Discussion 



Even though this study was relatively short, the 

 impact of several factors was evaluated on the prob- 

 ability of short-term tree fall or, conversely, length of 

 time standing after fire-caused mortality. Little re- 

 search has dealt with fall rates or retention of fire- 

 killed trees, especially following prescribed fires, 

 which typically take place on sites where snags for 

 wildlife habitat are important. Lyon (1984) studied 

 lodgepole pine snag retention following a severe wild- 

 fire in which most trees were killed immediately with 

 the passing fire. In the first 2 years after the fire, less 

 than 2 percent of the snags fell. During the next 13 

 years, the average annual fall rate was about 13 per- 

 cent, which is about 4 percent lower than the average 

 annual fall rate in the current study. Raphael and 

 Morrison (1987) calculated the probability of snag fall 

 in a mixed conifer stand that had experienced a severe 

 wildfire 15 years prior to their research. During a 

 5-year study, 55 to 65 percent of the dead Jeffery pine 

 and lodgepole pine fell, and little difference in fall 

 rate was noted for trees between 5 and 15 inches 

 d.b.h. Unique to the current study are the examina- 

 tion of the impacts of different seasons of tree injury, 

 different levels of injury, and the length of time be- 

 tween injury and death. These could have implications 

 for snag creation and retention in managed stands. 



With less mortality in the autumn treatment (fig. 1), 

 fewer trees were available for falling, but percent tree 

 fall differences were not statistically significant (fig. 

 2). Keen (1929) found that fall rates were similar for 

 trees killed by bark beetles in the summer and the 

 autumn. Lack of significant difference between the 62 

 percent autumn fall rate and the 78 percent spring and 

 summer fall rate in the current study may be partially 

 due to the small sample size in the autumn treatment, 

 as only 21 trees died and 13 fell. However, an indica- 

 tion based on two findings shows that trees may stand 

 longer if mortally wounded when dormant (autumn) 

 compared to when actively growing. First, figure 4 and 

 the associated analysis show that trees that died 

 within the first year of injury had a higher probability 

 of imminent falling than those that survived the first 

 postfire year but died later. Second, 74 percent of the 

 total tree mortality in the spring and summer burns 

 occurred in the first postburn year compared to only 

 48 percent of the total from autumn fire injury. There- 

 fore, the fact that severe injury in the spring and 

 summer led to a greater chance of early mortality than 

 that caused by autumn fires, implies that a lower 

 probability of early falling for the autumn-killed trees 

 has some credibility. 



With greater mortality in the smaller size classes, 

 more trees were available for falling, but tree fall rates 

 were similar among size classes (fig. 3). Following a 



wildfire in lodgepole pine, Lyon (1984) reported that 

 trees less than 3 inches d.b.h. had an average annual 

 fall rate of 27 percent, while trees between 3 and 8 

 inches d.b.h. fell at about one-third that rate. Dahms 

 (1949) found that about 75 percent of wildfire-killed 

 ponderosa pine between 8 and 20 inches d.b.h. fell in 

 the first 10 postfire years compared to 35 percent of 

 the 20- to 30-inch trees and 15 percent of the 30- to 

 42-inch trees. 



Greater surface area of decay-resistant heartwood 

 for larger, presumably older trees likely explains these 

 fall rate differences between size classes. In the cur- 

 rent study, the ranges of d.b.h. and ages by size class 

 may not have been different enough to expect dissimi- 

 lar amounts of heartwood. Random tree aging indi- 

 cated that many trees 3 to 5 inches d.b.h. were only 

 lOto 15 years younger than trees 12to 15 inches d.b.h., 

 which averaged 90 to 95 years old. 



Results of this study indicate a possible relationship 

 between level of crown scorch and length of survival 

 after injury. In the first postburn year, 51 percent of 

 the 10-year mortality occurred for low-scorched trees 

 compared to 66 percent for high-scorched trees. By the 

 third postburn year, 77 percent of the low-scorch trees' 

 10-year mortality was complete compared to 92 per- 

 cent for high-scorched trees. 



An assumption might be made that if crown scorch 

 and length of postfire survival are related, then only 

 one criterion is necessary to evaluate probability 

 of early tree fall. Indeed, high scorch regardless of 

 length of survival led to almost an 80 percent fall rate, 

 and first-year mortality led to an equally high fall rate. 

 However, even though trees that died following low 

 scorch but survived at least 2 years postburn had only 

 a 27 percent fall rate, low-scorched trees with first- 

 year mortality fell at a 75 percent rate. This indicates 

 the value of both characteristics. Also, based on a 

 small sample size, of the seven high-scorched trees 

 that died after the third postfire year, four (57 per- 

 cent) fell before the end of the 10-year study, but of 

 the eight low-scorched trees that died in that last 

 7 years, none fell. 



Trees in most size and vigor categories can be blown 

 down by high winds. But some structural decline 

 generally has to be present for trees, live or dead, to fall 

 under low to moderate winds. With the exception of 

 trees having burned-out basal scars or heavy fuel 

 consumption around the root crown, the structural 

 integrity of most fire-injured trees is initially un- 

 changed (Kimmey 1955). Other agents become in- 

 volved in response to fire injury leading to structural 

 deterioration, which increases tree fall potential. 



Fire-damaged or fire-killed trees typically attract 

 bark beetles and other insects (Miller and Keen 1960; 

 Mitchell and Martin 1980) that can assure tree mor- 

 tality by girdling the cambium or phloem. Insects that 



5 



