Fire Model Inputs

These inputs configure the fire model described in Fire Model. Every key is read once at startup from the erf.fire prefix (the three plotfile keys from erf). Defaults are those of the parameter struct in FireEngine/ERF_FireParams.H; the reference deck Exec/CanonicalTests/Fire/inputs_fire_master_reference lists every key with a comment but is a configured case, so its values are not all defaults. Keys marked “read but not consumed” are accepted for compatibility and have no effect at present.

Master control

Parameter	Definition	Acceptable Values	Default
erf.fire.enable	Enable the fire model	Boolean	false
erf.fire.engine	Fire engine behind the shared fire-atmosphere coupling (Source/FireCoupling); “flare” is FLARE, this repository’s own model (FireEngine/). Any other value aborts	“flare”	“flare”
erf.fire.fire_debug	Print per-step fire diagnostics to stdout	Boolean	false
erf.fire.grid_ratio	Fire grid refinement factor in x and y; every box length on the fire grid’s level must divide by it	Integer > 0	5
erf.fire.anchor_level	AMR level whose grid the fire grid refines; -1 is the finest level at start-up. Above level 0 the level must keep its boxes (erf.regrid_int = -1), cover one rectangle of full-height columns and run without the dust layer; a level below the finest prints a warning	-1, or 0 to the finest level	-1

Fuel and moisture

Parameter	Definition	Acceptable Values	Default
erf.fire.fuel_model_id	Anderson FBFM13 fuel model used everywhere without a fuel map, and for WAF and coefficients with one	1-13; 101-204 with fuel_set = scott_burgan40; 1000-1015 custom codes; others abort	1
erf.fire.moisture_1hr	1-hour dead fuel moisture [fraction]	Real 0-1	0.08
erf.fire.moisture_10hr	10-hour dead fuel moisture [fraction]	Real 0-1	0.08
erf.fire.moisture_100hr	100-hour dead fuel moisture [fraction]	Real 0-1	0.10
erf.fire.moisture_live	Live herbaceous and woody moisture [fraction]; the live classes start here (see moisture_live_model)	Real	0.60
erf.fire.moisture_live_model	Live classes with moisture_dynamic: the legacy dead-fuel update (a value above 0.40 drops to 0.40 on the first step) or held at moisture_live	“legacy”, “fixed”	“legacy”
erf.fire.moisture_dynamic	Advance the dead classes each step with the Nelson time- lag model	Boolean	true
erf.fire.precip_rate_mm_hr	Uniform precipitation rate wetting the dead fuel [mm/hr]	Real >= 0	0.0
erf.fire.precip_source	Rain wetting the dead classes: precip_rate_mm_hr everywhere, or each column’s change per step of the microphysics surface precipitation accumulation (needs a scheme with precipitation, moisture_dynamic, and precip_rate_mm_hr = 0)	“uniform”, “atmosphere”	“uniform”
erf.fire.moisture_model	Dead-class update: one time constant per class, or radial diffusion in a cylindrical stick (Nelson 2000 framework)	“timelag”, “stick”	“timelag”
erf.fire.emc_model	Equilibrium moisture curves of the dead-class update: the legacy quartics, or Van Wagner and Pickett (1985) drying and wetting EMC (temperature dependent, as in WRF-SFIRE)	“legacy”, “van_wagner”	“legacy”
erf.fire.stick.n_shells	Radial shells per stick	Integer 1-16	6
erf.fire.stick.radius_cm	Stick radii of the 1-h, 10-h and 100-h classes [cm]	3 Reals > 0	0.15 0.635 2.5
erf.fire.stick.rain_surface_moisture	Surface moisture held while it rains	Real 0.01-0.40	0.35
erf.fire.stick.diffusivity_scale	Multiplier on the lag-calibrated diffusivity	Real > 0	1.0
erf.fire.fuel_map.file	Spatial fuel map file; empty means uniform fuel. Rows of an ESRI ASCII map run north first; one code per fire cell, so the map must have the fire grid’s size	String	“”
erf.fire.fuel_map.format	Fuel map format	“ascii”, “lcp”	“ascii”
erf.fire.fuel_map.blending_fraction	ROS blending toward neighbours of a different fuel code	Real 0-1	0.0
erf.fire.fuel_map.nonburnable_codes	Fuel codes that are non-burnable; empty keeps the fall- through to fuel model 1	Integers	none
erf.fire.fuel_map.fuel_set	Fuel model set: the Anderson 13, or the Scott and Burgan 40 (LANDFIRE FBFM40 codes 101-204, 91-99 non-burnable) alongside them	“anderson13”, “scott_burgan40”	“anderson13”
erf.fire.fuel_map.sb40_crosswalk	Translate the Scott-Burgan codes to the Anderson 13 at load (the CFBM route) instead of using them natively	Boolean	false
erf.fire.fuel_map.load_from_map	Start each cell with its own model’s fuel load instead of the uniform model’s	Boolean	false
erf.fire.rothermel_per_fuel	Evaluate Rothermel coefficients per cell from the fuel map instead of fuel_model_id, on the isotropic and the level-set paths alike	Boolean	false
erf.fire.custom_fuel.codes	Fuel codes described in the deck instead of a published set; each must be in 1000-1015	Ints	none
erf.fire.custom_fuel.N.name	Label for code N in the start-up summary	String	custom_N
erf.fire.custom_fuel.N.w_1h_kg_m2	1-h dead fuel load of code N [kg/m2]; at least one load must be > 0	Real in [0, 500]	required
erf.fire.custom_fuel.N.w_10h_kg_m2	10-h dead fuel load [kg/m2]	Real in [0, 500]	0.0
erf.fire.custom_fuel.N.w_100h_kg_m2	100-h dead fuel load [kg/m2]	Real in [0, 500]	0.0
erf.fire.custom_fuel.N.w_live_herb_kg_m2	Live herbaceous load [kg/m2]; no curing transfer is applied	Real in [0, 500]	0.0
erf.fire.custom_fuel.N.w_live_woody_kg_m2	Live woody load [kg/m2]	Real in [0, 500]	0.0
erf.fire.custom_fuel.N.sav_1h_1_m	1-h dead surface-area-to-volume ratio [1/m]	Real in [10, 12000]	required
erf.fire.custom_fuel.N.sav_live_herb_1_m	Live herbaceous surface-area-to-volume ratio [1/m]; kept only where that load is > 0	Real in [10, 12000]	sav_1h_1_m
erf.fire.custom_fuel.N.sav_live_woody_1_m	Live woody surface-area-to-volume ratio [1/m]; kept only where that load is > 0	Real in [10, 12000]	sav_1h_1_m
erf.fire.custom_fuel.N.depth_m	Fuel bed depth [m]; the Balbi models return zero spread at or below 0.01 m	Real in (0.01, 50]	required
erf.fire.custom_fuel.N.moisture_ext	Moisture of extinction [fraction]	Real in [0.01, 2]	required
erf.fire.custom_fuel.N.heat_content_J_kg	Heat content [J/kg]	Real in [5e6, 5e7]	required
erf.fire.custom_fuel.N.density_kg_m3	Fuel particle density [kg/m3]	Real in [50, 2000]	512.0
erf.fire.custom_fuel.N.burnout_time_s	Flaming burn time [s]; required for code N when erf.fire.burnout_model = sfire	Real in [0, 1e5]	0.0
erf.fire.firebreak.N.type	Firebreak N (N = 0..9, consecutive): shape	“rect”, “circle”	none
erf.fire.firebreak.N.x_lo/y_lo/x_hi/y_hi	Rectangle firebreak bounds [m]	Reals	0.0
erf.fire.firebreak.N.cx/cy/radius	Circular firebreak centre and radius [m]	Reals	0.0
erf.fire.firebreak.use_mask	Hold firebreak cells in the non-burnable mask as well as stamping the sentinel; false (the form before 2026-10) warns when firebreaks are configured, since the stamp can erode	Boolean	true
erf.fire.structures.enable	Treat cells of the structure heightmap above min_height as non-burnable	Boolean	false
erf.fire.structures.file	Building heightmap in ERF terrain text format	String	hybrid.structure_file, then erf.buildings_fil e_name
erf.fire.structures.min_height	Sampled height above which a cell is a structure [m]	Real >= 0	0.5
erf.fire.structures.wind_open_columns	Renormalise the bilinear wind weights over the columns open at the wind height; needs structures.enable	Boolean	false
erf.fire.structures.ignition.enable	Structures ignite from the exposure accumulators, burn, radiate and launch brands (Structure Ignition and House-to-House Spread) ; needs exposure.enable	Boolean	false
erf.fire.structures.ignition.heat_load_J_m2	Largest accumulated heat load in the wall band at which a structure ignites [J/m2]	Real; <= 0 turns the criterion off	6.0e6
erf.fire.structures.ignition.ember_count	Brands landed on the footprint at which a structure ignites	Integer; <= 0 turns the criterion off	50
erf.fire.structures.ignition.intensity_kW_m	Fireline intensity in the wall band above which time counts toward residence_s [kW/m]	Real; <= 0 turns the criterion off	1000.0
erf.fire.structures.ignition.residence_s	Time above intensity_kW_m at which a structure ignites [s]	Real >= 0	60.0
erf.fire.structures.ignition.peak_flux_W_m2	Peak heat release of a burning structure per unit footprint area [W/m2]	Real > 0	2.5e5
erf.fire.structures.ignition.fuel_load_J_m2	Heat a structure releases over its whole burn per unit footprint area [J/m2]	Real > 0	7.8e8
erf.fire.structures.ignition.growth_time_s	Time from ignition to the peak flux (t-squared growth) [s]; 0 starts at the peak	Real >= 0; peak_flux_W_m2 * growth_time_s / 3 <= 0.7 * fuel_load_J_m2	600.0
erf.fire.structures.ignition.rad_fraction	Share of the release radiated onto the cells around the structure	Real in [0, 1]	0.3
erf.fire.structures.ignition.rad_radius_m	Distance within which a burning footprint cell’s radiation is added to the heat load [m]	Real >= 0; 0 turns the radiation off	100.0

Ignition

Parameter	Definition	Acceptable Values	Default
erf.fire.ignition_x	Ignition disc centre x [m]	Real	0.0
erf.fire.ignition_y	Ignition disc centre y [m]	Real	0.0
erf.fire.ignition_r	Ignition disc radius [m]; 0 = no disc (schedule, polygon, line or threshold ignites instead)	Real >= 0	20.0
erf.fire.ignition.schedule_file	CSV of timed ignition events; empty disables	String	“”
erf.fire.ignition.polygon_file	One or more CSVs of perimeter vertices, each one polygon or polyline ignition of the deck’s polygon_type and polyline_width; every file is stamped, so several files are several fires that merge as they grow	List of Strings	“”
erf.fire.ignition.polygon_type	Closed polygon or open line fire	“polygon”, “polyline”	“polygon”
erf.fire.ignition.polygon_time	Time [s] at which the polygon is stamped, so the atmosphere spins up first; 0 stamps it at initialisation	Real >= 0	0.0
erf.fire.ignition.polygon_interior_ros	Spread rate [m/s] behind the interior: arrival time polygon_time - d/R and fuel exp(-d/(R tau)); 0 = off	Real >= 0	0.0
erf.fire.ignition.polygon_interior_tau	Burnout e-folding time [s] of the interior; 0 = dx_fire/R	Real >= 0	0.0
erf.fire.ignition.polyline_width	Half-width of the polyline ignition zone [m]	Real > 0	10.0
erf.fire.ignition.threshold_enable	Ignite every burnable, unburned fire cell whose k = 0 air temperature (theta times the Exner function) exceeds threshold_temp	true, false	false
erf.fire.ignition.threshold_temp	Ignition temperature [K]	Real > 0	573.15
erf.fire.ignition.threshold_radius	Radius [m] of the disc stamped around each hot cell; 0 = one fire cell	Real >= 0	0.0
erf.fire.ignition.threshold_start_time	No threshold ignition before this time [s]	Real >= 0	0.0

Wind

Parameter	Definition	Acceptable Values	Default
erf.fire.wind_ref_ht	Height above local ground at which the atmospheric wind is sampled [m]	Real > 0	6.1
erf.fire.prescribed_wind	Bypass atmospheric wind interpolation with a constant reference wind, for controlled idealized-wind test cases; exact everywhere only with use_waf = false and use_terrain_wind = false	Boolean	false
erf.fire.prescribed_wind_x	Constant reference wind, x, when prescribed_wind [m/s]	Real	0.0
erf.fire.prescribed_wind_y	Constant reference wind, y, when prescribed_wind [m/s]	Real	0.0
erf.fire.use_per_fuel_wind_ht	Sample the wind at the per-fuel fcwh height (6.096 m for every fuel) instead of wind_ref_ht, with or without a fuel map; every reader of the wind’s height follows it	Boolean	false
erf.fire.wind_sample_ht	Sample the wind at this height above ground and bring it to wind_ref_ht (6.096 m with use_per_fuel_wind_ht) with a log profile (CFBM); 0 = off	Real > wind_ref_ht (6.096 with use_per_fuel_wind_ht), or 0	0.0
erf.fire.wind_sample_z0	Roughness length [m] of every log profile the fire reads: the two-height resampling, the wind below the first cell centre and the ember drift; below the height of the reference wind, which is 6.096 m with use_per_fuel_wind_ht	0 < Real < wind_ref_ht, or < 6.096 with use_per_fuel_wind_ht	0.1
erf.fire.wind_below_first_cell	Wind at a target below the first cell centre: “clamp” takes the centre’s wind unchanged; “log” scales it by the neutral log profile of wind_sample_z0 (recommended for a first cell above wind_ref_ht; 12 % lower under a 20 m first cell)	“clamp”, “log”	“clamp”
erf.fire.wind_interp	Horizontal mapping of atmospheric columns onto fire cells	“bilinear”, “nearest”	“bilinear”
erf.fire.use_waf	Apply the wind adjustment factor to the reference wind	Boolean	true
erf.fire.waf_formula	Wind adjustment factor formula, from the fuel bed depth: the domain model’s on a uniform fuel, each cell’s own wherever a spatial fuel map is read	“andrews”, “behaviorplus”	“andrews”
erf.fire.use_wind_limit	Bound the midflame wind by Rothermel’s (1972, eq. 87) limit U <= 0.9 I_R (Rothermel and BEHAVE kernels); false: no limit	Boolean	true
erf.fire.wind_limit	Which limit use_wind_limit applies: “rothermel” is 0.9 I_R ft/min; “fuel_class” this code’s earlier rule of 300 ft/min above a SAV of 1000 1/ft and 500 below (the form before 2026-10)	“rothermel”, “fuel_class”	“rothermel”
erf.fire.use_terrain_wind	Apply the terrain wind factors (ridge, shelter, valley, deflection); default false since 2026-10	Boolean	false
erf.fire.k_ridge	Ridge speed-up factor (needs use_terrain_wind = true)	Real > 0	1.5
erf.fire.k_shelter	Lee sheltering factor	Real > 0	0.6
erf.fire.k_valley	Valley channelling factor	Real > 0	0.8
erf.fire.k_deflect	Wind deflection factor	Real > 0	0.3
erf.fire.terrain_file_name	Fine-grid terrain file interpolated onto the fire grid for slopes	String	“”

Rate of spread

Parameter	Definition	Acceptable Values	Default
erf.fire.ros_model	Rate-of-spread model; “grass_simple” is the fit that “cheney_gould” selected before 2026-10	“rothermel”, “behave”, “macarthur”, “cheney_gould”, “grass_simple”, “fbp”, “balbi”, “hybrid”, “prescribed”	“rothermel”
erf.fire.rothermel_cell_moisture	With moisture_dynamic, rebuild the Rothermel coefficients from each cell’s own dead moistures; false takes the domain-mean moistures everywhere (the form before 2026-10)	Boolean	true
erf.fire.macarthur.ros_max	Cap on the McArthur rate [m/s] (WRF-Fire’s 6 m/s); 0 removes it. The exponent is bounded at 40 in every case, so an uncapped rate stops growing at a 47.5 m/s midflame wind (4e16 m/s, finite, long past physical)	Real >= 0	6.0
erf.fire.reaction_velocity_formula	Rothermel’s reaction-velocity exponent A: Albini’s (1976) revision, A = 133 sigma^-0.7913 (ERF’s form so far), or Rothermel’s (1972) Eq. 39, A = 1/(4.774 sigma^0.1 - 7.27), the form WRF-Fire uses. Only A changes, and through I_R the default wind limit 0.9 I_R (wind_limit = rothermel). Read by every run that evaluates Rothermel (ros_model = rothermel, or a hybrid with a rothermel member); BEHAVE keeps its own Albini form; other models ignore it with a warning	“albini”, “rothermel”	“albini”
erf.fire.wrf_bmst_compat	WRF-Fire comparison: deflate the fuel load fed to the Rothermel coefficients by 1 - M/(1+M), M the dead-load- weighted dead moisture, as WRF-Fire does to its load. Not a physical correction (ERF’s loads are oven-dry) and not full WRF-Fire parity (bed SAV, net load, heat content and wind/rate caps differ); fuel consumption and heat release keep the full load. Through I_R it lowers the default wind limit 0.9 I_R too. Same readers as the row above	Boolean	false
erf.fire.directional_ros	Evaluate the rate along the front normal on the level-set path; false spreads the head rate in every direction	Boolean	true
erf.fire.directional_shape	Form of that rate: the model along the front normal, or the support function of the ellipse with the model’s head, back and flank rates, which keeps a point fire’s head at the head rate; not for balbi or hybrid	“projection”, “ellipse”	“projection”
erf.fire.directional_wind_coupling	With directional_shape = “projection”: exponentiate the wind/slope factor from the projected wind (Rothermel’s own form), or from the raw wind and scale by the cosine to the front normal afterward (matching WRF-Fire’s fire_ros), which keeps a point or finite-line fire’s head at the head rate; needs ros_model = rothermel	“projection”, “advective”	“projection”
erf.fire.directional_split_hamiltonian	With directional_wind_coupling = advective: upwind the isotropic term and the wind and slope terms of the level- set Hamiltonian separately, each advective term by the sign of its own velocity, instead of one front-normal flux; removes the grid-orientation dependence at oblique wind angles; needs ros_model = rothermel	true, false	false
erf.fire.directional_ellipse_lw	Flank rate of that ellipse: the model’s no-wind, no-slope rate, or b over Anderson’s (1983) length-to-width ratio at the effective wind speed; head and back unchanged; needs directional_shape = “ellipse”	“model”, “anderson”	“model”
erf.fire.directional_ellipse_lw_max	Cap on that Anderson ratio (the fit itself saturates at 8)	Real >= 1	8.0
erf.fire.prescribed.ros	Rate of spread for ros_model = prescribed where by_fuel has no entry; no wind, slope or moisture dependence, and directional_ros is turned off [m/s]	Real >= 0	0.0
erf.fire.prescribed.gradient	Gradient of the prescribed rate, dR/dx and dR/dy [1/s]	Two Reals	0.0 0.0
erf.fire.prescribed.origin	Point where the prescribed rate equals its base value [m]	Two Reals	0.0 0.0
erf.fire.prescribed.min_ros	Floor on the prescribed rate [m/s]	Real >= 0	0.0
erf.fire.prescribed.by_fuel	Flat list of fuel code and base rate pairs overriding prescribed.ros per cell; needs fuel_map.file	Reals	none
erf.fire.cheney_gould.moisture	Cheney-Gould dead fine fuel moisture [%]	Real 1-40	10.0
erf.fire.cheney_gould.curing	Cheney-Gould degree of curing	Real 0-1	1.0
erf.fire.cheney_gould.pasture	Cheney-Gould pasture coefficients	“natural”, “grazed”	“natural”
erf.fire.cheney_gould.curing_curve	Curing coefficient: Cruz et al. (2015) or Cheney et al. (1998)	“cruz2015”, “cheney1998”	“cruz2015”
erf.fire.cheney_gould.wind_source	Wind the 1998 model reads: the reference wind at wind_ref_ht (set it to 10 m) or the WAF-reduced midflame wind	“reference”, “midflame”	“reference”
erf.fire.fbp.fuel_type	Canadian FBP fuel type (ros_model = fbp)	C1-C7, D1, M1-M4, S1-S3, O1A, O1B	“C2”
erf.fire.fbp.ffmc	Fine Fuel Moisture Code	Real 0-101	90.0
erf.fire.fbp.bui	Buildup Index	Real >= 0	60.0
erf.fire.fbp.curing	Degree of curing of the grass types [%]	Real 0-100	60.0
erf.fire.fbp.pc	Percent conifer of M1 and M2	Real 0-100	50.0
erf.fire.fbp.pdf	Percent dead fir of M3 and M4	Real 0-100	50.0
erf.fire.fbp.use_slope	Slope through the system’s equivalent wind	Boolean	true
erf.fire.fbp.wind_source	The reference-height wind (set wind_ref_ht = 10) or the midflame wind after the Wind Adjustment Factor	“reference”, “midflame”	“reference”
erf.fire.behave.dynamic_transfer_lo	BEHAVE live herbaceous moisture [fraction] at and below which the whole live herbaceous load transfers to dead, with a linear share up to dynamic_transfer_hi. Acts only on fuels with a live herbaceous load	Real < dynamic_transfer_hi	0.30
erf.fire.behave.dynamic_transfer_hi	BEHAVE live herbaceous moisture [fraction] at and above which none transfers; the share is (hi - M)/(hi - lo), clamped to [0,1]. Aborts unless greater than dynamic_transfer_lo	Real > dynamic_transfer_lo	1.20
erf.fire.behave.net_load	Net fuel load of a BEHAVE category: “weighted” is Rothermel’s surface-area-weighted load with Albini’s size classes; “sum” the plain sum of the classes (the form before 2026-10, 3.3x FM10’s reaction intensity)	“weighted”, “sum”	“weighted”

Balbi model

Parameter	Definition	Acceptable Values	Default
erf.fire.balbi.formulation	Steady explicit 2009 form or convective-radiative 2020 form	“2009”, “2020”	“2009”
erf.fire.balbi.T_a	Ambient temperature [K]	Real	300.0
erf.fire.balbi.T_f	Mean flame temperature [K], 2009 form	Real	1000.0
erf.fire.balbi.T_i	Ignition temperature [K]	Real	600.0
erf.fire.balbi.delta_H	Latent heat of vaporisation [J/kg]	Real	2.26e6
erf.fire.balbi.C_pf	Specific heat of dry fuel [J/(kg K)]	Real	1800.0
erf.fire.balbi.r_00	Radiation length scale [m]	Real	2.5e-4 (2009), 2.5e-5 (2020)
erf.fire.balbi.tau_0	Residence time coefficient [s/m]	Real	75591.0
erf.fire.balbi.chi_0	Radiant fraction, 2020 form	Real	0.3
erf.fire.balbi.C_pa	Specific heat of air [J/(kg K)], 2020 form	Real	1150.0
erf.fire.balbi.C_pw	Specific heat of water [J/(kg K)], 2020 form	Real	4180.0
erf.fire.balbi.T_vap	Water vaporisation temperature [K], 2020 form	Real	373.0
erf.fire.balbi.K1	Convective decay coefficient, 2020 form	Real	130.0
erf.fire.balbi.st	Stoichiometric air to pyrolysis-gas ratio, 2020 form	Real	17.0
erf.fire.balbi.rho_a	Air density [kg/m3], 2020 form	Real	1.2
erf.fire.balbi.sigma_B	Stefan-Boltzmann constant [W/(m2 K4)], 2020 form	Real > 0	5.670374e-8
erf.fire.balbi.max_iter	Root-solve iteration cap, 2020 form	Integer >= 1	40
erf.fire.balbi.tol	Root-solve bracket width [m/s], 2020 form	Real > 0	1.0e-4
erf.fire.balbi.directional	Direction-dependent Balbi spread on the level-set path	Boolean	false
erf.fire.balbi.use_surface_temp	Take the ambient temperature per cell from the atmosphere	Boolean	false
erf.fire.balbi.flame_temp_from_combustion	Derive the 2009 flame temperature from the combustion energy	Boolean	false
erf.fire.balbi.use_cell_moisture	Take the fuel moisture per cell from the moisture model	Boolean	false
erf.fire.balbi.use_moisture_extinction	Zero the rate at and above the fuel model’s moisture of extinction	Boolean	false
erf.fire.balbi.herb_curing	Fraction of live herbaceous load carried as cured dead fuel, 2020 form	Real 0-1	0.0
erf.fire.balbi.wind_source	Wind consumed by the model	“midflame”, “reference”	“midflame”
erf.fire.balbi.heat_flux_coupling	Add the fire-induced vertical velocity from the surface heat flux	Boolean	false
erf.fire.balbi.k_upward	Scale on the induced velocity	Real	1.0
erf.fire.balbi.hf_ref_height	Reference height for the induced velocity [m]	Real	10.0

Hybrid model

Parameter	Definition	Acceptable Values	Default
erf.fire.hybrid.primary	Model used where the weight is 0	a member model; see theory	“rothermel”
erf.fire.hybrid.secondary	Model used where the weight is 1	a member model; see theory	“balbi”
erf.fire.hybrid.selector	How the per-cell weight is set	“region”, “fuel”, “structure”, “wind”	“region”
erf.fire.hybrid.region	Rectangle x_lo y_lo x_hi y_hi [m] taking the secondary model (selector = region)	4 Reals	required
erf.fire.hybrid.blend_width	Linear ramp width across the region or structure edge [m]	Real >= 0	0.0
erf.fire.hybrid.secondary_fuel_codes	Fuel codes taking the secondary model (selector = fuel)	Integers	required
erf.fire.hybrid.structure_file	Building heightmap in ERF terrain text format (selector = structure)	String	erf.buildings_file_nam e
erf.fire.hybrid.structure_min_height	Sampled height above which a cell is a structure [m]	Real	0.5
erf.fire.hybrid.structure_distance	Distance from a structure within which the secondary applies [m]	Real >= 0	10.0
erf.fire.hybrid.wind_lo	Effective wind speed at and below which the weight is 0 [m/s] (selector = wind)	Real	8.0
erf.fire.hybrid.wind_hi	Effective wind speed at and above which the weight is 1 [m/s]	Real > wind_lo	12.0

Propagation

Parameter	Definition	Acceptable Values	Default
erf.fire.propagation_method	Front propagation scheme	“farsite”, “levelset”	“farsite”
erf.fire.farsite.phi_threshold	Legacy update: level-set value at or below which a cell is a front cell	Real	0.1
erf.fire.farsite.use_anderson_lw	Derive the ellipse coefficients from the Anderson length- to-width ratio	0, 1	1
erf.fire.farsite.coeff_a	Richards head coefficient when use_anderson_lw = 0 (the rectangle’s head rate a R; with farsite.shape = ellipse only the ratio (a + c) / (2 b) is read, as the length-to- width ratio)	Real > 0	0.5
erf.fire.farsite.coeff_b	Richards flank coefficient when use_anderson_lw = 0 (the rectangle’s flank rate b R; ellipse: see coeff_a)	Real > 0	0.25
erf.fire.farsite.coeff_c	Richards backing coefficient when use_anderson_lw = 0 (the rectangle’s back rate c R; the ellipse’s back is R / HB of the ratio, not c R)	Real > 0	0.1
erf.fire.farsite.gaussian_sigma	Legacy update: stamp radius [m], < 0 single cell, 0 automatic, > 0 fixed Gaussian	Real	-1.0
erf.fire.farsite.cfl_fire	CFL number of the FARSITE subcycle	Real > 0	0.5
erf.fire.farsite.front_update	Front update: “front_cell” burns the unburned cells next to the burned region at their Hopf-Lax arrival times; “legacy” is the stamping used before 2026-09, which ran about twice the rate of spread	“front_cell”, “legacy”	“front_cell”
erf.fire.farsite.shape	Spread shape of the front-cell update: the Richards (1990) ellipse with Alexander’s head-to-back ratio, or the rectangle of the Richards coefficients read as a support function (the shape before 2026-10, head/back ratio 5)	“ellipse”, “rectangle”	“ellipse”
erf.fire.levelset.cfl	CFL number of the level-set subcycle, dt = cfl h / R_max. Start-up aborts when it makes the substep unstable with the viscosity (see eps_visc). Without viscosity, on square cells, the substep is stable up to cfl 0.888 with first-order differences and 1.015 with WENO5	Real > 0	0.4
erf.fire.levelset.eps_visc	Artificial viscosity coefficient [m] away from the front: a length, so the smoothing in cells is eps_visc / dx. Start-up aborts, naming the largest stable eps_visc and cfl, when the substep is unstable for any front direction (a von Neumann scan of the RK3 substep; see the theory page). On square cells at cfl 0.4 the limit is 0.432 dx with first-order differences (gradient “upwind” or “weno5z_front”) and 0.597 dx with “weno5z”, so the default 0.4 m aborts on fire cells finer than 0.927 m; set eps_visc to about 0.4 dx there	Real >= 0, below the stability limit	0.4
erf.fire.levelset.gradient	One-sided derivatives of the level set: first-order everywhere, HJ-WENO5-Z everywhere, or HJ-WENO5-Z within the front band and first order elsewhere (WRF-Fire)	“upwind” / “weno5z” / “weno5z_front”	“weno5z_front”
erf.fire.levelset.weno_band_cells	Half-width of the front band, in fire cells, that uses WENO with “weno5z_front”	Integer > 0	4
erf.fire.levelset.eps_visc_front	Artificial viscosity within visc_front_cells of the front (WRF-Fire fire_viscosity_bg); negative keeps eps_visc everywhere. Checked at start-up like eps_visc: with WENO5 inside the WENO band of “weno5z_front” (0.597 dx at cfl 0.4, so the default 0.1 m aborts on fire cells finer than 0.168 m), and with first-order differences for the value just beyond the band when that is larger than eps_visc (a band that ends inside visc_front_cells or inside the blend)	Real; when >= 0, below the stability limit	0.1
erf.fire.levelset.visc_front_cells	Half-width, in fire cells, of the near-front viscosity band (WRF-Fire fire_viscosity_ngp)	Integer >= 0	2
erf.fire.levelset.visc_transition_cells	Width, in fire cells, of the linear blend from eps_visc_front to eps_visc beyond that band (WRF-Fire blends over fire_viscosity_band x fire_viscosity_ngp cells, 1 cell with its defaults)	Real >= 0	2.0
erf.fire.levelset.reinit_every	Reinitialise every N subcycles	Integer > 0	5
erf.fire.levelset.reinit_iters	Outer RK3 pseudo-time steps per reinitialisation; 0 skips the reinitialisation and its final clamp (see reinit_clamp)	Integer >= 0	1
erf.fire.levelset.reinit_dtau	Reinitialisation pseudo-timestep [m]; <= 0 selects 0.01 dx (WRF-Fire’s value) for both schemes	Real	-1.0
erf.fire.levelset.reinit_scheme	Reinitialisation scheme: WRF-Fire’s reinit_ls_rk3, or Jiang and Peng’s (2000) HJ-WENO5 with SSP-RK3	“wrf”, “jiang_peng”	“wrf”
erf.fire.levelset.reinit_clamp	Reference of the reinitialisation’s final clamp phi = min(phi, phi_ref), “the burned area only grows”: the level set at the start of the subcycle that ends with the reinitialisation, before its advection (WRF-Fire’s), or the advected field the reinitialisation starts from (the form before 2026-10, which never lets a too-steep burned side relax and moves the front forward). No effect with reinit_iters = 0	“step_start”, “post_advection”	“step_start”
erf.fire.levelset.reinit_jp_sign_eps2	jiang_peng only: term added to phi0^2 under the smoothed-sign square root; <= 0 selects dx^2	Real	-1.0
erf.fire.levelset.wall_extrapolate	Extrapolate the level set into non-burnable cells inside every stencil, so a masked wall is a zero-gradient boundary	Boolean	false
erf.fire.levelset.mask_mode	Non-burnable cells: “wall” holds them at phi >= 0 with a wall stencil, “wrf” (as WRF-Fire) only zeroes their rate (split Hamiltonian path): no hold, no wall stencil, reinitialisation acts on every cell	“wall”, “wrf”	“wall”
erf.fire.levelset.ellipse	Spread as a Huygens ellipse set by the midflame wind (FARSITE family); turns the directional_ros default off, and cannot be combined with an explicit true	Boolean	false
erf.fire.levelset.ellipse_lw	Fixed length-to-width ratio of that ellipse; 0 = Anderson (1983) from the wind	Real >= 0	0.0
erf.fire.levelset.ellipse_lw_max	Cap on the length-to-width ratio	Real >= 1	8.0
erf.fire.accel.enable	Startup acceleration of the rate of spread	Boolean	false
erf.fire.accel.use_temporal	Temporal per-cell model instead of the size-based model	Boolean	false
erf.fire.accel.L_acc	Size-based acceleration length scale [m]	Real > 0	50.0
erf.fire.accel.A_point	Temporal acceleration constant, point fire [1/min]	Real > 0	0.115
erf.fire.accel.A_line	Temporal acceleration constant, line fire [1/min]	Real > 0	0.886
erf.fire.accel.perim_limit	Perimeter length separating point and line fires [m]	Real > 0	500.0
erf.fire.accel.enable_wind_lag	Exponential lag on wind increases (legacy clock only)	Boolean	false
erf.fire.accel.tau_wind	Wind-lag time constant [s]	Real > 0	60.0
erf.fire.accel.clock	Clock of the temporal model: “legacy” per burned cell, restarted when the cell burns and written to burned cells only; “front” from ignition, carried with the front and applied to the rate that moves it	String	“legacy”

Heat flux and coupling

Parameter	Definition	Acceptable Values	Default
erf.fire.tau_residence_s	Fixed residence time [s]; 0 derives it from the cell crossing time and fuel SAV	Real >= 0	0.0
erf.fire.burnout_model	Fuel burnout e-folding time: the cell crossing time floored by the particle time (residence) or WRF-SFIRE’s burn time per fuel model / 0.8514 (sfire)	“residence”, “sfire”	“residence”
erf.fire.burnout_times_s	13 burn times [s] for Anderson models 1-13 overriding the table (sfire)	13 Reals > 0	SFIRE table
erf.fire.burnout_time_to_efold	Divisor turning a burn time into the e-folding time	Real > 0	0.8514
erf.fire.coupling_type	Wind used by the fire and whether its heat is injected	“passive”, “lagged”, “synchronous”	“lagged”
erf.fire.heat_flux_alfg	E-folding height of the injected heat [m]	Real > 0	45.0
erf.fire.fire_atm_feedback	Multiplier on the injected fluxes; 0 is one-way coupling	Real >= 0	1.0
erf.fire.inject_latent	Inject the latent flux into the vapour equation when moisture is active	Boolean	true
erf.fire.heat_flux_partition	Sensible flux handed to the atmosphere: the full heat release (legacy) or the CFBM dry-fuel share 1/(1+M_f)	“legacy”, “cfbm”	“legacy”
erf.fire.prescribed_heat.flux	Constant sensible heat flux added to the fire heat flux over a disc, whatever the fuel and front; injected like fire heat; 0 is off [W/m2]	Real >= 0	0.0
erf.fire.prescribed_heat.center	Centre of the prescribed heat-flux disc [m]	Two Reals	0.0 0.0
erf.fire.prescribed_heat.radius	Radius of the prescribed heat-flux disc [m]	Real > 0 when flux > 0	0.0
erf.fire.prescribed_heat.start_time	Time the prescribed heat flux switches on [s]	Real	0.0
erf.fire.prescribed_heat.end_time	Time it switches off [s]; negative keeps it on	Real	-1.0
erf.fire.source_mode	Replace the theta and vapour source slots (legacy) or add to the rebuilt stage source. Left at the default with another theta source on the fire level (radiation, Rayleigh damping of T, custom forcing or subsidence, immersed forcing) the run stops: set it explicitly	“overwrite”, “add”	“overwrite”
erf.fire.heat_open_fraction	Place injected heat only in the open part of columns with structures; needs structures.enable	Boolean	false
erf.fire.heat_tendency_density	Restore the historical density factor in the injected tendency (injects rho times the flux); false is energy- consistent	Boolean	false
erf.fire.heat_tendency_exner	Divide the rho theta tendency by the cell’s Exner function so the column enthalpy integrates to the fire flux; false (WRF-SFIRE’s form) injects Pi times the flux, 4.5 % short at 850 hPa	Boolean	false
erf.fire.smoke_enable	Add the passive smoke tracer	Boolean	false
erf.fire.smoke_emission_factor	Smoke mass per unit fuel burned [kg/kg]	Real >= 0	0.02
erf.fire.smoke_heat_of_comb	Heat of combustion used to convert flux to fuel burned [J/kg]	Real > 0	1.87e7
erf.fire.smoke_heat_from_fuel	Convert the flux to fuel burned with the fuel model’s own heat content instead of smoke_heat_of_comb (CFBM parity)	Boolean	false

Spotting

Parameter	Definition	Acceptable Values	Default
erf.fire.spotting.enable	Albini ember spotting	Boolean	false
erf.fire.spotting.I_B_min	Minimum Byram intensity to loft a brand [kW/m]	Real > 0	100.0
erf.fire.spotting.P_base	Launch probability per front cell per application	Real 0-1	0.01
erf.fire.spotting.terminal_velocity	Brand terminal fall velocity [m/s]	Real > 0	0.5
erf.fire.spotting.n_traj_steps	Forward-Euler sub-steps along the trajectory	Integer > 0	20
erf.fire.spotting.launch_from	Cells that may launch a brand: every burned cell, or only those with unburned fuel within front_band cells	“burned”, “front”	“burned”
erf.fire.spotting.front_band	Band width for launch_from = front [fire cells]	Integer >= 1	2
erf.fire.spotting.spot_radius	Ignition disc radius at the landing point [m]	Real > 0	10.0
erf.fire.spotting.P_catch	Probability that a landed brand ignites	Real 0-1	1.0
erf.fire.spotting.reentry_fuel_thresh	Minimum remaining fuel fraction for a landing to ignite	Real 0-1	0.05
erf.fire.spotting.random_seed	Random seed; <= 0 seeds from the clock (not repeatable)	Integer	0
erf.fire.spotting.spotting_interval	Apply spotting every N fire steps (one per atmospheric step)	Integer > 0	1
erf.fire.spotting.fuel_system	Fuel system of the maximum-distance table	“13”, “40”	“13”

Crown fire

Parameter	Definition	Acceptable Values	Default
erf.fire.crown.enable	Crown fire initiation and active crown spread	Boolean	false
erf.fire.crown.canopy_base_ht	Canopy base height [m]	Real > 0	5.0
erf.fire.crown.canopy_bulk_den	Canopy bulk density [kg/m3]	Real > 0	0.2
erf.fire.crown.canopy_depth	Canopy depth [m]	Real > 0	10.0
erf.fire.crown.foliar_moisture	Foliar moisture [fraction]	Real	1.0
erf.fire.crown.h_crown_BTU_lb	Canopy heat content [BTU/lb]	Real > 0	8000.0
erf.fire.crown.wind_10m_kmh	Fixed 10 m wind for the crown rate [km/h]; 0 derives it from the effective wind	Real >= 0	0.0
erf.fire.crown.ros_model	Active crown rate-of-spread model	“cruz”, “rothermel1991”, “van_wagner_proxy”	“cruz”
erf.fire.crown.use_passive_blend	Blend surface and crown rates through the crowning fraction	Boolean	false

Diagnostics and output

Parameter	Definition	Acceptable Values	Default
erf.fire.flame_temp_method	Flame temperature estimate	“byram_radiant”, “mcalpine_heat”, “nelson_emissivity”	“byram_radiant”
erf.fire.flame_temp_T_amb	Ambient temperature for the flame temperature [K]	Real	300.0
erf.fire.compute_flame_tilt	Compute the flame tilt diagnostic	Boolean	false
erf.fire.flame_tilt_T_amb	Ambient temperature for the tilt buoyancy velocity [K]	Real	300.0
erf.fire.flame_tilt_rho_air	Air density for the tilt buoyancy velocity [kg/m3]	Real	1.2
erf.fire.write_fire_stats_csv	Append burned area, perimeter and front statistics each step	Boolean	true
erf.fire.fire_stats_csv_file	Statistics CSV filename	String	“fire_stats.csv”
erf.fire.boundary_guard_cells	Width, in fire cells, of the band along every non-periodic edge of the fire grid that the boundary guard watches	Integer >= 0	2
erf.fire.boundary_guard_action	When a burning cell enters the band: “warn” prints once and records the contact time, “abort” stops the run, “none” checks nothing	“none”, “warn”, “abort”	“warn”
erf.fire.edge_reach_check	Print once at ignition the distance to each edge against the largest rate of spread times the time left in the run	Boolean	true
erf.fire.probes	Flat list of x y points [m] whose arrival time is printed once	Reals	none
erf.fire.exposure.enable	Per-structure exposure CSV and fields; needs structures.enable	Boolean	false
erf.fire.exposure.file	Exposure CSV filename	String	“fire_exposure.csv”
erf.fire.exposure.interval	Fire steps between exposure rows	Integer >= 1	100
erf.fire.exposure.ring	Width of the wall band around a footprint [fire cells]	Integer 1-4	1
erf.fire.suppression.enable	Read and apply the suppression action file (fire lines, drops, hold test, burnout); see Suppression	Boolean	false
erf.fire.suppression.file	The action file; required with suppression.enable	String	none
erf.fire.suppression.poll_interval	Fire steps between modification-time checks of the file; 0 reads it once at start-up	Integer >= 0	10
erf.fire.suppression.log	Event log (applied, completed, expired, hold_failed, burnout, reread)	String	“suppression_log.csv”
erf.fire.plot_file	Fire plotfile prefix	String	plt_fire_
erf.fire.plot_int	Steps between fire plotfiles; < 0 disables	Integer	-1
erf.fire.plot_per	Seconds between fire plotfiles; < 0 disables	Real	-1.0