ARTICLE IN PRESS
0301-4797/$ - se
doi:10.1016/j.je
�CorrespondE-mail addr
Journal of Environmental Management 84 (2007) 377–383
www.elsevier.com/locate/jenvman
Modeling military trampling effects on glacial soils in the humidcontinental climate of southern New York
LTC Kenneth W. McDonalda,�, LTC Andrew Glenb
aDepartment of Civil and Mechanical Engineering, United States Military Academy, Bldg 752, Mahan Hall, West Point, NY 10996, USAbDepartment of Mathematical Sciences, United States Military Academy, Thayer Hall, West Point, NY 10996, USA
Received 23 July 2004; received in revised form 7 June 2006; accepted 13 June 2006
Available online 2 April 2007
Abstract
The purpose of this research is to create a baseline model of soil compaction response to trampling and a methodology to model the
effects of trampling on soil. Although trampling studies have been conducted in the past, the analysis of military training in part provides
a different perspective and approach. The data showed bulk densities remained relatively constant for a time and then began to increase
at an increasing rate for several hundred passes and finally leveled and remained at or below 1.30 g/cm3 through the remainder of the
experiment. Mathematical models were created based on empirical data from a trampling experiment using a more standard logistical
growth curve as well as curves based on Weibull and gamma cumulative distribution functions (CDFs). The experiment and the resulting
models give quantifiable continuous inference on the effects of trampling, as opposed to the existing qualitative assessments. These
baseline models will be the foundation for future studies of land management when trampling occurs.
r 2006 Elsevier Ltd. All rights reserved.
Keywords: Soil compaction; Math modeling; Trampling; Bulk density; Military training; West Point
1. Introduction
The purpose of this research is to create a baseline modelof soil compaction response to trampling and a methodol-ogy to model the effects of trampling on soil. Surfacedisturbance caused by off-road vehicles and heavy foottraffic often leads to disruption of soil structure, reducedplant cover, degradation of biological and physical soilcrusts, soil compaction, reduced water infiltration, in-creased runoff, and accelerated erosion (Doe, 1993;McCarthy, 1996; Kade and Warren, 2002; Gilewitch,2004). By its nature, military training constitutes anextreme form of land use and degrades training landsusefulness if not maintained. Because of the limitedamount of available training land, maintaining this landoften conflicts with established military training practices.In this construct, the military must ensure it does every-thing possible to protect and maintain its training lands.The impact of military vehicle traffic on soil is well known;
e front matter r 2006 Elsevier Ltd. All rights reserved.
nvman.2006.06.012
ing author. Tel.: +1845 938 6511; fax: +1 845 938 5522.
ess: [email protected] (K.W. McDonald).
however, little is known concerning the impact of militarytrampling on soil. Compaction as a soil condition indicatoris well established. Soil compaction by its nature is a directmeasure of use and can be applied as an indirect measure ofthe impact of military training.The preponderance of research on trampling impact on
soil properties has focused on recreational use and impact.Research into the effect of induced foot traffic onsoil properties is limited. Work by researchers (Lutz,1945; Dotzenko et al., 1967; LaPage, 1962; Monti andMackintosh, 1979; Bryan, 1977) focused on measuringtrampling effects in recreation areas and public forestparks. A subsequent qualitative (low-, moderate-, high-use)labeling approach to describe the impact of trampling onthese areas was developed without definite qualitativeconstraints.In a study by Young and Gilmore (1976) on three
campgrounds in Illinois, they suggest that compaction may‘‘stabilize’’ and not increase with use. This assertion wassupported with evidence that although compaction tendedto be greater at the center of all campsites than at theperimeters, these differences were not statistically different.
ARTICLE IN PRESSK.W. McDonald, A. Glen / Journal of Environmental Management 84 (2007) 377–383378
A similar finding was noted in a study on military campingimpact on soil properties at Fort Leonard Wood, Missouri.Trumball et al. (1994) found that although bulk densitiesincreased at low- and high-use sites compared to undis-turbed sites there was not significant difference betweenlow- and high-use areas. In another military study,Whitecotton et al. (2000) conducted an extensive studyon the impact of military encampment at the US Air ForceAcademy’s Jack Valley Training Area. Their researchfound no significant differences between the bulk densitieson low- and high-use camp sites. In a cattle tramplingstudy, Guthery and Bingham (1996) looked at a theoreticalbasis for management of trampling by cattle. Of importand applicability, their study developed quantitativepredicative models from which they used empirical resultsto assist in developing range management techniques.
In a series of studies, David Cole and others analyzed theeffects of trampling and compaction. In 1992, Cole et al.found that differences in bulk density as a measure ofcompaction were not much different between moderate-and high-use sites. In a series of trampling studiesconducted in 1985, Cole looked at the effects of tramplingon vegetation cover and soil condition and found that themost rapid increase in compaction occurred within the first50–75 passes, and then a decreasing rate occurred up toabout 400 passes until the compaction resistance leveledout. In 1988, Cole concluded a 3-year trampling andrecovery analysis and noted that compaction increasedfrom year 1 to 2, but showed no marked increase incompaction from year 2 to 3. Cole noted this as anunexpected result when considering the trampling rateremained constant and continued throughout the 3-yeartrampling experiment.
Fig. 1. The West Point Militar
Finally several studies have looked at trampling’scompactive effect on vegetation. McNearney et al. (2002)investigated trampling’s effect on the germination andvigor of prairie grass. Their study found a direct correla-tion between foot traffic, compaction and reduced germi-nation and stunted growth of prairie grass. Likewise,Roovers et al. (2004) conducted a 2-year study in thechanges in vegetation structure and found short-termeffects consist of the mechanical deterioration of plantmaterial, whereas long-term effects of trampling includedirect as well as indirect effects on the total system ofvegetation and soil (e.g. compaction on root functions).These results clearly illustrate that vegetation coverdeclines when trampling intensity increases.
2. Methodology
2.1. Study area
The West Point Military Reservation is located on thebanks of the Hudson River in southeast New York,approximately 50 miles north of New York City andencompasses an area of approximately 16,080 acres(cantonment area—2250 acres and training area—13,830acres) (Fig. 1). The training land, southwest of thecantonment area, is characterized by mountainous terrainwith open fields, forests, training areas, weapon ranges,and impact areas. Additionally, several ponds and lakes areused for public recreation areas and reservoirs. The areaincludes the watershed for the West Point MilitaryReservation and the local community of Highland Falls(M. Anderson, 2002, USMA, pers. comm.).
y Reservation, New York.
ARTICLE IN PRESS
N
L1L2
L4L5
L3
L6L7
L8L9
T1
T2
T3 T4
T5 T6T7
T8T9
-Climbing Wall-Hurdle-Parry-Rock Wall-Bridge-Contour-Stream-Bunker-Culvert-Stairs-Low Crawl
230
236
230
Bataan Bayonet Assault Course
Experiment Site
1 : 2,500 ContourInterval 2 meters
Fig. 2. The Bataan Bayonet Assault Course on the West Point Military Reservation, New York.
K.W. McDonald, A. Glen / Journal of Environmental Management 84 (2007) 377–383 379
Precipitation is well distributed throughout the year andis adequate for all vegetation. The total annual precipita-tion is 121.9 cm of which 50% usually occurs in Aprilthrough September. In winter, the average temperature is�1.7 1C and the average daily minimum temperature is�6.1 1C. In summer, the average temperature is 22.8 1C andthe average daily maximum temperature is 28.9 1C (Olsson,1981).
This region is characterized by narrow elevated moun-tainous terrain composed mostly of metamorphic Precam-brian gneiss and granite formed during the Proterozoic era.This provides the basis for the mineral content associatedwith the soil. The predominant soil minerals are vermicu-lite, quartz, chlorite and iron oxide (US Geological Survey,1958; McDonald, 2003).
The soils in this area are Inceptisols specifically definedas Hollis soils having a thin very dark grayish browngravelly fine sandy loam topsoil and yellowish brownsubsoil. These soils form in well drained to somewhatexcessively drained, sloping and gently sloping areas inglacial till deposits derived from crystalline rock (schist,gneiss, and granite). Soil size distribution (45% sand; 40%silt; 15% clay-base on the International Soil ScienceSociety Size definition) was determined using manual sieveanalysis (Automatic Tray Shaker) and laser diffractionparticle size analysis (Beckman Coulter LS Series based onthe Fraunhofer and Mie theories of light diffraction).Gravel percentage was minimal to none. Due to time andavailable equipment, the organic matter content (8.5%)was determined by burning and washing the soil (Olsson,1981; McDonald, 2003).
The area vegetation is mainly deciduous trees withshrubs and low lying vegetation below the tree canopy. Thevegetation cover and ground litter consisted of deciduoustree leaf material, shrubs and vine vegetation. The
thickness of the material was approximately 2.5–5 cmbefore the experiment began.This research is part of a greater research effort to
measure the effects of trampling on the Bataan BayonetAssault Course (BBAC) (Fig. 2). Therefore, an experimentsite was selected southwest of the BBAC. The experimentalsite has similar soil as the BBAC, it is level without adiscernable slope and it has not previously been used fortraining. Additionally, placing the experiment site close tothe BBAC allowed the use of cadets without impacting thetraining. The BBAC consistently has a high degreeof training use. Every year, approximately 1500 cadetstraverse this training course during the summer trainingcycle. The course is rectangular in shape, oriented south-west to northeast (Fig. 2), and approximately 40m wide by300m long. The course consists of nine lanes, each lanecontaining identical ‘‘stationary obstacles’’ that a soldiermight commonly encounter in a wartime environment.Examples of the stationary obstacles include: ladder walls,parry structures, bunkers, and culverts (Fig. 2). All lanesare identical and all obstacles are positioned uniformlythroughout the course.
2.2. Laboratory analysis
The Standard Proctor test (hand method) was used togenerate a compaction curve for the soil sample at theBBAC test site. The Standard Proctor test is a geotechnicalengineering test that consists of compacting soil in a15.2 cm diameter mold by dropping a 2.5 kg hammer ontothe sample from a height of 30.5 cm replicating thecompactive effort of 600 kN-m/m3. A moisture-densitycurve is developed for determining the maximum compac-tion of a particular soil at the optimum moisture contentbased on the applied compactive effort. The test was
ARTICLE IN PRESSK.W. McDonald, A. Glen / Journal of Environmental Management 84 (2007) 377–383380
adjusted to reflect the compactive effort of the averageweight and shoe size of a cadet. The impulse momentumtheorem was used to determine the compactive effort of theaverage cadet (1624N). Applying this force to the forceequation ðmv ¼ FtÞ the fall height for the compactionhammer was determined to be 20.8 cm. Likewise, the effectof vegetation was replicated by using vegetation as acushioning layer during the compaction test and a finalcurve was generated for a total of three curves (Fig. 3).
2.3. Field methods
The trampling experiment conducted at the BBACfollowed a modified version of the standardized proceduresdeveloped by Cole and Bayfield (1993). Cadets arrive at theBBAC for initial instruction prior to negotiating thecourse. After the initial instruction, cadets walk throughthe experiment site and traverse a path approximately 1mwide by 10m long to get to their particular starting points.Cadets continued this procedure throughout the summertraining cycle for a total of 3088 passes. Bulk densitymeasurements were recorded approximately every 100–150passes. A total of 22 soil samples were taken periodically atthe end of each training event using an ELE Internationalhammer-driven core sampler: 7.62 cm diameter densitytube with a 2.83� 10�4m3 capacity. The depth of samplewas 7.62 cm and included the surface layer material.Sample holes were not backfilled but were marked withred flags along the path to ensure subsequent sampleswould not be taken from the same location. Samples weretaken to the laboratory, weighed, dried at 107 1C for 24 h,reweighed and bulk density determined. The average
1.15
15 20 25
Dry
Uni
t Wei
ght (
gm/c
m3 )
Compact
1.40
1.35
1.30
1.25
1.20
1.10
1.05
1.00
Standard 20.8 cm w/H
Moistur
Fig. 3. Standard Proctor test with modification to reflect actual impact in fi
moisture content was 32%. Bulk densities were plotted tocorresponding number of passes.
2.4. Numerical modeling
The modeling phase combined laboratory experimentaldata and the field data into mathematical models toreplicate the trampling response curve. The general shapeof the soil bulk density response curve is an increasingfunction that is bounded below at an initial bulk densityand above at a maximum bulk density based on a specifiedapplied load. Cumulative distribution functions (CDFs)from probability and statistics theory also carry the sameabove and below bounding characteristics (Devore, 2000).Two CDFs, the Weibull and the Gamma, can be used toportray a soil’s response to trampling. By shifting theCDFs, models were developed to replicate the soil’sresponse to trampling. Additionally, a logistical growthcurve can also replicate this type of bounding characteristicand, by adjusting the rate of growth; a soil-response curvecan replicate the trampling effect on the soil (Polking et al.,2002). These three approaches were adjusted to fit thenatural curve obtained at the BBAC and compared todetermine the ‘‘best fit’’ model.Laboratory experiments and the field collection data
mentioned earlier were required to determine the boundedlimits for the soil bulk density. Combining the data fromthe laboratory experiments (maximum density) with thefield data (minimum density and natural curve), thebounded upper and lower portions of the curve wereestablished and applied to the three mathematical functionsto identify an accurate model for soil response to
30 35 40 45
ion Test
eel 20.8 cm w/Heel and Vegetation
e Content (%)
eld conditions to determine maximum bulk density expected in the field.
ARTICLE IN PRESSK.W. McDonald, A. Glen / Journal of Environmental Management 84 (2007) 377–383 381
trampling. With these two bounding values, the threemodels were populated and compared statistically to theactual curve generated from the foot traffic experiment.
3. Results and discussion
3.1. Moisture density curves
The manual Standard Proctor test was modified twice tosimulate the compaction effort of cadet trampling and theeffect of vegetation (Fig. 3). There is a significant differencebetween the curve generated by the Standard Proctor andthe two curves generated by the modified Proctor. Themodified (heel) curve proved not statistically different thanthe second modified (heel-vegetation) curve. The maximumcompaction for the modified (heel) compaction test wasapproximately 1.30 g/cm3. This maximum compaction isthe upper bounding layer of the modeling portion of theanalysis. This maximum bulk density was also predicted asthe upper bounding limit in the trampling experiment. Theupper bounding limit is the expected ‘‘stabilized’’ bulkdensity referenced in previous research efforts.
3.2. Soil bulk density
The experiment revealed measurable changes in soil bulkdensity in response to trampling. Initial bulk density wasrecorded prior to the trampling experiment and as bulkdensity was plotted against the number of soldier passes, avery distinctive response was noticed (Fig. 4). Initial bulkdensity values remained relatively stable until 700–800passes. At 700–800 passes, soil compaction began toincrease at an increasing rate and continued to increasethrough 900 and 1400 passes but at a decreasing rate. After1400 passes, bulk density began to ‘‘stabilize,’’ remainingbelow 1.30 g/cm3 with the exception of one outlier. Duringthe traffic experiment, loss of vegetative cover and groundlitter occurred within the first 400–500 passes. The results
0 200 400 1000 1200 1400
Bul
k D
ensi
ty(g
m/c
m3 )
Soil Respon1.35
1.30
1.25
1.20
1.15
1.10
1.05
1.00
600 800
Numbe
Fig. 4. Induced foot traffic experiment at the BBAC, with opti
of the bulk density measurements indicate a distinct initialincrease followed by stabilization in bulk density as cadetsprogressed through the training cycle. Fig. 4 also shows themaximum bulk density (1.30 g/cm3) the BBAC shouldexperience based on the compactive effort cadets exert onthe soil.
3.3. Numerical modeling
Applying the results of the geotechnical engineeringlaboratory experiments and the field experiment to thethree numerical models (the logistical growth, the Gammaand the Weibull) produced distinctive curves (Fig. 5). Aregression analysis produced the following R2 values foreach model: Logistical growth ¼ 0.85, Gamma ¼ 0.95 andWeibull ¼ 0.94. The corresponding R2 values indicate theGamma and Weibull models fit slightly better than thelogistical growth model, accounting for approximately95% of the variation observed in the test data.
4. Discussion and conclusions
The traffic experiment revealed measurable changes insoil bulk density in response to trampling (Fig. 4). Theinitial 700 passes had minimal effect on soil compactionand remained relatively stable. At 700–800 passes, soilcompaction occurred at an exponential rate. Soil compac-tion continued to increase through 1400 passes and thenlevel off. A possible explanation for this bulk densitychange after 700 passes is vegetation cover loss as reportedin previous research (Fig. 6). During the foot trafficexperiment, the visible vegetation layer disappeared atapproximately 400–500 passes. The measurement ofvegetation loss was subjective in nature and was notmeasured. The organic layer and vegetation may act as acushion to lessen the impact of the traffic and assist inpreventing a quick compaction of the soil. Although soilloss was not measured, visual inspection indicates that
1600 1800 2000 2200 2400 2600 2800 3000
se to Foot Traffic
r of Passes
3200
mum compaction identified by the solid line at 1.30 g/cm3.
ARTICLE IN PRESS
0 500 1000 1500 2000 2500 3000 3500
Bul
k D
ensi
ty (
gm/c
m3 )
BBAC Data Logistical Growth Model Gamma Model Weibull Model
Passes
1.35
1.30
1.25
1.20
1.15
1.10
1.05
1.00
Model Comparison
Fig. 5. Comparison of soil bulk density measurements at the BBAC with data from the numerical models.
Fig. 6. Before and after photos of experiment site.
K.W. McDonald, A. Glen / Journal of Environmental Management 84 (2007) 377–383382
there was minimal soil loss, that which can be attributed to‘‘carry-off’’ from boots worn by cadets.
The prediction that a maximum bulk density could bedetermined through a modified Standard Proctor testproved accurate in this study. The trampling experimentresults supported the predicted ‘‘stabilizing’’ of the soilbulk density at 1.30 g/cm3.
An area lacking in this research is the effect of soil-moisture and vegetative cover on bulk density results. Soil-moisture levels have direct effects on bulk densities butwere not the focus of this research. Additionally, vegeta-tion cover levels although observed were measuredquantitatively.
The modeling results provide a good foundation forfuture research. A visual inspection of the results (Fig. 5)shows the foot traffic experiment data displaying adistinctive exponential growth curve, which generally canbe duplicated. The same procedures (modified Proctor Testand trampling experiment) can be used to generatecompaction curves, although uniquely different (differentgrowth slope, initial density and maximum density), inother soils. The curves generated from the models illustratetheir ability to replicate the observed results and have theflexibility to be applied to different soils.Finally, the purpose of this particular research is to
create a baseline model of soil response to trampling so
ARTICLE IN PRESSK.W. McDonald, A. Glen / Journal of Environmental Management 84 (2007) 377–383 383
that in subsequent years ‘‘recovery time’’ of soil can beidentified and the apparent land management implicationsand possible suggestions can be addressed. Our intent is tocompare the models of recovered soil and see if they onlydiffer on the ‘‘time-shift’’ parameter of original model.
References
Bryan, R.B., 1977. The influence of soil properties on degradation of
mountain hiking trails at Grovelsjon. Geografiska Annaler Series
A-Physical Geography 59 (1/2), 49–65.
Cole, D.N., 1985. Recreational trampling effects on six habitat types in
Western Montana, United States Department of Agriculture, Forest
Service Research Paper INT-350, Intermountain Forest and Range
Experiment Station, Ogden, UT.
Cole, D.N., 1988. Disturbance and recovery of trampled Montana
grassland and forests in Montana, United States Department of
Agriculture, Forest Service Research Paper INT-389, Intermountain
Forest and Range Experiment Station, Ogden, UT.
Cole, D.N., Bayfield, N.G., 1993. Recreational trampling of vegetation:
standard experimental procedures. Biological Conservation 63,
209–215.
Cole, D.N., Hall, T.E., 1992. Trends in Campsite Condition: eagle Camp
Wilderness, Bob Marshall Wilderness, and Grand Canyon National
Park, United States Department of Agriculture, Forest Service
Research Paper INT-453, Intermountain Forest and Range Experi-
ment Station, Ogden, UT.
Devore, J.L., 2000. Probability and Statistics for Engineering and the
Sciences. Duxbury Press, Pacific Grove, CA.
Doe III, W.W., 1993. Simulation of the spatial and temporal effects of
maneuvers on watershed response (spatial effects). Unpublished Ph.D.
Dissertation. University of Colorado, Boulder, CO.
Dotzenko, A.D., Papmichos, N.T., Romine, D.S., 1967. Effects of
recreational use on soil and moisture conditions in Rocky
Mountain National Park. Journal of Soil and Water Conservation
22, 196–197.
Gilewitch, D., 2004. The effect of military operations on desert pavement:
case study from Butler Pass, AZ. In: Caldwell, D.R., Ehlen, J.,
Harmon, R.S. (Eds.), Studies in Military Geography and Geology.
Kluwer Academic Publishers, Dordrecht, The Netherlands.
Guthery, F.S., Bingham, R.L., 1996. A theoretical basis for study and
management of trampling by cattle. Journal of Range Management 49,
264–269.
Kade, A., Warren, S.D., 2002. Soil and plant recovery after historic
military disturbances in the Sonoran Desert, USA. Arid Land
Research Management 16, 231–243.
LaPage, W.F., 1962. Recreation and the forest site. Journal of Forestry 60,
319–321.
Lutz, H.J., 1945. Soil conditions on picnic grounds in public forest parks.
Journal of Forestry 43, 121–127.
McCarthy, L.E., 1996. Impact of military maneuvers on Mojave Desert
surfaces: a multi-scale analysis. Unpublished Ph.D. Dissertation.
University of Arizona, Tucson, AZ.
McDonald, K., 2003. Military foot traffic impact on soil compaction
properties at the United States Military Academy. Unpublished Ph.D.
Dissertation. University of Missouri-Rolla, Rolla, MO.
McNearney, P., Riley, J., Wennersten, A., 2002. Trampling increases soil
compaction; soil compaction depresses vigor of Andropogon gerardii.
Tillers, vol. 3. Grinnell College, Grinnell, IA, pp. 25–28.
Monti, P.W., Mackintosh, E.E., 1979. Effect of camping on surface soil
properties in the boreal forest region of northwestern Ontario,
Canada. Soil Science Society of America Journal 43, 1024–1029.
Olsson, K.S., 1981. Soil Survey of Orange County, New York. Soil
Conservation Service, US Department of Agriculture.
Polking, J., Boggess, A., Arnold, D., 2002. Differential Equations
with Boundary Value Properties. Prentice-Hall Inc., Upper Saddle
River, NJ.
Roovers, P., Verheyen, K., Hermy, M., Gulinck, H., 2004. Experimental
trampling and vegetation recovery in some forest and heathland
communities. Applied Vegetation Science 7, 111–118.
Trumbull, V.L., Dubois, P.C., Brozka, R.J., Guyette, R., 1994. Military
camping impacts on vegetation and soils of the Ozark Plateau. Journal
of Environmental Management 40 (4), 329–339.
US Geological Survey, 1958. Military Geology of the West Point Area,
New York. Military Geology Branch, Engineer Intelligence Division,
Office of the Chief of Engineers, Department of the Army,
Washington, DC.
Whitecotton, R.C.A., David, M.B., Darmody, R.G., Price, D.L., 2000.
Impact of foot traffic from military training on soil and vegetation
properties. Environmental Management 26 (6), 697–706.
Young, R.A., Gilmore, A.R., 1976. Effects of various camping intensities
on soil properties in Illinois campground. Soil Science Society of
America Journal 40, 908–911.
