Difference between revisions of "WRF-Fire development notes"

Revision as of 16:36, 1 August 2012

Main article: WRF-Fire. To get the software, see How to get WRF-Fire.

This page tracks activity in the WRF-Fire and related software development. All members of the community are welcome to get an account and edit.

Wish list

Add your suggestions or plans here. If you want to comment on any of the items, please start a new section of the talk page and link to it here.

Interpolate by angle between different backing, flank, and forward spread rates.
Rework the propagation scheme so that the speed is computed only once per time step instead inside the right-hand-side in the Runge-Kutta method.
Add flux limiter option as in Rehm and McDermott 2009.
Flank spread rate as a function of base rate and TKE/turbulent fluxes
Turn off artificial diffusion for good
Implement full BURNUP algorithm instead of the simplification.
Keep standalone model working.
Add a canopy fire model for passive (colocated with surface fire) canopy fire.
Parametrize ignition startup. See discussion.
Input fuel moisture from the WRF surface model. See discussion.
Better level set method: High-order ENO and WENO, discontinuous Galerkin.
Rewrite the level set method so that the level set function is ignition time minus current time and get rid of the separate tign array.
Use the Fire SDK to compute ROS.
ROS model from Balbi et al 2009.
Force the fire to a given location from a file read on restart. This will require fire history replay and atmosphere spin-up.
Modify fuel from a file on restart to reflect suppression efforts.
Ignition & fire modification from MODIS fire detection.
Add fractional fuel volume to fuel categories and generate random fuel coverage.
Add other variables in the National Fire Danger Rating System as diagnostics to WRF output.

In progress

When you start on an item from the wish list, please move the item here and add a link to the branch where you work on it.

Fire area and sensible heat flux defined by bitmap images from a satellite or airborne sensor in real run. branch ronan
Connection of emissions products with WRF-Chem. Started, see Coupling with WRF-Chem.
Ignition from a given fire perimeter, with atmosphere and fuel consumption spin-up, by specifying ignition times. See discussion.
Generate KML files for visualization in Google Earth. See discussion
Real test problem with real fire data at both fine and coarse resolution.
Better quadrature for fuel left: The current scheme is second order accurate when all 4 corners of a fire mesh cell are on fire, and exact when all 4 corners are not on fire. In the case when only some of the 4 corners are on fire, the only requirements are that the transition when nodes ignite is continuous and monotonous. In that case, the scheme may not be very accurate. A better scheme would be accurate in more cases and have natural invariance properties. branch jm2/mkim last commit branch vk/test

This affects the amount of heat output from the cell in the timesteps when the fireline crosses the cell.

Set up fireflux test case: branch fireflux, submitted abstract
Data assimilation
Dynamically determine midflame height and interpolate the wind there. See discussion.
Conversion of atmosphere state to use outputs of unmodified WRF to drive the standalone model. See discussion.

Done

Move the item here when done and add a link to the contributor's last commit before the merge into master. Links to major milestone events can be also added to the timeline.

2004

Ned Patton and Janice Coen propose WRF-Fire in a workshop paper and formulate the fundamental principles of how to combine WRF with the fire spread model from CAWFE.

2005

Standalone fire model from CAWFE rewritten in F90 by Ned Patton received Mar 31 2005

2006

John Michalakes writes specs for support of fire grids in WRF Feb 6 2006

2007

Alternative implementation of fire grid in nested domains in WRF by Jon Beezley Feb 29 2007
WRF-Fire development started from the WRF+CAWFE tracers code received from Ned Patton, Jon set up the git repository Jun 6 2007
Matlab prototype of level set method working Aug 30, 2007
Imported a modular rewrite with a new standalone level set code, called SFIRE Sep 5 2007
WRF+spread by level set coupled code working Nov 2 2007
Paper with description of WRF-Fire and data assimilation posted on arXiv Dec 24 2007
Ignition controlled from namelist Dec 24 2007

2008

Change nodes on the fire mesh from corners to centers Feb 10 2008
WPS support 3 Aug 2008
MPI support Oct 23 2008
Cleanup, delete standalone driver and the CAWFE tracer code Nov 18 2008
Landfire data conversion Dec 15 2008

2009

Distributed memory (halos) Jan 7 2009
Approved by UCD and NCAR for public release Mar 17 2009
Set up nested ideal case May 4 2009
WPS support for fire grid and nesting 7 May 2009
Paper with a description of the WRF-Fire and data assimilation published in June 2009
Add namelist.fire to define fuel categories 20 Sep 2009
Interpolate in the computation of fuel_left locally to save stack memory Oct 10 2009
3D visualization in Matlab Nov 3 2009
Visualization in VisTrails and Supercomputing 2009 Demo, Nov 15 2009

2010

Support for WRF restart 16 Jan 2010
Control in namelist the initial atmosphere perturbation (bubble) in ideal run, off by default. 27 Feb 2010 See discussion.
Version 19 Feb 2010 released with WRF 3.2 on Apr 2 2010
Hyperbolic vertical mesh grading in ideal run not only exponential. Mar 9 2010
Fetch high-resolution geogrid data automatically Apr 27 2010
Walking line ignition (as the fireman walks igniting the fire) 23 May 2010
Basic visualization in VAPOR. May 26 2010
Merge git repository with WRF and WPS 3.2 Jun 12 2010 announcement
Interpolate the terrain gradient to the fire mesh instead of differencing an interpolated terrain height. Jun 14 2010 See discussion.
Replace includes for fire parameter arrays by a derived type. Jun 14 2010. See discussion.
Resurrected a standalone driver for the fire code that is independent of WRF for diagnostics. 20 Jun 2010 See discussion.
Surface initialization in ideal runs allows for defining surface properties like roughness, ground temperature, albedo etc., required by soil models and radiation codes. See discussion. Jul 6 2010
Reading from a file in an ideal run: land use, fuel categories, topography Jul 22 2010
Vertical wind profiles in Matlab for diagnostics Sun Jul 25 2010
Interpolate wind to a given height (the same for all fuels) from log profile, and set the height as fire_wind_height in the namelist. Sep 11 2010, added wind reduction factor with defaults per Behave 16 Sep 2010
Rate of spread diagnostics 16 Sep 2010.
Support for NetCDF 4 output files based on HDF5 16 Oct 2010.
Subscale ignition with zero initial radius. See discussion. Nov 2, 2010

2011

Vertical log interpolation of the wind to different heights for different fuels. See discussion. 24 Feb 2011
Ignition time interpolation for better quadrature of fuel left March 16, 2011. See discussion.
Added independent adjustment factors of base spread rate, wind correction, and slope correction to fuel description Mar 23, 2011
Perimeter ignition from specified ignition times on the whole domain, allowing for atmosphere and fuel consumption spin-up. Mar 24, 2011 in ideal run. See discussion.
WRF-Fire released with WRF 3.3 Apr 6, 2011. The release code is based on the Nov 20, 2010 version with bug fix Jan 17, 2011, with further changes made at NCAR.
Started using the name SFIRE instead of WRF-Fire, June 2011.
Reference paper published in GMD July 7, 2011.
Timing from ESMF structures to support variable time step and ignition in a restart run. Jul 24, 2011 See discussion.
Users guide moved to the wiki, including existing howtos as chapters. August 2, 2011
Spread vector and Byram's fire intensity to estimate the severity of a potential fire for a fire danger rating map Aug 26, 2011
Restart from wrfout (cycling=.true.) Sep 16, 2011 See discussion.
Spatially variable moisture as a part of WRF input Sep 27, 2011
Standalone model running from WRF atmospheric state. Oct 28 2011 See discussion.

2012

Integrated fuel moisture model with input from WRF variables, based on the Canadian fire danger rating system Mar 10, 2012. See discussion.
Multiple fire time steps in a single atmospheric step, for stability with highly heterogenous fuels Mar 30, 2012
Fire area and sensible heat flux defined by bitmap images from a satellite or airborne sensor, in an ideal run Jul 27, 2012
Fixed fire of a circular shape, stationary or moving along a line, with a time varying heat flux Jul 29, 2012

Back burner

Items from the wish list that are around for a long time or may never be done should be moved here.

Merge additional changes made independently in the WRF 3.3 release.
Convert ignition namelist variables into arrays for multiple ignitions.
Add running (not colocated with surface fire) canopy fire.
Support WRF-NMM in addition to WRF-ARW
Take the winds from a given distance behind the fireline, and set the distance in the namelist.
Output of emissions products into WRF for visualization of smoke transport and dispersion.
Add fuel models & fuel modeling schemes, esp. Scott-Burgan categories. Input the same fuel description files as BehavePlus and FARSITE.
Generate input data for WFDS.

External links

@@ Line 7: / Line 7: @@
 :''Add your suggestions or plans here. If you want to comment on any of the items, please start a new section of the [[Talk:WRF-Fire development notes|talk page]] and link to it here.''
+* Interpolate by angle between different backing, flank, and forward spread rates.
+* Rework the propagation scheme so that the speed is computed only once per time step instead inside the right-hand-side in the Runge-Kutta method.
+* Add flux limiter option as in [http://fire.nist.gov/bfrlpubs/fire09/art023.html Rehm and McDermott 2009].
 * Flank spread rate as a function of base rate and TKE/turbulent fluxes
 * Turn off artificial diffusion for good