New Look at Optimum Road Density for Gentle Topography

Transcription

1 30 Transportation Research Record 898 New Look at Optimm Road Density for Gentle Topography JOHN K. BOWMAN AND RCHARD A. HESSLER, JR. Past forest transportation planning techniqes and trends were reviewed. The needs of the Cheqamegon National Forest located in northern Wisconsin were analyzed. The most applicable method was then adapted to the Cheqamegon. t was fond that road-spacing eqations fit the needs as mch as any other available techniqe. The eqations were modified to inclde log-haling costs and two separate road standards. The eqations were then solved for typical conditions fond on the Cheqamegon. Roads cold be bilt at a 20 percent greater density on the Cheqamegon. This increase wold reslt in lower total costs. The timber-prodction acres will reqire an additional 5100 miles of road at a cost of abot $ This information will be combined with timber-harvest data in the preparation of a Cheqamegon National Forest plan to develop a timber harvest and roading schedle. Throghot the history of transportation planning in the logging indstry, many attempts have been made to determine the optimm logging-road density. n 1942, Matthews dealt with this problem in Cost Control in the Logging ndstry (1). He proposed a method for opt.imizatinn of logging and roaa constrction costs with respect to distance to minimize costs and maximize profits. Matthews, however, did not inclde hal cost in his attempt to define optimm road spacing. This paper expands his work by adding another set of terms, which allows two road standards, and applying them to a new set of eqations where hal is a factor. Several applications of the road-spacing theory can be observed as a reslt of these eqations. For example, timber sales are sally irreglar in shape. We can simlate the shapes as rectangles or sqares, and the road densities (miles of road divided by the nmber of sqare miles) can then be calclated by sing the eqations. An actal area can be selected and measred for road density. This vale can then be compared with the theoretical vale, and conclsions regarding crrent densities can be made (i.e., too few roads, too many roads). This system's main applicability will be related directly to gentle slope terrain de to the assmption that the effect of slopes on hal and skid costs will be minor. The small magnitde of vertical relief in this region makes this assmption valid. Use of this system in a montainos reg ion wold entail the formlation of new and additional terms or, at the very least, a cost for hal, skid, and constrction that reflects the increase created by the slopes encontered dring road bilding and logging. The new eqations give the land manager a clear view of the cost of land development options that can then be sed to compare environmental conseqences. As a reslt, valid and rational land management planning decisions can be made. PURPOSE AND SCOPE The prposes of this investigation were to stdy the transportation needs of the Cheqamegon National Forest in northern Wisconsin, to review past forest transportation planning techniqes in relation to these needs, and to adapt the most applicable method to the conditions on the Cheqamegon. The investigation was carried ot in for phases. The first phase consisted of reviewing the key forest transportation planning techniqes sed from 1900 to the present. Dring the second phase, the critical featres of the Cheqamegon National Forest related to planning forest road and logging systems were identified. n the third phase, the most applicable method was adapted to the precise conditions fond on the acre Cheqamegon National Forest. The last phase developed the adapted method to determine the qantity and cost of roads needed for the optimm road and logging system development on the forest. n all phases, the sitability of the methods was tested for the conditions fond on the Cheqamegon National Forest. The reslts are applicable to areas that have gentle terrain and se mltiple roarl standards. t is qite possible that the methods developed cold be sed for other natral resorce and agricltral applications where the activities are spread fairly evenly over a large area. REVEW OF PAST PRACTCES n the early years of forest transportation planning, it was common practice to focs on an intended activity sch as bilding a campgrond, harvesting a stand of older timber, or accessing a high forestf ire activity area, and then project the roads to these sites. Road standards and rotes were sometimes determined by estimating some combination of probable traffic, perhaps considering travel time, road costs, engineering economics, road impact, and bdget priorities. Most often, standards and rotes were determined by past practices or by the "gt" feelings of engineers or forest managers. n 1942, Matthews (1) recognized the neerl for a more systematic method of planning logging and roading costs. He developed formlas for balancing the cost of haling or skidding logs to forest roads, the cost of bilding roads, and the density of the timber. He considered the cost of prehaling (yarding or skidding), the volme per acre, and the cost of road constrction. Matthews pointed ot that the standard of main roads might change as the roads penetrated deeper into the forest becase of the diminished volme being haled. He stated that higher standards shold be jstified by the redced cost of hal. Later, in 1947, the Forest Service pblished the Cost of Haling Logs by Motor Trck and Trailer (2). For many years, this athoritative work was the basis for. calclating the cost of log hal in the western United States. This excellent work lacked a precise means of calclating road constrction, road maintenance, and logging costs. The 1960 revision, called the Logging Road Handbook (3), inclded pdated costs and haling costs for the eastern United States. Another significant addition to forest transportation planning was the Forest Eng i neering Handbook <il pblished in 1961 by the Brea of Land Management, U.S. Department of the nterior. The handbook otlined a systematic method for planning road and logging systems. A good transportation plan, the handbook pointed ot, is the reslt of concrrent consideration of the physical reqirements of logging methods1 the economics of yarding, reading, and trcking1 the silvicltral system1 ctting priorities1 protection of soil, water, and the nct stand1 and the safety of the entire opera-

2 Transportation Research Record tion. The final plan shold be the compromise reached after weighing all these factors. The necessity for on-the-grond reconnaissance to identify elements that do not show p on photographs or maps was also pointed ot. The fact that the planner cannot avoid field work was emphasized. Team transportation planning became somewhat poplar in the late 1960s. A grop of engineers and foresters on the Olympic National Forest became one of the first teams. t was felt that the participation of other specialists in the development of the forest transportation plan wold reslt in a more effective plan. Throgh this approach, the Olympic fond that an adeqate transportation plan reqired a complete logging plan and field verification. Hal costs were fond to be relevant in few instances in the steep, montainos terrain of that forest. The best plan was often the location that had the least impact on the land while providing a logging system that cold log a maximm amont of land. One reslt of the Olympie's efforts was one of the first systematic, planned skyline logging programs. n the late 1960s, the pblic's image of government changed. The feeling that the government was not responsive to their needs prevailed. A.W. Greely, associate chief, Forest Service, in his address at the 1970 ntermontain Logging Conference (Spokane, Washington, April 2, 1970), felt that the Forest Service wold become a target of the pblic's dissatisfaction. He recognized the trend toward direct confrontation with pblic agencies and demands for change. He also warned that increased accontability and satisfaction of pblic needs wold be essential in ftre management activities. The controversy on the Bitterroot National Forest was evidence of Greely's concerns. The Bitterroot <1l was accsed of having too many poorly located and improperly constrcted roads, which were redcing the natral beaty and casing sedimentation. After examining these isses, the Forest Service fond some of the concerns to be real. t was sggested that a coordinated transportation plan is essential to the orderly development and management of the many resorces on any national forest. The interrelation between the resorce base, the transportation system, the silvicltral systems, and the logging system mst be thoroghly known and analyzed to achieve the objective of qality land management. One obvios soltion was to reqire longer yarding distances with existing logging systems to eliminate some roads. By the end of tt.e 1960s, the complexities of transportation planning demanded a compterized method for analyzing the myriad combinations of road possibilities. n 1965, the Forest Service established the Transportation System Planning Project at Berkeley (~) and an Advanced Logging Systems Program at Seattle (].l. The main emphasis of the Berkeley project was to adapt newer rban transportation planning techniqes to the national forest transportation system. The main focs of the project was a compter program called the Timber Transport Model. The model converted earlier manal network methods to compter format, thereby making it easier to analyze some forest roading problems. The Seattle project was mainly directed at getting design data for advanced logging systems, inclding skylines, helicopter, and balloon. The tools of forest transportation planners have become more complex, and the need. for more precise forest road and logging system tl!chiqes more pressing. At the same time, the pblic has shown more awareness and concern over the environmental effects of logging and road bilding and the trade-offs between the costs and benefits of bilding these road systems. mproved methods of planning are needed to determine the nmber of roads and their dollar and environmental cost so that forest managers and the pblic can nderstand what is being analyzed, why the transportation decisions are important, and what the total transportation system might look like, TRANSPORTATON PROBLEM Log transportation starts after the trees are first ct, limbed, and left in tree lengths or sawed into log lengths. A tractor is sed to skid the trees or logs to a landing area that is accessible by an lb gross-vehicle-weight trck. The landing can be a clearing adjacent to a main road, along a low-standard road, or anyplace on or adjacent to an extremely-low-standard road (Figre 1). Landing areas sally are accessed by extremely-low-standard roads becase those roads are adjacent to the greater timber area. These extremely-low-standard roads are sch that logging trcks can barely travel on them. Trck speeds are sally less than 5 mph. Rocks 1. 5 ft in diamter, rts l ft deep, and large roots sally make these roads impassable to passenger cars. Recreationists, loggers, hnters, and other forest sers mst se light trcks or forwheel-drive vehicles to se these extremely-lowstandard roads. Low-standard roads are the next higher order of roads. Crves are not geometric bt are greater than a 100-ft radis. Maximm grade is 15 percent, trck speeds are abot 15 mph, and selected gravels are sed to reinforce weak areas. The road srface is 12 ft wide, with a 25- to 30-ft clearing width. n some cases, crshed gravel will be sed if it is most economical. Passenger cars can se these lowstandard roads dring most of the year. The main trnk roads are ft wide. Trck speeds are abot 35 mph. Passenger cars can se these roads from early spring to late fall. The minimm crve has a radis greater than 400 ft and grades are less than 8 percent. Abot 90 percent of these roads have a crshed-gravel srface. The highways that access the forest have a 24-ftwide paved srface. Speeds of 50 mph or greater are common for logging trcks. n montainos regions, the problem of planning a transportation system is often simplified becase of the limited choice of logging eqipment. A particlar type of cable logging eqipment may have a maximm of 1000 ft of an appropriate size cable. This fact limits the maximm road spacing to a 1000-ft slope distance. The 1000-ft maximm distance may not be optimm, bt it does limit the nmber of practical possibilities that need to be considered. Two or more logging systems, with qite different characteristics and costs, often need to be considered for the same area. n flat areas, by c.:>ntrast, logs are plled with large-tracked or rbber-tired tractors. There is no fixed distance that determines the needed spacing of roads. Under these conditions, road spacing shold be determined from a stdy of the break-even cost of skidding logs to roads with the cost of bilding more roads. t is the interaction of the cost of logging and the cost of road bilding that determines the amont of road needed and the amont of money spent on roads. As more and more low-standard roads are connected, the increased traffic will, at some point, demand a higher-standard main road. n the western montains, the main road location is often dictated by topographic featres sch as a major valley, ridge, or pass. Techniqes sch as network analysis are highly effective in these locations. A breakeven analysis, however, is more sefl in areas with

3 32 Transportation Research Record 898 Figre 1. dealized road layot on Cheqamegon Natlonal Forest ,, Main Trnk Collector Highway Existing Highways - are spaced at 9-12 miles Existing Main Trnk Collectors~ are spaced at 3-6 miles Low Standard Roads spacing to be de term i ne d Extreme ly Low Standard Roads spacing to be deter mined gentle topography. This analysis consists of calclating the point where increased road constrction will be offset by lower hal costs. Traffic is then estimated to see what the most efficient road standard will be. The qestion to be answered is not, Which road is the best?1 it is, How many roads of what standard is most efficient? Wlnm considering the lowest possible road standard, care shold be taken not to select a standard that is too low. n some cases, increased trck-haling costs more than offset the savings in road constrction. The cost of roads is often looked at as an independent cost element that shold be minimized. Care shold be taken to consider all the elements of transportation rather than jst one or two. Skidding, road constrction, and road maintenance, as well as haling, mst be considered in order to optimize the cost of log transportation. Timber on the Cheqamegon is sold by sealed bid action. n order for national forest timber to be competitive, the cost of logging mst be favorable when compared with other potential sorces near northern Wisconsin. These sorces inclde other government forests sch as state, conty, and town forest lands1 small private landowners1 and forests in northern Michigan, Wisconsin, and arond the Great Lakes in Canada. Roads on the Cheqamegon mst also compete with roads on other national forests. Only those roads that retrn the greatest amont of money along with other considerations are normally bilt on the national forests. The roads on the Cheqamegon mst be highly cost effective in order to be fnded. n order for timber to sell, an adeqate road density shold be available so that the combination of skidding and reading costs are minimized. METHOD SELECTED Matthews' method seemed to fit the conditions of gentle topography, evenly dispersed timber, and small variation of road-bilding costs from one area to another. The difference in the road costs of one location, as compared with another, is often insignificant in flat areas. Hal-cost differences can also be small. One approach to this sitation is a road-spacing-standards analysis that reslts in approximate road spacing for the road standards being considered. Matthews developed a road-spacing eqation that can be sed to develop a gide for road spacing in areas where road locations are not sensitive to cost [Figre 2 (!)]: S=Vl7.4R/VC (!) where S a road spacing in nits of 100 ft, R ~ road cost per 100 ft of road, V 0 volme of timber to be removed per acre, and C cost of skidding logs per 100 ft of rond trip. Trck-haling costs were not inclded in Matthews' eqation becase he fond them insignificant. This was tested for extraordinarily-lowstandard roads where maximm trck speeds are 5 mph or less. t was fond that, for very low roadbilding costs ($5700/mile) and very high haling costs [$5.15/nits of 100 ft of timber (Ccf)/ mile), optimm spacing changes from 840 to 890 ft. When considering a large nmber of these extremelylow-standard roads, this difference in spacing may be significant. Althogh Matthews' eqation gives some very good general direction for road spacing on gentle topography with timber evenly spaced on the land, it does not fit the most common sitation fond on the Cheqamegon National Forest. Rather than one standard of road, the most common condition is extremely-lowstandard roads sed in combination with low-standard roads. The cost of the extremely-low-standard road is often less than half that.of the low-standard road. Adding another layer of roads greatly increases the complexity of a mathematical analysis. n considering soltions to this problem, both mathematical and compter analyses presented themselves as possibilities. A compter program cold be sefl in developing total costs and in solving the program for some of the variables to see how they affect total costs. Unfortnately, sccessfl se of sch a program is predicted on the ser's nderstanding of the mtal interdependency of the vari ables. t was decided that a mathematical soltion wold be more systematic and wold reslt in more consistent findings that cold be clearly presented to the decisionmakers. NEW EQUATONS We decided to modify Matthews' eqations so that mltiple road standards and hal costs wold be considered. Or goal was to develop an eqation that reflected total costs of road bilding and haling for two road standards and logging costs for varios timber densities. Road maintenance also needed to he cons idered. To s i mplify the eqations, road constrction costs were increased by the present vale of 20 years of road maintenance. First, each cost element was identified (Figre 3): Extremely-low-standard-road cost= (Re/52.8) x Lex Ne (2) where Re cost of extremely-low-standard road ($/ mile), Le length of extremely-low-standard road (nits of 100 ft), Ne nmber of extremely-low-standard roads in the stdy area, and 52.8 coversion of miles to nits of 100 ft. Low-standard-road cost= (R/52.8) x L x N (3) where R cost of low-standard road ($/mile) L "' length of low-standard road (nits per 100 ft), and

4 Transportation Research Record Flgre2. Problamdlagram. Llmlt of timber being logged + D s q -... s 4 -s-... s -... s - t Existing Road D = Depth of Timberbel t S = Road Spacing of Branch Roads ~=Average Prehal ing or Skidding Distance Figra 3. Chaqamegon National Forest problem diagram.... i ==== ==== e -e.,, "' N l--~ "' --- -~ c: ~ <> ==== Limit of timber being logged t D Se.. ~ -,.3 ~ ==== :===~ ==== 1 UJ + 1- Se -..j _, ~ Existing Collector Trnk Road D =Depth of timberbelt S = Road spacing of ow standard roads L = Length of low standard roads Se= Road spacing of extremely low standard roads Le = Length of extremely low standard roads Se. -11 =Approximate verage skidding distance N nmber of low-standard roads in the stdy area. Hal on extremely-low-standard road= (He/52.8) x (Le/2) x V (4) where He is the hal cost on extremely-low-standard roads ($/mile/ccf) and v is the total timber volme in the stdy area (Ccf). Hal on low-standard road= (H/52.8) x (L/2) x V (5) where H is the hal cost on low-standard roads ($/ mile/ccf). Hal on collector trck road= (Hh/52.8) x (L/2) x V (6) where Hh is the hal cost on the collector trnk road ($/Ccf/mile). Skidding costs = C x (Se/4) x V (7) where C is the skidding cost ($/nit of 100 ft/ccf/ rond trip) and Se is the spacing of extremely-lowstandard roads (nits of 100 ft). The total cost is eqal to the sm of the factors above: i.e., total cost eqals skidding cost pls hal cost over the varios road standards pls road constrction costs. The nknowns were replaced with terms that reslted in only one nknown factor. All the individal terms were placed into a sing le total-cost eqation (the fll eqation derivation is available from the athors) : Total cost= (CSeV/4) + (HeV/211.2)(S- Se)+ (HhLeV/105.6) + (HLV/105.6) + (2ReL 2 /105.6S.S) x (S - s.) + (RL 2 /52.85) (8)

5 34 Transportation Research Record 898 The first partial derivative with respect to spacing of the low-standard roads (S) was taken and the derivative was set to zeto and solved for S: s =V4(RL 2-2ReL)lieV (9) The tirst partial with respect to the spacing of the extremely-low-standard road (Se) was taken, set eqal to zero, and solved for Se, which reslted in the following eqation: Se = V2ltoL 2 / l3.2v[(c - Ae)f52.8)) (10) n the total-cost eqation, the spacing of the extremely-low-standard roads mst be less than the spacing of the low-standard roads or negative vales will reslt. Empirically, the spacing of the higher-cost roads wold always be greater than for lower-cost roadsi when sing Matthews' eqations, this is always tre. we feel this assmption is safe. A defined area was sed to solve the eqations in order to make L and Le constants. An area x ft was sed to develop the relations. This was checked by sing an area x ft, The spacing checked ot to within 0.02 ft of the vales for the smaller area. To verify that the eqations were realistic, the total-cost eqation was solved on a programmable calclator for varios vales of road spacing and road costs. The reslts were graphed for extremelylow-standard-road spacings of 500, 1000, and 1500 ft. The minimm points on the crves correlate well with the calclated optimm vales. When both the low-standard-road spacing and the extremely-lowstandard-road spacing are near optimm, the total cost is at a minimm. A frther check of the eqations was made for a rectanglar area that had a depth of 8000 ft. This is ~ representation of the actal cndi tions fond on the Cheqamegon National Forest becase the main trnk roads are spaced abot three miles apart. The reslts are very close to the eqations developed for a sqare area. RESULTS To stdy the effects of the new eqations, skidding costs were changed while all other elements were fixed. The reslts were calclated for a variety of road spacings, We fond that skidding cost will not significantly change the optimm road spacing for low-standard roads even thogh it has a large effect on total cost. Timber volme had a sbstantial effect of road spacing (Figres 4 and 5). For a sitation where low-standard roads cost $11 000/mile, extremely-lowstandard roads cost $5700/mile, and the skidding costs $1.74/Ccf, the optimm spacing wold be as follows: Optimm s12acin9 (ft) at S Ccf 10 Ccf JS Ccf ~ ~r Acre 12er Ac re eer Acre Low-standard roads Extremely-low-stan dard roads Figra 4. Coit van1 1paclng for an extremely low 1tlndard road. ;;:- ;;; f-- V 0...J <( f-- 0 f T. ct+ 60)~/i2. T Total Costs, Dollars Cost of Skidding Logs, Dollars per Ccf 6,032 Cost of Road Constrcting, Doi la rs per Ml le Road Spacing, Units per 100 Feet ~)= Volme of Timber to be Removed 100 Cbic Foot Units per Acre Calclated Optimm Spacing V 5 Ccf/Acre V 8 Ccf/Acre V 11 Ccf/Acre ROAO SPACNG UN TS OF 100 FEET

6 Transportation Research Record Figre 5. Cost verss spacing for a low-standard road.... ;;, >- "' T = C~ / vs T =Total Costs, Dollars (~ C =Cost of Skidding Logs, Dollars per Ccf 11,188 =Cost of Road Constrcting Dollars per Mi le S =Road Spacing, Units of 100 Feet V = Vo me of Timber to be Removed 100 Cbic Foot Units per Acre (Ccf/Acre) =Calclated Optimm Spacing ~ V = 5 Ccf/Acre V = 8 Ccf/Acre V = 11 Ccf/Acre 4 o_~--,-----.,.---,---.,--,--,----,-----.,,..--.., ~ S SPACNG OF ROADS, UNTS OF 100 FEET Skidding costs had a sbstantial effect of the spacing of extremely-low-standard roads, The totalcost crve for road spacing of extremely-low-standard roads does not have the broad, flat character of the crves of total-cost crves compared with spacing of low-standard roads, This indicates that the spacing of the extremely-low-standard roads is more critical and has a narrower range of acceptable vales (Figres 4 and 5), The vales of road length needed mst be adjsted from the perfectly shaped geometric model to allow for road locations that fit the topography, soils, or other important featres. This was accomplished by comparing the effective straight-line length with their actal length, A nmber of existing roads that have the same standards as the proposed roads were stdied, t was fond that the actal distance exceeds the theoretical distance by 12 percent, This indicates that the theoretical road lengths shold be increased by 12 percent so that on-theg rond conditions will be allowed for. Road constrction costs shold be increased by 12 percent so that the eqations are correct for actal conditions. The effect of nonprodctive land was also considered. Sch nonprodctive areas as small open marshes, wildlife meadows, peat bogs, and other featres are fond throghot the Cheqamegon National Forest. These non-timber-prodcing areas have few or no roads, A stdy of the forest fond that nontimber-prodcing land ranged from abot 0.5 to abot 3,3 percent of otherwise timbered land. A stdy of the road layot for logging arond these small nontimber areas showed that they had little effect on optimm road density. A qestion that seemed important was, How do crrent road densities compare with theoretical optimm road densities? To answer this qestion, several areas of different timber types were selected, A complete road system was projected by sing crrent techniqes of projecting a normal road density that considered timber, soils, streams, etc. This projected road density was compared with the calclated optimm road density, t was fond that the actal road density ranged from 79 to 82 percent of the optimm calclated density. This indicated that abot 20 percent more roads cold be constrcted and wold reslt in a lower total cost of timber prodction. The lower cost wold be abot $1. 50/Ccf or abot $ /year. The road-density calclations allow s to estimate total road needs for the Cheqamegon National Forest. There are abot acres of national forest land on the Cheqamegon. Of this, abot acres are open water and acres are open or non-timber-prodctive brsh fields. Another acres are reserved for ses other than growing timber, inclding recreation. Becase recreation roads are sally also sed for timber prodction, they do not sbstantially affect the total mileage, Considering these deletions, acres (abot 1172 mile 2 ) are available for timber prodction. Assming an average yield of 15 Ccf/acre and an optimm road density of 6, 5 miles/mile 2, the total road needs for the Cheqamegon National Forest are 7600 miles. There are crrently 2274 miles of forest roads and 231 miles of highways nder state or conty jrisdiction. This shows that there are an additional 5100 miles of roads needed to log timber at an optimm least cost on the forest. The estimate record is that a total of 3500 miles of road are needed on the Cheqamegon. This wold mean that only 1000 more miles are needed. Abot 50 miles are constrcted on the Cheqamegon each year. t is apparent from the progress being made that the additional 1000 miles needed is qite low. Of the 5100 miles needed, abot half will be low-standard roads and half will be extremely-low-standard roads, The total cost will be abot $ Road-density calclations have direct application to the type and age of timber stands, Most of the timber stands on the Cheqamegon are now years old. Some species sch as aspen and jack pine have a rotation of years. These species are now being clearct. Other species have a rotation of 120 years. These stands will ndergo commercial thinnings beginning near age 35 and periodical thinnings ntil age 90, Road-density calclations will

7 36 Transportation Research Record 898 be modified by disconting ftre costs for these stands to allow for the redced early yield and still consider the mch heavier final ct. A foc&st plan is crrently being developed for the Cheqamegon National Forest. This plan considers pblic opinion and data abot the natral resores rond on the forest, The analysis portion of the plan will inclde calclations to determine the optimm time to ct the varios timber stands while maintaining an even, sstained yield of timber and other resorces from the forest. The road-density calclations will be matched with the harvest plans to develop the miles of road needed for timber harvest for the first 10 years of the plan. The constrction cost by year, and the cost of road maintenance, will be a part of the plan and will be sed to develop the forest bdget. CONCLUSONS The road-spacing eqations work well for gentle topography. Optimm road spacing is fairly easy to calclate for different road and hal costs for two road standards, skidding costs, and timber densities. The minimm total cost of logging and haling timber will be achieved at the optimm road spacing. Roads are not laid ot in straight lines. They crve arond obstrctions, which increase constrction cost. The theoretical vales can be compared with actal sitations so that the tre road lengths can be estimated. The Cheqamegon National Forest will need an additional 5100 miles of road to remove timber at an optimm least cost. These additional roads will cost abot $ Having a greater road density shold save $ /year. The optimm road-spacing eqations, in conjnction with timber stand prodction data, will form the transportation planning portion of the Cheqamegon' s forest plan. The road-spacing, road, and road maintenance costs will be calclated for varios planned timber harvests so that a firm bdget can be develped. REFERENCES 1. D.M. Matthews. Cost Control in the Logging ndstry. McGraw-Hill, New York, , J.J. Byrne, R.J. Nelson, and P.H. Googins. Cost of Haling Logs by Motor Trck and Trailer. Forest Service, U.S. Department of Agricltre, Oct revised May J.J. Byrne, R.J. Nelson, and P.H. Googins. Logging Road Handbook: The Effect of Road Design on Haling Costs. Forest Service, U.S. Department of Agricltre, Agricltre Handbook 183, Dec J. K. Pearce. Forest Engineering Handbook. Brea of Land Management, U.S. Department of the nterior, Oregon State Office, Management Practices on the Bitterroot National Forest--A Task Force Appraisal. Forest Service,.s. Department of Agricltre, Missola, MT, April 1970, 6. The Advanced Transportation Planning Programs of the USDA, Forest Service. Forest Service, U.S. Department of Agricltre, Field Notes, Vol. 1, No. 6, Nov W.E. Fren. Advanced Engineering Technical Specialty Training Program. Forest Service, U.S. Department of Agricltre, Field Notes, Vol. l, No. 3, Ag Classification of Unpaved Roads in Ontario E.F. DOBSON AND L.J. POSTLL Unpaved roads constitte a large portion of the total road network in Ontario. Crrently, there is no program available to the mnicipalities that will enable them to classify their npaved-road network and apportion their existing maintenance fnds across the road system in a cost-effective manner. This paper otlines a simple bt rational approach to classifying npaved roads into three distinct classes by sing for qality-of-service characteristics. A formla was developed that permits the existing maintenance expendi tres to be apportioned over the three classes of npaved roads. The concl sion s that classification of the npaved roads in Ontario will permit the n paved-road network to be elevated to the level of maintenance management now practiced on paved and srface-treated roads. Sch a program will en hence the task of the roads manager and aid the taxpayer, traveler, and those who 1trive to conserve or natral resorces for ftre generations. Unpaved or aggregate-srfaced roads have been an integral part of the road and street network in Canada for the better part of a centj:y. As the nation has grown, ~.. bas--the-roa:a - s ystem, giving ever greater access to the frontier areas of Canada. nitially, roads were hewn ot of the forests, and the existing soil fm"meotiie- ro aa srf ace. Maintenance reqirements were nominal. The development of the motorized carriage triggered a new era, and along with this indstrial development came a need for improved roads. ndstry and the engineering commnity responded, and today we have a great network of roads. n the Province of Ontario, there are km of npaved roads, of which km are maintained by individal mnicipal governments. The responsibility for maintaining the npaved-road network may be the contyi the individal township within a contyi the individal city, town, or village within a township1 or the provincial body--the Ministry of Transportation and Commnication (MTC). The greatest percentage of npaved roads is nder the jrisdiction of the township governments, as shown in Table l (from 1980 MTC data). The network of roads contines to expand as the poplation grows and its needs increase. t is not practical to transform all of these npaved srfaces into hardtop srfaces and, as a reslt, npaved roads will contine to form an integral part of the total road network for the foreseeable ftre. The costs for maintaining the road system have escalated dramatically since the mid-1970s and, in recent years, the mnicipalities have not been able to raise these increased costs from the ratepayers. As a reslt, there have been arbitrary cts in some road programs to offset escalating costs in other