0040794 Journal of Applied Ecology (1978), 15, 451-457 TRAMPLING EFFECTS OF HIKERS, MOTORCYCLES HORSES IN MEADOWS AND FORESTS AND T. WEAVER AND D. DALE Department ofbiology, Montana State University, Bozeman, Montana 59715 SUMMARY (1) The effects of experimental trampling due to hikers, horses, and motorcycles were studied in a Festuca idahoensis-poa pratensis meadow and in a Pinus albicaulis-vaccinium scoparium forest. (2) Horses and motorcycles were more damaging than hikers. Motorcycles were most damaging when going uphill while hikers and horses were most damaging when going downhill. (3) Damage was less rapid in grassy vegetation than in shrubby vegetation. (4) Damage was generally greater on slopes than on level ground. INTRODUCTION The management of trails should be based on knowledge of the effects of humans on them; most of the available information has been reviewed by Speight (1973), Stankey & Lime (1973), and Liddle (1975). Observations of existing trails suggest the following conclusions. (1) Vegetation cover is reduced by trampling and some plants are more resistan to trampling than others (Speight 1973; Liddle 1975; Dale & Weaver 1974; Davidson & Fox 1974). (2) Trail width increases linearly with increasing slope, wetness, roughness and the logarithm of the number of users (Bayfield 1973; Dale & Weaver 1974), but decreases linearly with the logarithm of the roughness of trailside vegetation and terrain (Bayfield 1973). Vegetation more than 2 m from the edge of a trail is often little effected (Dale & Weaver 1974). (3) Trail depths depend on compaction and erosion and therefore on climate, vegetation type, soil and substrate type, slope, and type of user (Helgath 1975; Dale & Weaver 1974). (4) Soil compaction is usually greater, i.e. bulk densities average 02-06 g cm-3 greater, in trampled areas than in untrampled areas. Several experimental studies.have shown the quantitativeffects of wear on vegetation (Bell& Bliss 1973; Liddle 1975; Liddle & Greig-Smith 1975; Rogova 1976) but there has been no experimental comparison of the effects of different modes of trail travel. The experiments reported below compare the effects of hiker, motorcycle, and horse trampling on level and sloping sites in both a meadow and a forest with a dwarf shrub understorey. METHODS The site The two habitat types studied represent conditions commonly found in the northern Rocky Mountains. The first area was a Poa pratensis-festuca idahoensis grassland at an 0021-8901/78/0800-0451$02.00? 1978 Blackwell Scientific Publications 451
0040795 452 Trampling and vegetation altitude of 2070 m near the Battle Ridge U.S. Forest Service Ranger Station in the Bridger Range, Montana. The area was a lightly grazed example of Daubenmire's (1970) Festuca idahoensis habitat type. Soils were deep, stonefree, sandy loams (7000 sand, 700 clay) probably of the Teton series. At the nearby Bozeman 12NE weather station, annual precipitation averages 850 mm, growing season precipitation averages 293 mm, January temperatures average -7 0C, and July temperatures average 13 'C. The second area was a Pinus albicaulis forest at an altitude of 2770 m in the Hidden Creek drainage near Hidden Lake, Gallatin Range, Montana. It is similar in vegetation, soil, and probably climate to those described by Weaver- & Dale (1974). Its understorey was dominated by the low shrub (20 cm) Vaccinium scoparium. Its soils were deep stony sandy loams (730% sand, 90 clay). A level (0 0 slope) and a sloping site (15 0 slope) was chosen in each study area. In each case, the two sites were within 150 m of each other and had similar soils and vegetation. At each of the four sites a nest of parallel, rectangular, and overlapping trails was established. The hiker trails were 30 x 12 m rectangles, the motorcycle trails were rectangles of the same dimensions offset 3 m upward and 3 m to the right; and the horse trails followed similarly offset rectangles of the same size. Method Each trail was trampled 1000 times during the summer of 1973. Trips were recorded with hand talleys. The hikers weighed 91 and 82 kg and wore hiking boots with cleated soles. The motorcycle was a Honda 90 running in second gear at speeds less than 20 km h- 1. The horses weighed 500, 550 and 579 kg and were shod with uncleated shoes. At the forest site, half the passes along the trail were made in one direction and half in the other to stimulate two-way traffic. At the meadow site trampling was all in one direction so uphill and down-hill effects could be contrasted on the sloping site. Soil water during the trampling period was always below field capacity and ranged from 8 to 3000 (weight basis) on the forest site and 27 to 370 on the meadow site. One should consider these differences when comparing data from the grassland and forest sites. Ten transects were marked across each trail. Beforeach trail was used, and after each successive 100 tramplings, the following were estimated each transect: (1) percentage of bare ground in a 2 x 5 dm quadrat placed at trail center and parallel to it, (2) bare trail width and (3) trail depth as the vertical distance from a measuring stick down to the trail centre; (4) bulk densities were also measured after each 100 passes, using three cylindrical cores (5 cm diameter x 10 cm) taken between the vegetation sampling sites. The experiment was replicated three times. The second and third replicates were trampled 500 and 100 times respectively. Essentially identical effects were noted but are not reported here. Statistical analysis Means and standard errors for each measurement were calculated for each site, treatment, and sampling period. These standard errors and simple sign tests (Moroney 1951) have been used to test the significance of differences seen in Figs 1 and 2. Using the sign test, two treatments differ significantly (P < 0 05) if the following proportion of the observations for one treatment are above (or below) those for another treatment: 9/10, 15/20, 21/30 or 27/40 (Moroney 1951). Coefficients of variation were generally similar between vegetation types, slopes,- and modes of transport; they decreased markedly with
0040796 T. WEAVER AND D. DALE 453 TABLE 1. Standard errors for data typical of that appearing in Fig. 1. N = 10 for bare ground, width, and depth. N = 3 for bulk density, no standard error is shown for bulk densities of forest soils because the three cores were pooled before analysis Hiker Motorcycle Horse Passes 0? 150 00 150 00 150 Meadow Bare ground 100 2 4 2 5 0 3 6-4 2 2 7-6 500 61 108 74 04 3-6 10-3 1000 8 1 105 103 00 70 00 Width 100 3 2 1.9 17 1.9 1-4 1-8 500 51 24 47 4-5 4-1 18 1000 14 22 5.1 37 40 2-9 Depth 100 03 07 03 12 06 04 500 03 07 03 09 07 05 1000 04 0-8 04 08 07 0-8 Bulk density 100 106 2-4 8-0 8 7 134 4-8 500 116 99 49 86 53 38 1000 4.9 3-3 7-7 132 3-7 7-9 Forest Bare ground 100 5 7 8-1 7-3 4 9 6 9 5 5 500 4.3 5-3 2-2 00 05 00 1000 09 1.1 20 00 00 00 Width 100 0 8 15 1-7 54 1.9 49 500 16 2-9 3-2 27 18 39 1000 06 08 22 3 1 30 65 Depth 100 02 04 0-2 05 06 02 500 02 0-4 05 06 07 05 1000 03 03 07 07 1.0 1-7 increasing numbers of passes. After 1000 passes, coefficients of variation averaged 18% for measurements of bare ground, 110% for width, 3900 for depth, and 30 for bulk density, by contrast, after 100 tramplings the coefficients of variation averaged 65% for bare- ground, 50o% for width, 66% for depth, and 90% for bulk density. RESULTS AND DISCUSSION Vegetation cover The percentage of bare ground increased with increasing numbers of passes (Fig. 1); this effect was more rapid on sloping sites than on level sites, and was more rapid in dwarf shrub than in grassy vegetation. On level ground, horses were most destructive and hikers least destructive but on grassy slopes motorcycles were more destructive than horses. The greater effect of trampling on slopes than on level areas is expected, since there is more sliding, spinning and braking on sloping sites. The greater effect of trampling on shrubs is expected, since grasses are more resistanto trampling than shrubs are (Dale & Weaver 1974). Alpine vegetation, like forest vegetation and unlike meadow vegetation, is very sensitive to trampling (Bell& Bliss 1973). Hikers had smaller effects on vegetation than either horses or motorcycles on both level and sloping sites because the hiker applies less pressure than do either horses or motorcycles. There was no significant difference between motorcycles and horses in the dwarf shrub sites; the greater destructiveness of.motorcycles on the sloping meadow site is probably due to the fact that the impact was concentrated a relatively narrow trail.
0040797 454 Trampling and vegetation GRASSLAND FOREST Level Slope Level Slope H1ker H0 iker 0 -Motorcycle /1/ 50- O~~~~~~~~~~10 10 500 1000 loo- 00 ++_1/! / 0 + L 500 000 500 1000 ) 500 00 500100 co /s /-.- 50 + Li: a-.- 5.c Hors 10+??\"++++_+0+ 1v41 4. 10- v +_~~+/? 00 00 100-7 0_-? * e+ E~ +----+ + O_--+++-- FIG. &1. The. /fets..he. ), c )......Bars. 5a 0 e ri 500o1000 g000 500 560 - ~~~~~~~~~~~~~~~~~~~~-0/_ C 00 00., 50-0 000 00 1000 0 C...~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~... 2 - _ +.... Trail w l.0 width. inrae wt irs l- e -.0 use (i 1). oas i on sa o- j 10* _ 1s- le s lel an s n 0 soo "odo o 5,00 1000 500. O a.~~~~~~~~~~~~~aa 1.e FIG Thmerefet of hikers; dat, motrcyeste here andg hoss() onppgrasln tand forestrvsites. Bars r atldthes lowrais rhighto each grahnhow(ef to right)e th0ues mannstandyarderrr foprohimaerswic motorcycl tesean horsesrespectvly Tablel1 afurthr gives information; athrbue on variabilityenof thes trailgwidthr ineased with ioneprieaseaus (Faig. 1).t Ather fany given numberso pases ntrwalk widthst wer ousull gxprieater.o slopingrpot thanto levelstes. Onde bot leelandsopithng sites torails width watreaest 'frc hokrse and leastforhikers. Dale & Weaver (1974),ia study nof abnfreast in our experiments. Previu grepotsr withato trails wer widering mieadows thoanl ine to 'searching for footing' (cf. Bayfield 1973). The tendency of path widths to be greatest
0040798 T. WEAVER AND D. DALE 455 Uphill Downhill 100? 500 --+ 100? 50 0 + '1K H Ike D50 505/1/ / l I ~ ~ ~ ~ ~ /1 + ~~~~~~~ ~Motorcycle ~~~~~~~~~~~~+ Horse 500 1000 500 1000 1005 0 50 1001 N T T..+- + standard errors for hike,-- m a + + 0- + 0 oa.-01 + 0'-I +. o-w L V 5Q- + /P0-IT50 I- 6 0~~~~~ 10 T/ ~~ 10 T - To 0 ~- -+ 0 2 0 00Q - a lpn 15 0) medo ste Basa telwe ih ofec- rp hw(ettih)tema standard rrors fo hies moocce0n ossrsetvl.vria asaon oedt E 0~pit ersnttelretidvdulsadr ros a5 la passes; ( tehit 1000 for horses wildern"s and ''sao''idoo les o hkrbohgnral n inteeeprmns mus be duoh d) -ol e~~trildet at least pass 0 i)tri 1000 50 etstedt egeae nsoe thno lvlste;(ii0ri
0040799 456 Trampling and vegetation depths tend to be greater in a stone-free meadow soil than in a stony forest soil, at least for hikers and cycles; (iv) trail depths were greatest under horse use and least under hiker use at all sites. Increases in depth were generally most rapid in the first few hundred passes, regardless of user, and could be due to both soil compaction and erosion. Erosion is most rapid on sloping sites (Helgath 1975). We attribute the relatively shallow trails of the forest site to its stoniness. The different effects of hikers, motorcycles and horses is likely due to the increases in plowing action with increasing weight per unit of ground area. Soil compaction Soil compaction increased with increasing numbers of passes, was greater on slopes than on level sites and was generally greatest for horses and least for hikers. Compaction is greatest on slopes because a greater downward force is applied there. Compaction is greater under horses than hikers or cyclists because horses exert the greatest downward pressure on the soil. Upslope v. downslopeffects Only on the meadow trails were upslope and downslop effects compared. In general, motorcycle damage was greatest when traffic was upslope while horse and foot traffic was most damaging when the traffic was downslope (Fig. 2). Motorcycles removed vegetation at simila rates whether going up or downhill; motorcycles moving uphill established a narrow rut which later became wider, shallower, and more like downhill trails. Horses and hikers removed vegetation and soil at greaterates when going downhill than uphill; the contrast with cycles is probably due to the halting action of an animal walking downhill. Downhill hiker trails were slightly wider than uphill trails while downhill horse trails were slightly narrower. Trail management In trail management the following points should be considered. Trampling is damaging to a site. Damage generally increased from hiker to motorcycle horse in our study; if ridden at less conservative speeds, i.e. greater than 20 km h 1, motorcycles might be more damaging than horses. Damage increased as the slope of the site increased; it is doubtful that ascents steeper than 15? should be permitted in natural areas. Motorcycle damage was greatest when going uphill while hiker and horse damage was greatest when going downhill; where a choice is available, wear on trails would be minimized if motorcycles ascended gentle slopes and descended steep slopes while hikers and horses ascended steep slopes and descended gentle slopes. Physical damage is probably least on grassy and stony sites. The quality of a trail depends on its physical condition as well as on factors not considered here, such as noise pollution, air pollution, trailside grazing, manure, etc. Areas will be needed for each mode of transport, since each provides different satisfying experiences. If one area is to be used by hikers, motorcycles and horses, the carrying capacity should be calculated according to the reciprocal of the damage done by each mode of transport. ACKNOWLEDGMENTS Our experiments were supported by National Science Foundation Grant GI 39592.
0040800 T WEAVER AND D. DALE 457 REFERENCES Bayfield, N. (1971). A simple method for detecting variations in walker pressure laterally across paths. Journal of Applied Ecology. 8, 533-6. Bayfield, N. (1973). Use and deterioration of some Scottis hill paths. Journal of Applied Ecology, 10, 635-44. Bell, K. & Bliss, L. (1973). Alpine disturbance studies: Olympic National Park, USA. Biological Conservation, 5, 25-32. Dale, D. & Weaver, T. (1974). Trampling effects on vegetation of the trail corridors of North Rocky Mountain Forests. Journal of Applied Ecology, 11, 767-72. Davidson, E. & Fox, M. (1974). Effects of off-road motorcycle activity on Mojave Desert vegetation and soil. Madrono, 22, 381-90. Daubenmire, R. (1970). Steppe vegetation of Washington. Washington Agricultural Experimental Station Technical Bulletin 62, Pullman. 131 pp. Helgath, S. (1975). Trail deterioration in the Selway-Bitterroot Wilderness. USDA Forest Service Research Note INT 193. Intermtn. Forest and Range Expt. Sta., Ogden, Utah. 15 pp. Liddle, M. (1975). A selective review of the ecological effects of human trampling on natural ecosystems. Biological Conservation, 7, 17-36. Liddle, M. & Greig-Smith, P. (1975). A survey of tracks and paths in a sand dune ecosystem. Journal of Applied Ecology, 12, 893-930. Moroney, M. (1951). Factsfrom Figures. Penguin, Baltimore, Maryland. 472 pp. Rogova, T. (1976). Influence of trampling on vegetation of forest meadow and whortleberry-moss pine forest cenoses. Soviet Journal of Ecology, 7, 356-9. Speight, M. (1973). Outdoor Recreation and its Ecological Effects: a bibliography and review. Discussion Papers in Conservation No. 4, University College, London. 35 pp. Stankey, G. & Lime D. (1973). Recreational carrying capacity: an annotated bibiography. USDA Forest Service General Technical Repts. INT3. Intermtn. Forest and Range Expt. Sta., Ogden, Utah. U.S. Weather Bureau. (1973). Climatological Data for Montana. US Dept. of Commerce, Washington, D.C. Weaver, T. & Dale D. (1974). Pinus albicaulis in central Montana: enviroment, vegetation, and production. American Midland Naturalist, 91, 222-9. (Received 29 July 1977)