensure the safety of firefighters and improve operational decision making. However, existing radar-based techniques are primarily limited to detection and basic extrapolation of boundary motion. This study develops a data-driven framework for predicting the motion and location of the gust front using observations from polarimetric weather radar. Each radar-detected gust front at a given time is represented as a fitted curve and subsequently vectorized into normalized anchor points. By associating those points from adjacent scans, the motion of the gust front is estimated. Two velocity models, point-wise and mean motion, are developed and evaluated. In addition, their results are fused to produce probabilistic forecasts of the position of the gust front. The predicted gust front effectively captures its most likely location and the uncertainty associated with transient or interacting fronts. The results of multiple cases demonstrate significant prediction skills up to approximately 30 minutes of lead time. Although the current implementation relies on manually labeled data, integration with the Neuro-Fuzzy Gust Front Detection Algorithm (NFGDA) is underway to enable fully automated real-time operation. The framework represents a first step toward operational short-term gust front forecasting to improve the situational awareness of wildfire hazard.

Our objective in the present study is to provide basic insights into the coupling between external-gas and solid biomass vegetation processes that control the dynamics of flame spread in wildland fire problems. We focus on a modeling approach that resolves processes occurring at vegetation and flame scales, i.e., the formation of flammable vapors due to the thermal degradation of the solid biomass, the subsequent combustion in ambient air, the thermal feedback to the biomass through radiative and convective heat transfer, and the possible transition from flaming combustion (taking place outside of the solid biomass) to smoldering combustion (taking place inside the solid biomass). The capability uses a multiphase combustion framework and treats external-gas processes through a Large Eddy Simulation solver and solid biomass processes through a discrete particle model. The discrete particle model adopts a one-dimensional porous medium formulation, includes descriptions of drying, thermal pyrolysis, oxidative pyrolysis, and char oxidation, as well as a description of the external-gas-to-solid-biomass diffusion of oxygen mass; the discrete particle model thereby provides a treatment of in-depth oxidative processes and allows the simulation of smoldering combustion. The modeling capability is applied to the simulation of fire spread across a surrogate biomass vegetation bed corresponding to a discrete array of cylindrical-shaped, vertically-oriented, pine wood sticks, characterized by a monomodal size distribution, in horizontal flat terrain and under wind-aided conditions. The numerical results demonstrate that the model can simulate successful flaming-to-smoldering transition followed by complete biomass consumption. Novelty and Significance statement: • A new computational model is proposed to simulate wildland fire behavior at high levels of resolution capable of capturing vegetation- and flame-scale phenomena. • Solid biomass vegetation processes are treated using a discrete particle model that features a porous medium formulation and includes a description of in-depth oxygen mass diffusion, and exothermic oxidative pyrolysis and char oxidation reactions. • The computational model is shown to be capable of simulating the transition from flaming to smoldering combustion, often observed in wildland fire spread problems.

Our objective in the present study is to provide fundamental insights into the main factors that control the thermal degradation processes occurring inside the biomass vegetation during wildland fire spread. These processes determine the formation of flammable volatiles through pyrolysis reactions and the intensity of the internal heat release through char oxidation reactions; both aspects are central to flaming and smoldering combustion. We adopt here a simplified framework in which the biomass vegetation is viewed as a population of discrete elongated particles that are made of pine wood, are cylindrical-shaped and feature different diameters. The exposure conditions experienced by these particles in an assumed fire are characterized by a given intensity of the external thermal load, a given duration during which this load is applied, and a given external flow velocity that controls the rates of convective heat transfer and oxygen mass transfer. The response of the biomass particles is studied using a one-dimensional computational model, called the particle burning rate (PBR) model. The PBR model adopts a porous medium formulation, and provides a description of drying, (thermal and oxidative) pyrolysis, and char oxidation. We focus in the present study on the conditions required for complete "fuel consumption", i.e., for complete particle degradation from virgin solid to ash. It is found that in thin biomass particles, complete biomass degradation requires sufficiently high intensities of the external thermal load and sufficiently low external flow velocities; high flow velocities correspond to excessive cooling effects. Similarly, in thick biomass particles, complete biomass degradation requires sufficiently high intensities of the external thermal load, sufficiently long application times of the thermal load, and sufficiently high external flow velocities; low flow velocities correspond to insufficient oxygenation of the (exothermic) oxidative pyrolysis and char oxidation reactions.

Background Terrain-induced turbulence is dangerous for smokejumpers parachuting into complex terrain and results in numerous serious accidents annually.Aims We quantify wind modelling system WindNinja's ability to reproduce terrain-induced effects on the mean wind speed and turbulence in complex terrain. We assess WindNinja's suitability for use in identifying safe jump spots during smokejumper operations in complex terrain.Methods We evaluate the model's ability to reproduce mean wind speed, mean wind direction and turbulence kinetic energy (TKE) measured by sonic anemometers and lidar scanners over a ridge-valley-ridge system collected under near-neutral atmospheric conditions during the Perdig & atilde;o field campaign. We conduct a WindNinja simulation to examine the wind and turbulence conditions during the 2021 Eicks Fire smokejumper accident.Key results WindNinja can reproduce both mean wind speed and turbulence characteristics induced by the terrain. WindNinja revealed critical turbulence information that could have been useful to smokejumpers during the Eicks Fire jumping operation.Conclusions WindNinja's ability to reproduce key features in the mean wind speed and turbulence fields induced by the terrain make it suitable for use as an aid in identifying safe jump spots in complex terrain.Implications Findings from this work will reduce parachute accidents and increase the safety of aerial firefighter operations.

The authors are a team of fire whirl researchers who have been actively studying whirls and large-scale wildland fires by directly observing them through fire-fighting efforts and applying theory, scale modeling, and numerical simulations in fire research. This multidisciplinary research-background team previously conducted scale model experiments to reconstruct hazardous large-scale fires in the laboratory, then conducted numerical simulations and developed fundamental theories to translate these findings into a basic understanding of combustion science and fluid dynamics. This article, a mix of reviews of the state of art experiments, theories, numerical modeling and artificial intelligence, and two case studies, is intended to address some safety concerns and raise awareness of large-scale fire whirls and forest fires with knowledge of thermodynamics, chemical kinetics, fluid dynamics, design, and practical fire-fighting experience, offering gaps that should be filled and future research to be conducted in each field, and crucial new observations and insights on large-scale fire incidents. We believe, this timely topic is of interest not only to fire research community but also to general readers, as the frequency and intensity of large-scale forest fires and fire whirls have increased, possibly due to the continuing global warming trend and human-induced changes in fuels. Each section and case study was written by one or two individual researchers based on their field of expertise which allows them to critically review progress made in their section of large-scale fire-whirls and forest-fires. Crucial observations and insights on the historical Great-Kanto-Earthquake-generated Hifukusho-Ato Fire-whirl (HAFW) and the slow rotations observed during recent forest firefighting efforts are presented. The first case study occurred in downtown Tokyo on 1 September 1923, as a result of the Great-Kanto-Earthquake, which claimed over 38,000 deaths within 15 min. The second case study discusses large-scale slow rotations observed during recent forest fires, which might had been responsible for the injuries and deaths of experienced firefighters.

Our general objective in the present study is to develop tools to better describe the coupling between solid phase and gas phase processes that control the dynamics of flame spread in wildland fire problems. We focus on a modelling approach that resolves processes occurring at flame scales, i.e., the formation of flammable vapors from the biomass vegetation due to pyrolysis, the subsequent combustion of these fuel vapors with ambient air, the establishment of a turbulent flow because of heat release and buoyant acceleration, and the thermal feedback to the solid biomass through radiative and convective heat transfer. The modelling capability is based on a general-purpose Computational Fluid Dynamics (CFD) library called OpenFOAM and an in-house Lagrangian particle model that treats drying, thermal pyrolysis, oxidative pyrolysis and char oxidation using a one-dimensional porous medium formulation that allows descriptions of thermal degradation processes occurring during both flaming and smoldering combustion. The modelling capability is calibrated for pine wood and is first applied to simulations of fire spread across a surrogate vegetation bed corresponding to thin, monodisperse, cylindrical-shaped sticks of pine wood with prescribed particle and environmental properties (i.e., bed height, surface-to-volume ratio, packing ratio, moisture content, and wind velocity). While the model can be used in sloped terrain, the present simulations are limited to a flat ground surface. The current emphasis is on determining threshold conditions for successful spread, differences between the plume-dominated and wind-driven flame regimes, possible transitions to a steady or time-dependent flame structure, and differences in the relative weights of the flaming and smoldering regions.

We demonstrate the use of a deep learning (DL) approach for representing the behavior of a high-resolution physics-based wildland fire spread model. The ultimate objective is being able to efficiently use the DL model for intensive simulations of large fires while retaining fidelity to the fine-scale physical processes. We begin with a fire model that reduces the spatial domain of the fire spread problem to one dimension (1D). The 1D model explicitly resolves cm-scale fuel variations, heat transfer and heating/drying dynamics of individual fuel particles and burning behavior of the bed. We then ran the fire model for 78,125 factorial combinations of fuel, weather, and topographic conditions as training data for the DL algorithm. The results of the DL analysis show overall agreement of 96% of the variation in fire behavior as represented by steady state rate of spread, flame length and flame zone depth. Exceptions to the DL regression indicate areas where more work is required in refining the resolution in training cases and use of advanced methods of embedding the fire model inside the DL algorithm loop.

Live foliage for some tree and shrub species can support flaming fire spread at much higher moisture content than dead fuel materials. However, the role of live fuels in forest fires has been controversial in the past decades. Although ignition and spread statistical data for live and dead fuels exist in the literature, a clear understanding of the fundamental difference in the burning behavior is missing. To illuminate the role of live fuel on forest fire spreading, a laboratory ignition experiment was designed to examine the burning behavior of live Norway spruce needles. A Schlieren-Infrared combined measurement apparatus was developed with a spatial resolution of 0.75 mm and a time resolution of 0.0025 s, to visualize/measure the ignition behavior of live fuels. Schlieren and IR images revealed that the ejection of live fuel volatiles can alter the flame direction and induce previously unaccounted heating of the nearby fuel. Depending on the conditions, these interferences could heat and modify the heat flux received by the adjacent fuels. To analyze each of these outcomes, a scaling analysis using the law approach was performed. First, theoretical equations were developed and validated against a set of previously published experimental data. After the characteristic equations were verified, we used them to assess the volatile ejection phenomenon. We found that adjacent fuels were preheated by hot volatiles ejected from the heated live needle, and direct flame contact ignited the adjacent fuels. Our IR experiments confirmed the outcomes of the scaling analysis. The rapid ejection of volatiles was also found to propel burning needles far from the burning branch, resulting in micro-spotting.

Wildland fires have an irreplaceable role in sustaining many of our forests, shrublands and grasslands. They can be used as controlled burns or occur as free-burning wildfires, and can sometimes be dangerous and destructive to fauna, human communities and natural resources. Through scientific understanding of their behaviour, we can develop the tools to reliably use and manage fires across landscapes in ways that are compatible with the constraints of modern society while benefiting the ecosystems. The science of wildland fire is incomplete, however. Even the simplest fire behaviours – how fast they spread, how long they burn and how large they get – arise from a dynamical system of physical processes interacting in unexplored ways with heterogeneous biological, ecological and meteorological factors across many scales of time and space. The physics of heat transfer, combustion and ignition, for example, operate in all fires at millimetre and millisecond scales but wildfires can become conflagrations that burn for months and exceed millions of hectares. Wildland Fire Behaviour: Dynamics, Principles and Processes examines what is known and unknown about wildfire behaviours. The authors introduce fire as a dynamical system along with traditional steady-state concepts. They then break down the system into its primary physical components, describe how they depend upon environmental factors, and explore system dynamics by constructing and exercising a nonlinear model. The limits of modelling and knowledge are discussed throughout but emphasised by review of large fire behaviours. Advancing knowledge of fire behaviours will require a multidisciplinary approach and rely on quality measurements from experimental research, as covered in the final chapters.

We present novel in‐field vegetation fire observations and the analyses using brightness temperatures recorded by longwave infrared camera and thermal image velocimetry. The brightness temperatures from a wind‐driven stubble wheat fire were obtained with a 60 frames per second (fps) video acquisition. Multi‐level sonic anemometers on a 10m in‐fire tower were used for measurements of turbulent velocity and air temperatures, while fuel level air and flame temperatures were collected by an array of thermocouples (TCs). The camera's image pixel resolution was adequate to discern thermal structures in accordance with the in‐fire TC measurements and its spacing distances. The in‐situ and remotely measured flaming zone dynamics were derived using two different methodologies, Thermal Image Velocimetry (TIV) and Image Segmentation (IS). The results highlight spatial and spectral information of coherent turbulent and mean velocity structures. The power spectra decomposition of the TIV showed similar spectral characteristics to the sonic speed measurements during the fire passage under the tower with a similar inertial subrange slope. This result reveals plausible evidence of interaction between the flaming zone and wind turbulence for a prescribed rapidly moving stubble wheat fire. This research presents a new field measurement methodology for understanding fire‐atmospheric interactions between the flaming zone and the immediate overlying atmospheric turbulent boundary layer.

An open source computational fluid dynamics (CFD) solver has been incorporated into the WindNinja modeling framework. WindNinja is widely used by wildland fire managers, as well as researchers and practitioners in other fields, such as wind energy, wind erosion, and search and rescue. Here, we describe the CFD solver and evaluate its performance against the WindNinja conservation of mass (COM) solver, and previously published large-eddy simulations (LES), for three field campaigns with varying terrain complexity: Askervein Hill, Bolund Hill, and Big Southern Butte. We also compare the effects of two model settings in the CFD solver, namely the discretization scheme used for the advection term of the momentum equation and the turbulence model, and provide guidance on model sensitivity to these settings. Additionally, we investigate the computational mesh and difficulties regarding terrain representation. Two important findings from this work are: (1) CFD solver predictions are significantly better than COM solver predictions at windward and lee side observation locations, but no difference was found in predicted speed-up at ridgetop locations between the two solvers, and (2) the choice of discretization scheme for advection has a significantly larger effect on the simulated winds than the choice of turbulence model.

With emerging research on the dynamics of extreme fire behavior, it is increasingly important for wind models, used in operational fire prediction, to accurately capture areas of complex flow across rugged terrain. Additionally, the emergence of ensemble and stochastic modeling frameworks has led to the discussion of uncertainty in fire prediction. To capture the uncertainty of modeled fire outputs, it is necessary to recast uncertain inputs in probabilistic terms. WindNinja is the diagnostic wind model currently being applied within a number of operational fire prediction frameworks across the world. For computational efficiency, allowing for real-time or faster than real-time prediction, the physical equations governing wind flow across a complex terrain are often simplified. The model has a number of well documented limitations, for instance, it is known to perform poorly on leeward slopes. First, this study is aimed at understanding these limitations in a probabilistic context, by comparing individual deterministic predictions to observed distributions of wind direction. Secondly, a novel application of the deterministic WindNinja model is presented in this study which is shown to enable prediction of wind direction distributions that capture some of the variability of complex wind flow. Recasting wind fields in terms of probability distributions enables a better understanding of variability across the landscape, and provides the probabilistic information required to capture uncertainty through ensemble or stochastic fire modeling. The comparisons detailed in this study indicate the potential for WindNinja to predict multi-modal wind direction distributions that represent complex wind behaviors, including re-circulation regions on leeward slopes. However, the limitations of using deterministic models within probabilistic frameworks are also highlighted. To enhance fire prediction and to better understand uncertainty, it is recommended that statistical approaches also be developed to complement existing physics-based deterministic wind models.

RESEARCHERS ARE GETTING CLOSER TO PREDICTING WHEN AND WHERE THESE LETHAL VORTICS WILL APPEAR

This work investigates gap winds in a steep, deep river canyon prone to wildland fire. The driving mechanisms and the potential for forecasting the gap winds are investigated. The onset and strength of the gap winds are found to be correlated to the formation of an along-gap pressure gradient linked to periodic development of a thermal trough in the Pacific Northwest, USA. Numerical simulations are performed using a reanalysis dataset to investigate the ability of numerical weather prediction (NWP) to simulate the observed gap wind events, including the timing and flow characteristics within the canyon. The effects of model horizontal grid spacing and terrain representation are considered. The reanalysis simulations suggest that horizontal grid spacings used in operational NWP could be sufficient for simulating the gap flow events given the regional-scale depression in which the Salmon River Canyon is situated. The strength of the events, however, is under-predicted due, at least in part, to terrain smoothing in the model. Routine NWP, however, is found to have mixed results in terms of forecasting the gap wind events, primarily due to problems in simulating the regional sea level pressure system correctly.

Wildland fire managers in the United States currently utilize the gridded forecasts from the National Digital Forecast Database (NDFD) to make fire behavior predictions across complex landscapes during large wildfires. However, little is known about the NDFDs performance in remote locations with complex topography for weather variables important for fire behavior prediction, including air temperature, relative humidity, and wind speed. In this study NDFD forecasts for calendar year 2015 were evaluated in fire-prone locations across the conterminous United States during periods with the potential for active fire spread using the model performance statistics of root-mean-square error (RMSE), mean fractional bias (MFB), and mean bias error (MBE). Results indicated that NDFD forecasts of air temperature and relative humidity performed well with RMSEs of about 2 degrees C and 10%-11%, respectively. However, wind speed was increasingly under-predicted when observed wind speeds exceeded about 4 ms(-1), with MFB and MBE values of approximately -15% and -0.5 ms(-1), respectively. The importance of accurate wind speed forecasts in terms of fire behavior prediction was confirmed, and the forecast accuracies needed to achieve "good'' surface head fire rate-of-spread predictions were estimated as +/- 20%-30% of the observed wind speed. Weather station location, the specific forecast office, and terrain complexity had the largest impacts on wind speed forecast error, although the relatively low variance explained by the model (similar to 37%) suggests that other variables are likely to be important. Based on these results it is suggested that wildland fire managers should use caution when utilizing the NDFD wind speed forecasts if high wind speed events are anticipated.