If you have Javascript enabled, move your mouse over the map to see a thumbnail of the smoke forecasts for each region. To see the current forecast for a region or to view an archived forecast in detail, click on a fire location area on the map.

NATIONAL FORECAST VERIFICATION REGIONAL FORECAST
		Current National 24 Hour Analysis and 48 Hour Forecast Loop
		Archived National GIS Smoke Forecasts & Fire Location Images

Regional Output Archive

The model output graphics consists of a series of "gif" images with the filename "smoke{HH}.gif", where HH is the two digit UTC hour representing the end of the averaging period. The HH field varies from 01 to 24 for analysis simulations and 01 to 48 for forecast simulations. The file "loop.gif" is an animation of the individual hourly smoke images. The files are written to a directory on the FTP server corresponding to the region and simulation day: ftp://gus.arlhq.noaa.gov/pub/web/fires/{REGION}/{DD}, where REGION corresponds to the domain directories indicated in the previous table. The forecast simulations writes the output to directory {DD}, while the analysis is written to directory {DD-1}, replacing yesterday's forecast products. Therefore, execept for the current day, the directory contents will consist of analysis products for the first 24 hours and the next day's forecast for hours 25 through 48. In this way forecast and analysis products can be compared if desired. Note that if there are no fires in a region, the "old" products remain in the directory.

Archive and forecast graphics for specific geographic areas:

Model Changes

For a history of model changes refer to the smoke forecast model change log.

Overview

Smoke forecasts are produced by NOAA's Air Resources Laboratory using the HYSPLIT dispersion model. The dispersion model is preconfigured to run over the entire country once-a-day using the daily 0600 UTC NAM and GFS meteorological forecast. Hourly average output maps of primary PM2.5 air concentration are produced using the actual fire locations from the previous day. The dispersion simulation consists of two parts: 1) a 24 h analysis simulation run for the previous day, and 2) a 48 h forecast simulation, which assumes that yesterday's fires will continue to burn today and tomorrow unless fire duration information is available for a particular fire. The smoke particle positions at the end of each analysis period are used to intitialze the next days's analysis simulation.

Fire Locations

Fire locations for the dispersion simulation are obtained daily for yesterday's fires from the NESDIS Hazard Mapping System. A preprocessor reads the fire position data file representing individual pixel hot-spots that correspond to visible smoke and aggregates the locations on a 20 km resolution grid. Each fire location pixel is assumed to represent one sq-km and 10% of that area (10 ha) is assumed to be burning at any one time. The aggregated gridded fire locations file is saved each day. The file is read the next day by the preprocessor to be added to that day's fire locations. The new fires identified are assumed to be the ones valid for that day and through the forecast period. However, if a grid cell has no new fires, but had substantial fires the previous day, then the previous day's fires are assumed to decay at a rate of 75% per day until the fire grid cell has less than 0.75 pixels burning. At that point the number of fires per cell is set to zero. The decay rate is NOT applied to the forecast period.

Wildfire Emissions

The amount of particulate emission associated with biomass burning is quite uncertain and depends upon many factors: biomass type, moisture content, wildfire or slash and burn agricultural clearing. Emissions can change depending on whether there are flames or just smoldering. Levine (1994) estimated total particulate matter emissions from biomass burning in the tropics to be 36-154 x 10^9 kg/yr. The authors estimate 240 million ha under slash and burn agriculture with an annual clearing of 24 million Ha in tropical secondary forests. Another estimate, from Crutzen and Andrea 1990, determined emission rates of aerosol measured at around 12g aerosol /kg C fuel for a tropical forest. Using these ranges, the normalized emissions are estimated to be from 5 kg/hr/Ha to 50 kg/hr/Ha. (Crutzen, P.J. and M.O Andreae, 1990, Biomass Burning in the Tropics, Impact on Atmospheric Chemistry and Biogeochemical Cycles, Science, 250, 1669-1678; and Levine, J.S., 1994, Biomass Burning and the Production of Greenhouse Gases, Climate Biosphere Interaction, Biogenic Emissions and Environmental Effects of Climate Change, Edited by R. G. Zepp. John Wiley and Sons, Inc)

HYSPLIT Emission Estimates

Meteorology

The dispersion model calculation uses hourly NAM (12 km resolution) meteorological data fields over the CONUS and 3-hourly one-degree resolution Global Forecast System data fields for any fire locations outside of the NAM domain. Dispersion calculations are run once-a-day using the 0600 UTC forecast cycle. The dispersion model run is started about 0700 LDT (East) and takes about one hour to complete. The 24 h NAM analysis data file consists of a combination of initial fields (+0 h to +5 h forecasts) four times-per-day (00, 06, 12, and 18 h UTC). The NAM model fields have have been previously interpolated to sigma coordinate system with a vertical resolution of about 10 levels within the first 850 hPa from the ground.

Dispersion Model Configuration

The HYSPLIT simulation was configured to release puff-particles at each time step over the duration of the simulation and over all the aggregated emission locations. In a typical calculation about 10,000 particles would be used to track the emissions from each fire location. The smoke particle release height is assumed to equal the final bouyant rise height as computed using Briggs (1969). The final rise is a function of the estimated heat release rate from the BlueSky emission algorithm and the forecast stability and wind speed at each release hour. Smoke particles are terminated after 72 hours. Smoke particles may lose mass due to dry (gravitational settling) and wet removal. Smoke particles are assumed to have a diameter of 0.8 um with a density of 2 g/cc. Wet removal is much more effective than dry deposition and smoke particles in grid cells that have reported precipitation may deposit as much as 90% of their mass within a few hours. Two air concentration grids are defined for each simulation, each having a grid resolution of 15 km. One grid creates hourly averaged air concentrations (ug/m3) from the ground to 5 km for comparison with satellite smoke plume observations. The second grid defines the layer in the lowest 100 m as hourly average air concentrations (ug/m3) for air quality applications. Note that deposition and/or a small vertical concentration layer may result in patchy looking simulation results.