tion in all but one case was southwesterly or 

 west-southwesterly, with some sort of upper 

 trough located near the Washington coast. 



We have thus far looked at wind indica- 

 tions from upper-air observations. An obvious 

 question is how do these winds aloft relate to 

 those actually observed or expected at 

 ground- or tree-level locations. Winds in 

 mountain areas may, of course, vary consider- 

 ably according to exposure to the general air- 

 flow. In addition, surface friction acts to re- 

 duce the vvdndspeed near the ground, but 

 terrain-induced turbulence (proportional to 

 gradient windspeed) may bring strong gusts. 

 Locally, winds may be increased by terrain- 

 enforced convergence of airflow. Figure 25 

 compares the free-atmosphere winds observed 

 above Spokane with the winds at several sur- 

 face stations. 



In the Sundance Fire vicinity, hourly wind 

 data were available, except for several night- 

 time hours, from Lunch Peak, at 6,400 ft. 

 m.s.l. 20 miles to the southeast. A mobUe 

 fire-weather station, sent from Boise, was in 

 operation here for the Plume Creek Fire. A 

 rather abrupt rise in the Lunch Peak wind- 

 speed is noted between 1600 and 1700 P.d.t. 

 on September 1. Winds from the southwest 

 soon reached an average of approximately 35 

 m.p.h. at 1800 and 1900 P.d.t.; gusts (not 

 plotted in the diagram) were up to 50 to 55 

 m.p.h. for several hours. 



Roman Nose Lookout, at 7,264 ft. m.s.l. 

 on the northern edge of the fire, had a 12 

 m.p.h. southerly wind at the regular 1600 

 P.d.t. fire-weather observation time; a steady 

 speed of 35 to 40 m.p.h. from the southwest 

 was reported at 2015 P.d.t. (note point 

 plotted in figure 25). A lookout atop Mt. 

 Henry (7,235 ft. m.s.l.), in extreme north- 

 western Montana, reported winds averaging 

 40 m.p.h. and gusting to 60 m.p.h. (time not 

 given) . 



Among the mountaintop fire-weather sta- 

 tions reporting once daily, four of the seven 

 in northeastern Washington (at elevations 

 ranging between 3,800 and 5,900 ft. m.s.l.) 

 had windspeeds between 27 and 33 m.p.h. at 

 1600 P.d.t. Speeds at five such stations in the 

 Idaho panhandle (at elevations between 5,200 

 and 6,400 ft. m.s.l.) were as yet only between 

 12 and 16 m.p.h. To illustrate possible 

 terrain-exposure effects, Squaw Peak, 



Montana (at 6,200 feet), nearby to the south- 

 east, had 29 m.p.h. winds at this time. 



The Lunch Peak (and Roman Nose Look- 

 out) windspeeds fit in reasonably well with 

 the 5,000- to 7,000-ft. wind trend at Spokane 

 during the late afternoon and early evening of 

 September 1. The same can be said for the 

 winds atop a 400-ft. tower at the Hanford 

 A.E.C. Works, ground elevation 700 ft. m.s.l., 

 about 120 miles southwest of Spokane. The 

 available Lunch Peak and Hanford speeds dif- 

 fer considerably, however, from those of the 

 free-atmosphere winds at Spokane later in the 

 evening. This difference is probably associated 

 with variations in the vertical, turbulent trans- 

 fer of momentum through the "friction 

 layer." This layer would extend mainly to 

 near 5,000 ft. m.s.l. at Spokane and to corre- 

 spondingly higher altitudes above the moun- 

 tainous area in northern Idaho. Although ref- 

 erence is made in this paper to free- 

 atmosphere winds at 5,000 ft. m.s.l. in the Sun- 

 dance area, such winds are largely hypothetical. 



It may be well to clarify here the relation- 

 ship between the strong winds in the Sun- 

 dance area and the movement of a cold front. 

 The strong, or relatively strong, and gusty sur- 

 face winds observed in eastern Washington 

 and northern Idaho on September 1 began by 

 middle or late afternoon. This onset took 

 place, at high and low elevations, approxi- 

 mately 100 to 150 miles in advance of the 

 cold front that had developed east of the Cas- 

 cades. Cooler-air movement, representing a 

 gradual change of airmass, was, however, al- 

 ready accompanying these prefrontal winds. 

 The front in this case thus appears to repre- 

 sent a secondary steepening of horizontal 

 temperature gradient imbedded within a 

 broader gradient. The peak in 5,000-ft. m.s.l. 

 free-atmosphere windspeed, which was 

 reached in the late evening, occurred near or 

 slightly behind the surface map projection of 

 this eastward-moving front. But, as indicated 

 earlier, the front was rather diffuse or un- 

 detectable in northern Idaho and extreme 

 eastern Washington. 



In the absence of more detailed data cover- 

 age, the actual effect of this diffuse front (or 

 frontal zone) on the Sundance surface winds 

 remains a matter for speculation. It is, how- 

 ever, apparent that the front alone could not 

 have accounted for the unusually strong, 



30 



