stump-to-truck costs explained by the models (labeled 

 "R 2 "). Both models performed exceedingly well. The 

 model developed for the Northern Region explained 68 

 percent of the variation in stump-to-truck costs. The 

 Intermountain Region model explained 65 percent of the 

 variation. Tree size (ADBH) and the amount of steep land 

 (STEEP) were the most important variables explaining 

 stump-to-truck costs in the Northern Region. Percent of 

 the volume tractor yarded (%TRA) and ADBH were most 

 important in the Intermountain Region. 



In effect, the equations portray stump-to-truck costs as 

 a function of terrain, slope, and tree size. STEEP and the 

 amount of flat land (FLAT) obviously relate to slope. But 

 so does %TRA, because tractor operations typically take 

 place on flatter lands. Steepness is costly, shown by the 

 positive sign on the coefficients. As the percentage of the 

 timber sale tractor yarded goes down or steep land goes 

 up, stump-to-truck costs go up. The ADBH variable indi- 

 cates that regardless of slope, it is less costly (per thou- 

 sand board feet) to log larger trees. 



Transportation 



Transportation costs could also be called log haul costs. 

 The term refers to the costs of moving logs from the tim- 

 ber sale to the initial milling site — haul costs and road 

 maintenance costs. These costs averaged $36.10/M bd ft 

 in the Northern Region and $42.99 in the Intermountain 

 Region. 



Cost estimation models developed for transportation 

 costs explained the most variation of any cost model de- 

 veloped. The equation shown in table 3 for the Northern 

 Region accounts for 76 percent of the variation, while the 

 Intermountain Region model accounts for 91 percent. In 

 both Regions, miles of unpaved haul road (UHAUL) was 

 the single most important variable, accounting for about 

 half of the variation explained. Along with unpaved haul, 

 paved haul (PHAUL) and log size (ADBH) were consis- 

 tently important, in that order. 



The cost equations explain transportation costs by dis- 

 tance traveled. For both Regions, the distance of unpaved 

 and paved haul are significant variables. Together, they 

 sum to total haul distance. These variables always have 

 positive coefficients, meaning that as distance increases 

 so do transportation costs. The size of the coefficients 

 suggests that each mile of unpaved haul is about twice as 

 costly as paved haul. The ADBH variable shows that, 

 regardless of distance, it is less costly (per thousand board 

 feet) to haul large logs than small ones. Apparently, this 

 reflects the notion that smaller logs have less board-foot 

 volume per unit of log weight than larger logs. Because 

 dust control restrictions (DUSTR) is a binary (0/1) vari- 

 able, transportation costs in the Northern Region increase 

 by $4.92/M bd ft when it is present. 



Slash 



Slash disposal costs refer to costs of cleaning up logging 

 debris to prepare the site for regeneration activities. 

 Slash disposal costs averaged $14.44/M bd ft in the North- 

 ern Region and $12.00 in the Intermountain Region. 



The relatively low R 2 's shown in table 4 for the slash 

 cost models suggest that slash cost allowances were more 

 variable than other costs discussed. Only about half of 

 the variation in slash costs allowances was explained by 

 these models. Moreover, the importance of variables 

 differed widely between Regions. Corridor spacing 

 (SPACE) was the single most important explanatory vari- 

 able in the Northern Region, while volume per acre (VPA) 

 was most important in the Intermountain Region. VPA 

 and the number of cutting units (#UNITS) were second 

 important in the Northern Region and percent of the area 

 clearcut (%CC) was second in the Intermountain Region. 



The cost equations portray slash costs in terms of ease 

 and efficiency of operation. Narrower corridor spacing, 

 flatter ground, and either more group selection or less 

 clearcut all make for easier operations and lower unit 

 costs. Similarly, greater volumes per acre and fewer cut- 

 ting units both promote efficiency, and hence cost econo- 

 mies of scale in slash removal operations. But so does 

 removing slash from larger diameter trees, because sites 

 with these trees will carry a disproportionately large 

 volume per unit of land. 



Permanent Roads 



In the Forest Service, these permanent roads are called 

 "specified" roads. They are major roads that access the 

 timber sale and will remain after the sale is completed. 

 These roads can become part of an official, numbered road 

 network for a National Forest. The average cost of per- 

 manent roads was $19.62/M bd ft in the Northern Region 

 and $15.93/M bd ft in the Intermountain Region. 



The cost estimation models developed for permanent 

 roads explained about 35 and 55 percent of the variation 

 in costs in the Northern Region and Intermountain Re- 

 gion, respectively. Table 5 shows these models. The 

 miles of new roads to be constructed (NEW) was clearly 

 the most important variable in explaining permanent 

 road costs. It alone accounted for about half of the vari- 

 ation explained. Note that total road construction is sim- 

 ply the sum of new road miles and reconstructed road 

 miles. 



Cost equations for permanent roads show that miles of 

 road constructed are useful in predicting road costs per 

 thousand board feet. Assume that in addition to fixed 

 costs, the cost per mile of road constructed is constant. 

 The road coefficients identified in this study then imply 

 that the amount of timber accessed by new roads in- 

 creases faster than do the miles of access roads, on a per- 

 centage basis. For under these circumstances, permanent 

 road costs, expressed in terms of a unit of timber harvest, 

 increase as a decreasing function of miles of access roads. 



4 



