CONCLUSIONS 



The double ellipse formulation developed from Fons' wind 

 tunnel data is providing useful estimates of fire size, shape, and 

 other physical dimensions for field use. The development allows 

 the fire size and shape to be estimated from the downwind 

 distance traveled in a specified time and the average wind in- 

 fluencing the fire. Equations are provided that can express the 

 backing and flanking fire rates of spread so fires that have 

 grown from a spot fire to the line fire or to an irregular-shaped 

 area fire can be projected. This allows the opportunity for pro- 

 jecting fire growth from existing fire lines. 



The accuracy of the equations for area and perimeter is 

 within 10 percent of the area and perimeter determined 

 graphically for Fons' fire sizes and shapes. The greatest uncer- 

 tainty is in selecting the windspeed and the forward rate of 

 spread. Since upper winds and terrain effects must be con- 

 sidered along with the vegetative cover to assess wind, estimates 

 of windspeed will have considerable uncertainty about them. In 

 addition, wind is an input to predicting the rate of spread of a 

 fire. However, the use of historical fire data may help deter- 

 mine the resolution that is possible for field situations. If we 

 assume that the model is accurate, working backward from fire 

 size and spread distances, the average winds on the fire can be 

 estimated. Then fuel models and fire spread mathematical 

 equations can be exercised to provide comparisons to field 

 data. This way confirmation of developed models can be ac- 

 complished and correlations developed to allow updating 

 assumptions made in the formulations of equations. Thus, fire 

 size and shape equations may be useful research tools as well as 

 operational aids. 



Other uses have been evaluated, including use of the model 

 with historical fire, fuel, and weather records to establish the 

 rate of spread and wind necessary to have produced what is 

 documented. This allows an examination of the wind reduction 

 model (Albini and Baughman 1979) by establishing the wind at 

 midflame height and an estimate of the free-stream winds 20 ft 

 (6.1 m) or more above the vegetation cover. These values can 

 be compared to predicted or observed National Weather Serv- 

 ice windspeeds and winds measured at the fire. 



The rate of spread necessary to produce the size and shape 

 of the fire can be tested against fire spread models and the 

 various fuel models to determine if any of the fuel models pro- 

 duce predicted values similar to those measured. If none of the 

 fuel models provides a rate of spread as fast as the observed/ 

 recorded field documentation, the threshold where spotting is 

 contributing to fire growth can be established. If one or more 

 fuel models equal or exceed the observed rate of spread, fuel 

 model representativeness should be examined. 



Either a double ellipse or a simple ellipse fire shape can be 

 used with the equations, and little difference in fire size (acres), 

 perimeter, or fire shape is apparent. However, the most realistic 

 representation seems to be the double ellipse. With either model 

 there will be an error if a backing fire is not possible. The 

 model assumes there is a backing portion to the fire and would 

 overestimate the area and perimeter. 



Historical fire data and maps are being assembled to more 

 thoroughly analyze the double ellipse fire shape model; these 

 will be reported later. Weather and fuel data will be acquired 

 so othe? models can also be tested. Crowning situations can be 

 defined by the wind reduction coefficient needed to match the 

 observed behavior. Investigations on these will complement 

 other work that is addressing the problem of modeling crown 

 fires. 



