level on a ridge or mountaintop. Relative 

 humidity is, of course, a function of both 

 temperatiire and actual water-vapor content 

 of the air. The values given here were derived 

 by use of Spokane surface and upper-air ob- 

 servations and the hygrothermograph charts 

 from both Gisbome Mountain Lookout 

 (about 10 miles south of the Sundance Fire 

 area) and Big Spot mobile weather station 

 (about 25 miles north of Sundance). Good 

 consistency was found among the tempera- 

 ture and vapor pressure tendencies from these 

 data sources. The Lunch Peak data, however, 

 were not used here; the observed temperature 

 values and behavior showed discrepancy with 

 those at the above stations and at 6,424-ft. 

 Sunset Peak, Idaho (70 miles SSE of the Sun- 

 dance Fire), and were considered unrepre- 

 sentative of the general area. 



A relative humidity threshold of about 30 

 percent has been found empirically by some 

 researchers to be critical in the precondition- 

 ing of fine fuels. The estimated relative 

 humidity for the Sundance Fire environment 

 is seen to have remained below 30 percent on 

 September 1 until after 2200 P.d.t., about the 

 time of maximum estimated gradient wind. A 

 steady increase then occurred, raising the 

 humidity to 75 percent the following morn- 

 ing. For comparison, the low relative humid- 

 ity that persisted for several preceding days 

 and nights is shown in figure 27. The relative 



humidity increase (September 1-2) was re- 

 lated primarily to the decrease in "modified" 

 free-atmosphere temperature (to 51° F. at 

 5,000 ft. m.s.l.), rather than to any large in- 

 crease in vapor content of the air as compared 

 with preceding nights. 



Concluding Remarks 



In summary, we have described the general 

 weather situation associated with the Sun- 

 dance Fire run and have examined more 

 specifically the related wind and relative 

 humidity. Additional research, not reported 

 here, examined other considerations stated in 

 the literature to be possibly important in 

 rapid or "explosive" fire spread. Primarily 

 these are (a) vertical wind profile (change of 

 speed and direction with height); (b) airmass 

 instability; (c) subsidence; and (d) downdrafts 

 from adjacent thunderstorms. The last-named 

 factor may be safely regarded as absent in the 

 Sundance Fire case. Findings with respect to 

 the first three, not particularly crucial, will 

 appear in a forthcoming separate, more com- 

 plete meteorological paper. 



The above-presented weather factors must 

 be considered together with other factors, 

 such as fuels and topography, in accounting 

 for the fire behavior. With this in mind, it is 

 nevertheless quite apparent that unusually 

 strong winds occurred in the Sundance 

 vicinity at a most unfortunate time. 



70 r 



60 



50 



40 



30 - 



20 - 



10 - 



TIME 



FRI. 

 9/1 



THURS. 

 8/31 



WED. 

 a/30 



TUES. 

 8/29 



00 P d,t. 



Figure 27. - Time graph of estimated "modified" free-atmosphere relative humidity at 5,000 ft. m.s.l., Sundance 

 Fire area, for several days prior to and including September 1, 1967. 



32 



