MODELING WATER QUALITY IMPACTS OF OFF-ROAD
VEHICLE?S IN FORESTED WATERSHEDS
Except where reference is made to the work of others, the work described in this
thesis is my own or was done in collaboration with my advisory committee.
Christian John Brodbeck
Certificate of Approval:
Timothy P. McDonald Dan A. Brown
Co-Chair Co-Chair
Associate Professor Gottlieb Associate Professor
Biosystems Engineering Civil Engineering
John P. Fulton Puneet Srivastava
Assistant Professor Assistant Professor
Biosystems Engineering Biosystems Engineering
Kathryn Flynn Prabhakar Clement
Associate Professor Associate Professor
School of Forestry and Wildlife Sciences Civil Engineering
Stephen L. McFarland
Acting Associate Provost and Dean
Graduate School
MODELING WATER QUALITY IMPACTS OF OFF-ROAD
VEHICLES IN FORESTED WATERSHEDS
Christian John Brodbeck
A Thesis
Submitted to
The Graduate Faculty of
Auburn University
in Partial Fulfillment of the
Requirements for the
Degree of
Master of Science
Auburn, Alabama
August 8, 2005
iii
MODELING WATER QUALITY IMPACTS OF OFF-ROAD
VEHICLES IN FORESTED WATERSHEDS
Christian John Brodbeck
Permission is granted to Auburn University to make copies of this thesis at its discretion,
upon request of individuals or institutions and at their expense. The author reserves all
publication rights.
Signature of Author
Date
iv
VITA
Christian John Brodbeck, son of Martin and Diane Brodbeck, was born on
September 7, 1979 in Reutalhuleu, Guatemala. He grew up in Santa Maria de Jesus,
Guatemala, with his brothers Beau and Michael. After graduating from the Inter-
American School in June of 1997, he entered Faulkner State Community College in Bay
Minnette, Alabama. In June of 1999 he entered Auburn University in Auburn, Alabama.
He graduated cum laude with a Bachelors of Biosystems Engineering on December 15,
2002. He then entered Auburn University?s graduate program in Civil Engineering in
August of 2003, and worked as a graduate research assistant and graduate teaching
assistant in the Biosystems Engineering Department. He is currently employed as a
Research Associate in the Biosystems Engineering department of Auburn University.
v
THESIS ABSTRACT
MODELING WATER QUALITY IMPACTS OF OFF-ROAD VEHICLES IN
FORESTED WATERSHEDS
Christian John Brodbeck
Master of Science, August 8, 2005
(B.B.S.E., Auburn University, 2002)
140 Typed Pages
Directed by Timothy P. McDonald and Dan A. Brown
Erosion from off-road vehicles can cause negative effects on water quality by
impairing fish habitat and shortening reservoir life. Due to an increasing level of off-road
vehicle use throughout the country, this impact has risen to a level that has become a
cause for concern. Interest has been raised by the USDA Forest Service to quantify
sediment loads and determine management practices that may aid in reducing the current
sediment delivery rates.
A bridged stream trail crossing on the Kentuck ORV trail system in the Talladega
National Forest was equipped with water sampling equipment to measure total suspended
sediment and flow rates. Equipment was also installed to measure rainfall and traffic
volumes. From this data, sediment loads were calculated and used to calibrate the Water
vi
Erosion Prediction Project (WEPP) model. The model was then utilized to simulate
sediment yields for varying management practices.
A total sediment load of 120.9 kg was calculated for the entire data collection
period. Sediment yield proved to be only significant from storm events that had a one
year return interval or longer. During peak season, traffic volumes reached 180 passes
per day with an average throughout the riding season of 25 passes per day. During
calibration of the WEPP model, a Nash-Sutcliffe R? of 0.92 was achieved. Using the
WEPP model, it was determined that in order to achieve target sediment loads,
management practices should have a minimum forest buffer length of 20 m with a
minimum water bar spacing of 6 m for slopes between 13 and 20%. The use of proper
BMPs , such as water bar spacing, slope grade, and buffer lengths, can aid in minimizing
the degradation of water quality.
vii
ACKNOWLEDGEMENTS
The author would like to thank his parents for their economic and moral support
throughout his educational career. The author would also like to thank Dr. Tim
McDonald for providing the opportunity for this project and also for his continuing
support and encouragement through it all. The author would also like to thank Dr. John
Fulton and Dr. Puneet Srivastava for their guidance in accomplishing various tasks
throughout the project. The author would also like to recognize fellow students Layne
Owen and Corey Kichler who helped with the collecting and processing of field data.
The author would also like to thank Dennis Block and Upton Hatch of the Auburn
University Environmental Institute for the funding to conduct the research project.
Finally and most importantly, the author would like to thank God for everything He has
given us and done for us.
viii
Style manual used Auburn University Graduate School Guide to Preparation and
Submission of Theses and Dissertations
Computer software used Microsoft Office 2000 --- Microsoft Word 2000, Microsoft
Excel 2000 ---- Water Erosion Prediction Project (WEPP) model
ix
TABLE OF CONTENTS
LIST OF FIGURES ?????????????????????????...xii
LIST OF TABLES ?????????????????????????..xiv
INTRODUCTION ??????????????????????????.1
Problem Statement ???????????????????????.1
Objectives ??????????????????????????.5
REVIEW OF LITERATURE ???????????????????????.6
Introduction ??????????????????????????.6
Soil Erosion Amendments ????????????????????11
Erosion Modeling ???????????????????????15
Summary ??????????????????????????22
RESEARCH PROCEDURES ???????????????????????24
Introduction ??????????????????????????24
Site Description ???????????????????????24
Equipment Description ????????????????????27
Stream Water Sampling ?????????????????27
Stream Flow Rate ????????????????????28
Rainfall Measurement ????????????????????29
Traffic Measurement ????????????????????30
Laboratory Analysis ????????????????????32
Soil Amendments ???????????????????????33
Control Plots ???????????????????????33
Treated Plots ???????????????????????34
Data Analysis ???????????????????????37
x
Erosion Prediction Modeling ????????????????????37
Watershed and Subcatchment Delineation ???????????38
Weather Generator ????????????????????39
Subcatchment Modification and Model Calibration ????????40
Management Simulations ?????????????????45
Statistical Analysis ????????????????????45
RESULTS AND DISCUSSION ????????????????????46
Stream Water Sampling ????????????????????46
Trail Closed Period ????????????????????46
Trail Maintenance Period ?????????????????51
Trail Open Period ????????????????????55
Overall Sediment Loading ?????????????????60
Modeling ??????????????????????????64
Weather Generation ????????????????????65
Parameter Calculation and Model Calibration ????????66
Development of Best Management Practices ???????????70
Water Bar Spacing and Slope Gradient ????????71
Buffer Length ????????????????????72
Soil Amendment Assessment ????????????????????75
SUMMARY AND CONCLUSIONS ????????????????????81
Introduction ??????????????????????????81
Objective 1 ??????????????????????????81
Objective 2 ??????????????????????????82
Objective 3 ??????????????????????????83
Objective 4 ??????????????????????????83
General Conclusion ???????????????????????84
Future Research ???????????????????????84
BIBLIOGRAPHY ??????????????????????????86
xi
APPENDICES ??????????????????????????90
Appendix A: Trail Closed Period ?????????????????90
Appendix B: Maintenance Period ?????????????????97
Appendix C: Trail Open Period ????????????????..104
Appendix D: Trafficking Data ????????????????..119
Appendix E: Cumulative Rainfall and Sediment Loading ???????..121
Appendix F: Soil Amendment Assessment ?????????????..122
xii
LIST OF FIGURES
Figure Page
3.1 Site Location ??????????????????????????25
3.2 Blue Trail with Study Site ????????????????????26
3.3 24 1 Liter ISCO Bottles ????????????????????28
3.4 Stream Profile ??????????????????????????29
3.5 Initial Video Housing Setup ????????????????????31
3.6 View from Traffic Monitoring Station ??????????????31
3.7 Control Plot White-A in Ideal Condition ??????????????34
3.8 Application of Envirotac II ????????????????????35
3.9 Plot Soaked in Envirotac II ????????????????????36
3.10 Compaction of Treated Plot ????????????????????36
3.11 Watershed and Subcatchments ?????????????????39
3.12 Insloped Trail Leading into Stream ?????????????????42
3.13 Insloped Trail with 20m Forest Buffer ??????????????42
3.14 Outsloped Trail with Forest Buffer ?????????????????43
4.1 Cumulative Rainfall for February 12, 2004 Storm Event ????????47
4.2 Hydrograph for February 12, 2004 Storm Event ???????????48
4.3 TSS for February 12, 2004 Storm Event ??????????????48
4.4 Sediment Load for February 12, 2004 Storm Event ???????????49
4.5 Cumulative Rainfall for March 6, 2004 Storm Event ???????????52
xiii
4.6 Hydrograph for March 6, 2004 Storm Event ??????????????52
4.7 TSS for March 6, 2004 Storm Event ?????????????????53
4.8 Figure 4.4 ? Sediment Load for March 6, 2004 Storm Event ???????....53
4.9 Cumulative Rainfall for April 30, 2004 Storm Event ???????????56
4.10 Hydrograph for April 30, 2004 Storm Event ??????????????56
4.11 TSS for April 30, 2004 Storm Event ?????????????????57
4.12 Figure 4.4 ? Sediment Load for April 30, 2004 Storm Event???????....57
4.13 Traffic Volumes for April 2004 ?????????????????59
4.14 Average Difference in Sediment Concentrations for Categorized Storm
Events ?????????????????????????????61
4.15 Average Flowrate for Categorized Storm Events ???????????61
4.16 Relationship between Cumulative Rainfall and Sediment Load ?????64
4.17 All Storms and Storms with Sediment Data ??????????????65
4.18 All Measured vs. Predicted Sediment Load Values ???????????68
4.19 Measured vs. Predicted Values for Smaller Storms ???????????69
4.20 Annual Average Soil Loss and Yield with 10 m Forest Buffer ?????74
4.21 Annual Average Soil Loss vs. Forest Buffer Length ???????????75
4.22 Cross-section Profiles of Upper Section of Control Plot A ????????77
4.23 Cross-section Profiles of Lower Section of Control Plot A ????????77
4.24 Cross-section Profiles of Upper Section of Treatment Plot A ?????78
4.25 Cross-section Profiles of Lower Section of Treatment Plot A ?????78
xiv
LIST OF TABLES
Table Page
4.1 Sediment Load Summary of 12-Feb-04 Storm and the Closed Period ??50
4.2 Sediment Load Summary for Storm Event on March 6, 2004 and Trail
Maintenance Period ???????????????????????54
4.3 Sediment Load Summary for April 30, 2004 Storm Event and the Trail
Open Period ??????????????????????????58
4.4 Daily Traffic Totals and Averages for Trail Open Period ????????60
4.5 Sediment Load for Cumulative Rainfall Categories ???????????62
4.6 Sediment Load for Each Period ?????????????????63
4.7 Portion of Modified Weather File ?????????????????66
4.8 Calculated Soil Parameters ????????????????????67
4.9 Typical Soil Parameter Ranges ?????????????????70
4.9 Recommended Water Bar Spacing with a 20 m Forest Buffer ?????72
4.10 Recommended Water Bar Spacing with a 10.67 m Forest Buffer ?????73
4.11 Net and Average Soil Loss or Deposition for Each Section ????????80
1
INTRODUCTION
Problem Statement
The use of Off-Road Vehicle (ORV) trails as a form of outdoor recreation is
rapidly becoming more popular and is one of the fastest growing forms of recreation in
the United States. All-terrain vehicles (ATV?s) and 2-wheel motorcycles provide
entertainment for outdoor enthusiasts seeking the thrill and excitement of challenging
trails not traversable by ordinary vehicles. The use of ORV?s has increased dramatically
in the last three decades. According to the USDA Forest Service, the use of ORV?s has
increased from 5 million in 1972 to 36 million in 2002. In 2003, 5 percent of the visitors
to National Forests and Grasslands consisted of ORV users (USDA Forest Service,
Office of Communication). An increase in use, such as this, leads to the need for
management in order to protect the land and its natural resources for the benefit of all
users.
In 2003, the USDA Forest Service identified four threats to the Nation?s Forests
and Grasslands. The threats were as follows: fuel and fire, invasive species,
fragmentation, and unmanaged recreation (USDA Forest Service, Office of
Communication). Since ORV trails fit under the category of unmanaged recreation,
significant strategic changes must take place with the intent of refocusing attention. Due
to the identification of these threats, the USDA Forest Service has changed its approach
on some management issues, but their mission remains the same, ?to sustain the health,
2
diversity, productivity of the Nation?s forests and grasslands to meet the needs of present
and future generations? (USDA Forest Service, Office of Communication).
ORV trails that are unmanaged have the potential to have various adverse impacts
on the environment. Factors that affect the level of degradation include soil erosion
potential, terrain, and type of vegetation. The two most common types of adverse
impacts are severe soil erosion and damage of riparian areas and species. Through site
specific management and maintenance techniques, these impacts may be greatly reduced
to meet the Forest Service?s mission of sustainability.
The delivery of sediment into a stream system is always a cause of concern due to
the environmental impact that it may cause. In areas where salmonid species are present,
stream bottoms covered in gravel are required for spawning, so sedimentation can be a
serious problem. Also, water supply systems and reservoirs dependent on quality water
from surface sources are also affected due to a shortened life (Elliot et. al. 1999).
Because these types of impacts may occur, managing a trail system to reduce erosion
becomes a very important, but challenging task
Site-specific management of ORV trails is applicable all over the nation, but of
particular interest is the Kentuck ORV trail system, located in the Talladega National
Forest of Alabama. The Kentuck ORV is of particular interest because of some concerns
occurring downstream in relation to water quality degradation. The first step in trail
management is proper trail location, layout, and construction. According to Strom and
Wilkins (1990) a list of criteria were set forth before construction could begin on the trail
with the intention of reducing any environmental impacts, in particular erosion and water
quality degradation, that the trail may cause. This list of criteria included the following:
3
riding loops would be created using existing closed roads and newly constructed trails,
archeological survey needed to be completed prior to any soil disturbance, highly
erodible soils were to be evaluated by soil scientists and proper actions taken, degradation
of water quality would be avoided by use of proper stream crossings, avoid areas for
wildlife habitat improvement, areas known to have red-cockaded wood pecker were also
avoided, and timber management will be modified as is deemed appropriate. Following
this list of criteria during trail construction helped reduce some adverse impacts an ORV
trail produces, but did not eliminate them.
Although the Forest Service has attempted many different maintenance
techniques, problems still arise that are not only causing some environmental impacts, but
are economically costly as well. The foremost problem is erosion which is magnified by
puddling of water and the formation of ruts. On the Kentuck ORV, the use of water bars,
broad based dips, and water turn-outs have been used in an attempt to shed water from
the trails. Some maintenance techniques used, with the intent of increasing infiltration
rates, involved the use of gravel and ?geoblock? on the running surface. Due to high
traffic volumes during both wet and dry conditions, elevated levels of tire slip, and the
presence of steep slopes, maintenance practices and techniques are short lived. These
conditions also lead to the formation of ruts. The formation of ruts disregards the
usefulness any maintenance technique may have had as well as making conditions unsafe
for riders. The presence of ruts also magnifies erosion levels because of increased energy
as a result of channelized flow. Effective maintenance techniques for use, particularly on
ORV trails, are necessary because of current techniques not withstanding the wear and
tear caused by trafficking.
4
The development of new maintenance techniques requires background
information on which to base management and maintenance decisions. Since
documentation of sediment yield and delivery from ORV trails in the Eastern United
States is limited, the collection of field data is necessary. The most obvious location for
data collection to determine sediment delivery would be at stream crossings. Collecting
water samples for lab analysis of total suspended solids (TSS) levels at the stream
crossing would allow for the documentation of sediment delivery rates as a result of ORV
trail use. As a manager, knowing the levels of sediment yield and delivery are very
useful in determining the effectiveness of certain maintenance practices.
Another tool that proves to be useful in making management decisions is the
ability to predict the levels of soil yield and delivery. The ability to predict erosion
accurately from a trail system, or particular sections of a trail system, allows the manager
to simulate maintenance techniques or trail rerouting. Basing management decisions on
erosion prediction simulations generally will lead to a better outcome that will aide in
reducing environmental impact due to sediment production and delivery. Another benefit
of modeling sediment yield is, that by predicting erosion, the placement of various
management and maintenance techniques and designs can be better cited. Sediment yield
and delivery modeling can have numerous applications that may be utilized by managers
and road designers for the layout and upkeep of low-volume roads or trails.
The use of chemical amendments applied to the soil for the maintaining of ORV
trails is an area that has limited research but is believed to be a viable method for the
reduction of maintenance. The use of the three soil amendments for trail stabilization
was studied by Davis (Unpublished thesis. 2004). Davis (2004) found, through traffic
5
and rainfall simulation, that the soil amendment known as Envirotac? had superior
performance compared to lignin and control plots. Davis (2004) also found that
Envirotac? was useful in reducing rutting and TSS levels. The use of Envirotac? still
requires additional research to determine its applicability and usefulness in conditions
that involve extended wet periods along with high traffic levels.
The need for research on ORV trails and their impact on water quality in the
Eastern U.S. is the reason for conducting this project. Design and maintenance
techniques currently used on trails tend to follow those set forth for low-volume roads but
are not always applicable. Due to the increase of ORV use, the need for management of
ORV trail systems has become essential. The goal of this project is to develop Best
Management Practices for Off-Road Vehicle trails. The goal of this project will be
accomplished by addressing the following objectives.
Objectives
1. Quantify sediment delivery at a stream crossing and meter traffic volume
levels from an Off-Road Vehicle trail system.
2. Calibrate the WEPP model using field data to simulate and predict sediment
yield and delivery for ORV trails.
3. Use calibrated WEPP model to recommend Best Management Practices for
ORV trails.
4. Conduct an assessment on the application and testing of the soil amendment,
known as Envirotac?, to directly reduce trail maintenance and indirectly
reduce erosion.
6
REVIEW OF LITERATURE
Introduction
Studies involving erosion from low-volume roads have been conducted
throughout the entire United States, whereas, studies involving the effects of Off-Road
Vehicles (ORV?s) have been conducted mainly in the Western United States. The
principles and theories resulting from multiple research projects are applicable to the
Southeastern United States, but due to soil variation is difficult to account for in
determining if similar results would be attained in different locations across the country.
ORV?s affect many aspects of the environment including infiltration rate and sediment
production (Eckert, et. al. 1979), damage to soils and vegetation (Sparrow, et. al. 1978),
soil compaction and trail width (Weaver and Dale, 1978) and water pollution. Water
pollution is generally classified as point source, arising from a distinct outlet, or non-
point source, arriving in streams from diffuse areas. Point source pollution is easier to
identify whereas non-point source pollution is not. Land activities that create runoff are a
source for water contamination. Non-point source pollution has many ?sources?,
including agriculture, mining, forestry operations, landfills, and runoff (Liban, 1998).
Erosion from roads and road construction has been shown to yield 95 tons/hectare/year
(Brooks et al., 2003) of sediment. Erosion from roads and trails may be reduced through
the use of soil amendments.
7
Weaver and Dale (1978) conducted a study on the effects of hikers, motorcycles,
and horses in forests and meadows. They noted that the trail width increased linearly
with increasing roughness, wetness, slope, and number of users. Sparrow et. al. (1978)
noted that trail width increased in relation to wetness and that in areas where soils
became easily saturated, water would pond resulting in the formation of a quagmire.
Rider attempts to circumvent these wet areas, led to a gradual widening of the trail. In
Weaver and Dale?s (1978) study, they also discovered that on level ground, horses were
more destructive than hikers or motorcycles, but this did not hold true on steeper slopes.
Damage during uphill climbs was much higher for motorcycles than horses or hikers. In
general, when the motorcycles were ridden at conservative speeds (less than 20 kph),
they caused more damage than hikers, but less than horses. Problems occur when
outdoor enthusiasts enjoy riding on steep slopes at non-conservative speeds causing trail
damage which can lead to escalated erosion levels.
A study conducted by Eckert et al. (1979) draws the same conclusions found by
Elliot et al. (1999). In their paper titled ?Impacts of Off-Road Vehicles on Infiltration
and Sediment Production of Two Desert Soils?, they simulated traffic and rainfall on sites
in southern Nevada. Traffic simulation was conducted using motorcycles and 4-wheel
drive trucks traveling at 30 kph. Motorcycles were trail-bikes, with knobby tires and
weighing approximately 155 kg, and trucks were ? ton pickup operated in 4-wheel drive.
Results indicated that infiltration and sediment production was mostly due to the type of
surface soil. Infiltration rates, or hydraulic conductivities, for coppice soils were 3 to 13
times greater than that of interspace soils. Alternatively, sediment production for
interspace soils were10 to 20 times greater than coppice soils. Coppice soils consist of a
8
rocky structure that is well-aggregated and rapidly transmits water. With interspace soils,
not only is the structure weak and unstable, saturation occurs quickly which leads to the
susceptibility of particle dispersion in the form of runoff. Elliot et al. (1999) stated that
erosion rates were affected by hydraulic conductivity and soil erodibility and it backs the
conclusions reported by Eckert et al. (1979). Soils with higher hydraulic conductivities
will be able to drain water at a higher rate, leading to reduced runoff, which in turn
reduces sediment production. Eckert et al. (1999) discovered that after ORV traffic, soil
properties are altered, leading to varying results. Negative results due to vehicular traffic
include shear damage, fine soil material being powdered and compaction. In general,
Eckert et al. (1999) found that infiltration rates were reduced and sediment production
increased after soil disturbance due to vehicular traffic.
Other effects of vehicular traffic are soil shearing and compaction (Eckert et al.,
1979) resulting in soil damage and destruction of vegetation (Sparrow et al., 1978). The
shearing of soil is a physical disturbance that causes an increase in both water and wind
erosion in more arid regions. In the more arid regions, shear damage causes the
protective soil pavement to be destroyed, powdering fine soils, and finally filling in
cracks in the surface polygons (Eckert et al., 1979). Eckert et al. (1979) also discussed
the role that motorcycle and other vehicular traffic have in soil compaction. Several
studies (Lull, 1959; Wilshire and Nakata, 1976; Davidson and Fox, 1974) demonstrated
how compaction caused a decrease in pore space and an increase in bulk density. In
these studies, intense motorcycle use as well as traffic by other off-road vehicles was
directly correlated to increased soil compaction. A study conducted by Sparrow et al.
(1978) in Alaska investigated the effect that ORV?s had on soils and vegetation. They
9
discussed that the degree of vegetative destruction was related to the reported traffic
volumes. Soil impact was influenced mainly by soil depth and drainage. The presence of
gravelly or cobbly soils, whether shallow or deep, were less susceptible to erosion than
soils that were deep, but gravel free. Other studies (Egan, 1999; Patric, 1978; Brinker,
1995) emphasize the findings of Sparrow et al. (1978), regarding the negative impacts
that wet soil conditions have on roads and trails.
Researchers agree that it is best to avoid wet areas and/or areas containing steep
terrain for the location of low-volume roads or trails because of the increased probability
of runoff and sediment production. Sometimes the designer does not have the liberty to
avoid such areas, so specific design and maintenance techniques can be followed to help
reduce the level of sedimentation and therefore potential negative environmental impacts.
Numerous studies have been conducted through the years on the measuring of
sedimentation from forest roads. Several different types of road designs exist that aid in
reducing the levels of sediment transport. Croke and Hairsine (2001) stated that in a
forestry environment, a major sediment source is unsealed or low-volume roads. Specific
road designs are necessary to help reduce sediment production. Generally, low-volume
roads are designed to be either insloped, outsloped, or crowned (Tysdal et al., 1999). The
purpose of any one of these road designs is to shed water off the road surface. With an
insloped road, water flows into a ditch running parallel to the road, and then into a cross-
drain or water bar. Flow in this ditch is described as concentrated. On an outsloped road,
water is drained evenly across the road prism and down the hillslope without the
development of concentrated water flow.
10
A crowned road is a combination of both insloped and outsloped roads. Crowned
roads will shed water into a channel on one side while evenly draining water across the
road prism on the other side. Some crowned roads may drain water into ditches on both
side of the road. The effect of insloping, outsloping or a crowned road is overshadowed
when the formation of ruts commences (Elliot et al., 1999). Once ruts begin to form,
water is no longer shed across the road prism into a ditch or down a hillslope in the
manner of the designed flow for the road. The high level of erosion from roads, reported
by many researchers, is due to the concentrated flow of water in the ruts causing an
increase in rut size, thus sediment transport.
Brinker (1995) discussed the erosive potential that water with high kinetic energy
can have. He stated that the kinetic energy of water is controlled by mass, or volume, of
water and its velocity. Water velocity and volume are controlled by slope steepness on
which the road or trail is built and the ability to move water off the road or trail (Brinker,
1995). In order to remove water from the road?s traveled surface, Brinker (1995)
suggests four diversion devices: water turn-outs, cross-drain culverts, broad-based dips,
and water bars. The use of water diversion devices that are properly spaced along with
suitable road prism shape can greatly aide in the reduction of sediment yield from low
volume roads or ORV trails. By minimizing water volume and velocity, the kinetic
energy possessed by the water is reduced, thereby reducing its erosive potential.
Limited research has been conducted on specific trail maintenance techniques for
ORV trail systems. The Soil Ecology and Research Group (2002) installed five types of
erosion control devices on the San Clemente Island ATV trail system. These were
installed for research and demonstration and to test the compatibility with ATV?s. The
11
erosion control methods that were installed were: ?wood trail drain, rubber water bar,
rock water bar, rolling dip, and Soil Sement.? Wood trail drains consisted of ditches with
boards on the sides to prevent the ditch from collapsing in on itself. Rubber water bars
utilize strips of conveyer belt placed perpendicular to the trail to divert flow from the
running surface. Rock water bars function in the same manner as the rubber water bar,
but rocks are used rather than rubber stripping. A rolling dip consists of a rock-lined
drain perpendicular to the trail which also aides in removing water from the trail surface.
Soil Sement is a soil coagulant similar to Envirotac II. The only preliminary results from
this study consisted of erosion-control cost analysis. The costs per 100 m of trail for each
erosion control type are as follows: wood trail drain is $2245, rubber water bar is $1267,
rock water bar is $1225, rolling dip is $900, and soil sement has an estimated cost of
$1050. Results on the effectiveness of each device for erosion control are yet to be
reported. On ORV trails, the spinning of tires causes rutting to occur more readily,
making erosion control a difficult task as well as making it a challenge to keep water
diversion devices intact. The use of soil stabilizers, such as Soil Sement used by the Soil
Ecology and Research Group (2002), may provide a good alternative for minimizing the
formation of ruts and destruction of water diversion devices.
Soil Erosion Amendments
Fly ash, a type of soil stabilizer, is a by-product of pulverized coal combustion
and has been tested for its shear strength when mixed with soils (Porbaha, et al., 2000).
Porbaha et al. (2000) reported that the shear strength parameters of soils treated with fly
ash were higher than soils without any form of soil stabilization, making it useful in areas
with soft grounds. The effectiveness of fly ash is a contradictory issue. Some research
12
has reported increases in permeability for soils with a maximum ash content of 10%
while other research has reported large decreases in permeability for soils with 5 to 10%
fly ash content (Porbaha et. al., 2000). Further research should be conducted to
determine if fly ash may be a viable alternative for ORV trails in the southeast.
Additional forms of soil stabilization include the use of cement kiln dust,
polyacrylamide, and acrylic copolymers. Research has been conducted on cement kiln
dust to determine its effectiveness as a soil stabilizer (Miller and Azad, 2000). Cement
kiln dust is a by-product of cement manufacturing. It is collected from kiln exhaust gases
and is unsuitable for recycling by cement manufactures, so it is disposed of as an
unusable byproduct. Cement kiln dust stabilizes soil and it increases the potential
strength of the soil while decreasing the soil plasticity index. It is thought that by
measuring soil pH, one can rapidly determine the potential increase in soil strength.
Miller and Azad (2000) conducted a study on cement kiln dust and its
effectiveness on soil stabilization and their results are used in the following discussion.
In regular cement, such as Portland cement, there are four phases that govern the strength
and curing time. The four phases are alite (Ca
3
SiO
5
), belite (Ca
2
SiO
4
), aluminate
(Ca
3
Al
2
O
6
), and ferrite (Ca
2
(Al
x
Fe
1-x
)
2
O
5
). Aluminate and ferrite are fast reacting but do
not yield high strengths, whereas alite and belite are slower reactors but yield high
strengths (Portland Cement Clinker, 2004). From this, one can gather that it would be
important for cement kiln dust to have high percentages of alite and belite. These two
phases when reacting with soil and water would yield a higher-strength soil. Data on the
chemical composition of cement kiln dust showed that it was primarily composed of
13
Silicon dioxide and calcium oxide. Silicon dioxide and calcium oxide are important
components in forming alite and belite.
Polyacrylamide (PAM) is a synthetic organic polymer that was developed for the
clarification of drinking water. PAM has a high molecular weight and, depending on its
molecular composition, it is classified as either a linear or cross-linked polymer. PAM
with a linear molecular structure is effective in controlling erosion by stabilizing soils and
removing the fine suspended sediment found in storm water runoff. Cross-linked PAM,
usually a granular crystal, can absorb hundreds of times its weight in water. There are
many applications for PAM, but erosion control is of primary interest. PAM is a long
chain organic polymer. Its effectiveness is influenced by soil structure, texture, and
salinity. Surface attraction of PAM to soil particles during irrigation by Van der Waals
and coulombic forces make it effective in forming floccules thereby preventing erosion
(Wu, 2001). The surface attractions, due to these forces, helps soil particles resist
detachment by shear-inducing forces, preventing the transport of particles in runoff thus
stabilizing the soil. PAM also enhances infiltration by improving pore continuity. Linear
PAMs used in erosion control are generally anionic and water soluble. Cross-linked
PAMs are generally anionic as well, but cross-linked in order to minimize solubility and
maximize water adsorption. The use of linear PAM as a soil amendment has been tested
primarily on erosion caused by irrigation in an agricultural application (Sojka et al.,
1998). In agriculture, PAM is applied to fields through furrow irrigation. It was also
discovered that while helping reduce erosion, PAM also aided in increasing the
infiltration rate of water. The Washington State Department of Transportation tested the
usefulness of PAM in reducing erosion from highway construction sites (WSDOT).
14
Cross-linked PAM has proven unsuccessful in gardens and houseplants for conservation
of water (Wu, 2001). Due to evapotranspiration, the same amount of water is used with
and without cross-linked PAM. The difference is that with cross-linked PAM, higher
volumes of water are required at longer intervals, rather than smaller volumes of water at
shorter intervals.
Many types and uses of acrylic copolymers exist ranging from glues, fire
retardants, carpet backings, to erosion control. Envirotac is a specific type of acrylic
copolymer that is used as a soil stabilizer. Some of the applications of Envirotac are
unpaved road stabilization, erosion and dust control, and landfill stabilizer (Envirotac,
2005). Other types of acrylic copolymers include Soil-Sement, Soiltac, and
PennzSuppress D. These types of acrylic copolymers generally are used on construction
sites and other unpaved surfaces. They are not used for erosion control on agricultural
lands.
The use of acrylic copolymers such as Envirotac for erosion control is a new
concept. These acrylic copolymers create a hard layer, or crust, on the surface that binds
the soil particles together to prevent erosion. The use of the polymer for erosion control
in areas with little or no traffic, such as on stream banks or hillsides, and dust control
need only a light application that still permits water and air penetration. Heavier
applications of polymer are used to create an impervious, durable layer on unpaved
surfaces experiencing traffic. The soil particles are bound together by resins resistant to
breakdown by water, alkaline, and UV (Envirotac, 2005). US Troops at Camp Rhino,
Afghanistan were experiencing trouble due to the formation of dust clouds generated
during helicopter landings along with the development of rutting on aircraft runways.
15
Dust was being produced from a clayless 3 ft deep soil underlain with a clay soil
(Sawyer, 2002). Envirotac diluted in water was applied to the surface to create a hardpan
layer for aircraft landings. The acrylic copolymer created a hard plastic type resin bond
with the soil particles. Due to the bond between the particles, the development of dust
was suppressed and the runway surface was stronger minimizing rutting. Minimal
studies have been conducted on Envirotac and its chemical reactivity with soils, its
relation to soil pH, and the surface attraction to sediment. Davis (unpublished thesis)
reported that research plots treated with Envirotac and then compacted had an increase in
California Bearing Ratio from 5.1 to 11.2. The same study also reported significantly
lower total suspended sediment (TSS) values for plots with Envirotac compared to
control plots. It was concluded that Envirotac II performed better than lignin treated
plots and control plots with respect to TSS levels, bearing ratio, and presence of rutting.
Erosion Modeling
Methods for quantifying erosion from roads have been studied by engineers for
years. Recent research has involved using computer models to predict sediment yield
within a watershed. The application of erosion prediction models varies depending upon
the area of focus such as agriculture, forestry, or road construction (Elliot, et al., 1999).
One of the most popular erosion prediction models is the Universal Soil Loss Equation
(USLE) (Elliot, et. al., 1999). This model was developed for application in modeling
agriculture practices but has been applied towards forest harvesting conditions with little
success in predicting sediment yield from forest roads. An alternative to the USLE is the
Modified Universal Soil Loss Equation (MUSLE). The MUSLE provides an advantage
over the USLE due to its ability to account for sediment delivery and downslope
16
deposition (Elliot, et. al., 1999). A problem with both the USLE and MUSLE is that they
were designed to average or smooth out ecological and natural variability. Since the
USLE and MUSLE are not event models, peak storm events are averaged with smaller
events, therefore reducing the natural variability. The USLE and MUSLE are not
representative of the naturally occurring variability in the ecosystem (Baffaut, et. al.,
1998). Another type of erosion prediction models are cumulative effect models, in
particular the WATSED model. The WATSED model was developed by the USDA
Forest Service in the Northwestern United States. It is useful for predicting erosion from
roads in all or part of a defined watershed. The drawback with the WATSED model, is
its inaccuracy when used for areas outside of the Northwest US (Elliot et al., 1999).
Another model used for water surface profiles in both subcritical and supercritical flow is
the Hydrologic Engineering Center?s River Analysis System (HECRAS), but is not
applicable to water erosion modeling (NRCS, 2005). Another widely used model, which
is not applicable on a small-scale watershed, is the Soil and Water Assessment Tool
(SWAT) which is used on large-scale, ungaged river basins to predict the effect of
management decisions on nutrients, pesticides, and water yields (Blackland Research
Center, 2005).
Another erosion prediction model that is becoming popular is the Water Erosion
Prediction Model (WEPP). As described by Tysdal et. al. (1999), it is ?a physically
based erosion and sedimentation model used for prediction of erosion from forest roads
that can be described as hillslopes.? The WEPP model continuously simulates daily soil
loss due to irrigation, snowmelt, or rainfall but requires many input parameters for
accurate sediment yield prediction. Parameters include rainfall amounts and intensities,
17
soil textural characteristics, soil erodibility, land management practices, plant growth, and
topography (Baffaut, et. al. 1998). The WEPP model was created with the idea that it
could be parameterized and usable for crops, soils, management practices and
topographies to which it is applied (Laflen, et. al. 1991). Because of the required inputs,
the WEPP model appears most suited and applicable for modeling soil erosion from ORV
trails.
The WEPP model requires the input of several parameters as well as the use of
other models for applications such as weather simulations. One of the models WEPP
utilizes for climate simulation is known as the CLIGEN Model. The CLIGEN model
uses statistical weather data from more than 1400 weather stations across the United
States to generate weather simulations for the desired number of years (Baffaut, et. al.
1998). In order for WEPP to predict sediment yield and deposition, it uses six different
processes: erosion processes, hydrologic processes, plant growth and residue processes,
water use processes, hydraulic processes, and soil processes (Laflen, et. al. 1991). By
looking independently at the role and impact each process has on the overall sediment
production and deposition, WEPP can be applied to a wider range of natural conditions as
well as management activities.
Inputting various parameters into the WEPP model may prove to be lengthy as
well as difficult. A study conducted by Flanagan et. al. (2000) focused on using digital
geographic information to facilitate the input of parameters. Flanagan et. al. (2000)
attempted to build a Geographic Information System (GIS) and utilized Digital Elevation
Models (DEM) for delineation of watershed boundaries, channel delineation by means of
flow accumulation, hillslope locations, and use the myriad flowpath data to determine
18
hillslope profiles that were identified as ?representative?. Multiple automatic WEPP
simulations were run for both the Hillslope method and Flowpath method. They predicted
comparable runoff and sediment losses and produced similar results from the manual
application of WEPP. Flanagan et al. (2000) concluded that there was no significant
difference between the two automatic simulations or between the automatic and manual
simulations. Attempts to model soil erosion dynamically through the use of WEPP have
also been conducted (Wu et al. 1992). They encountered difficulties similar to Flanagan
et. al. (2000) in handling and inputting large amounts of data containing temporal and
spatial variability. Through the use of GIS, the process of inputting large-scale spatial
and temporal data is simplified and the visualization capabilities of a GIS allow for
making decisions based on model result displays (Wu et al. 1992). Studies, such as these,
lead to the development of GeoWEPP which incorporates a GIS interface to WEPP using
ArcView.
Sediment prediction associated with erosion using WEPP has been applied to
several areas, including timber harvest areas (Elliot, et. al. 1996), disturbed forests (Elliot,
et. al. 1993), rangeland and croplands (Laflen, et. al. 1991), and erosion from various
road designs (Morfin, et. al. 1996; Tysdal, et. al. 1999; Elliot, et. al. 1999). Laflen et. al.
(1991) set up experimental plots to determine soil erodibility values for rangeland and
cropland soils. It was determined that important variables for soil erodibility were
organic matter content, particle size distribution, rill presence and spacing, and slope
gradient. WEPP predicted some extremely high erosion rates in rills that were realistic
on freshly tilled soils, especially if the slopes were steep and high flow rates existed
(Laflen, et. al. 1991). Another study conducted by Elliot et. al. (1994) aimed at
19
modifying the WEPP model used for rangeland and cropland and applying it to timber
harvest areas, including forest roads. Various model components required change, such
as forest soil estimation and parameters, management techniques, hydrology, and the
introduction of roads. In this study, they were able to complete some of the model
components, but stated that other components such as rainfall causing sediment
detachment, runoff, overland sheet flow, residue composition, and plant growth were still
under development. Elliot et. al. (1999) conducted a study focusing on sediment
prediction from forest roads. Some factors discussed in this study were the difficulty in
distinguishing erosion from only forest roads versus sediment within the watershed that
generated from other sources. They stated that if the road erosion rates and sediment
plume length predictions are acceptable, then the predicted sediment reaching the stream
will also be an acceptable value. This is following the assumption that if you accurately
predict the amount of sediment leaving the road and how far it travels, then it would be
expected that the predicted amount of sediment reaching the stream will also be accurate.
When modeling forest road erosion, surface conditions generally overshadow the effect
of soil properties. As discussed by Elliot et al. (1999), once the soil on a road surface is
compacted the infiltration rate approaches zero, in which case the surface cover
dominates the soil properties.
As discussed earlier various types of road designs exist aiding in shedding water
from the road prism. When using WEPP, the user must specify the road design type,
including parameters such as outsloping or insloping, cross-drain spacing, and if rutting
exists. WEPP can be calibrated to help predict erosion under various conditions. Elliot
et al. (1999) suggests using WEPP to determine erosion from bike trails or footpaths, by
20
simply narrowing the width, or sediment yield prediction from parking lots, log landings,
or other cleared areas that are less than 30 m wide and in a state of erosion.
Some more specific studies have also been conducted on modeling and prediction
of erosion with the aide of WEPP. Tysdal et al. (1999) conducted a study where WEPP
watershed model was only used to predict erosion from low-volume insloping roads. As
described earlier, with an insloping road, water flows across the road prism, into a ditch
parallel to the road, and then across the road by means of a water bar or culvert. An
insloping road incorporates complex topography making it better to model as a small
watershed, rather than a hillslope (Tysdal et al. 1999). The fact that observed sediment
yield data is highly variable demonstrates why modeling an insloping road is very
complex and challenging. They stated that ?WEPP predictions fall in the range of 24 to
74 percent of the maximum measured values.? In this study, field measurements and
WEPP predictions were very similar in showing that longer, steeper roads produce higher
levels of sediment, as well as higher sediment production when grading within the road
ditch occurs. WEPP also predicted that on long, steep slopes higher sediment yields were
predicted due to a larger contribution area. Therefore, they concluded that the WEPP
watershed model can be useful for road engineers and managers in predicting sediment
yield and runoff when used correctly with the use of appropriate variables.
Determining the level of detail for input into the WEPP model for a realistic
output and prediction is an area of study addressed by Rhee et al. (2004). In this study,
buffer geometry, hillslope topography, and road geometry of an outsloped road for
sediment yield prediction using WEPP was explored using methods consisting of low,
intermediate, and high detail. A 4.4-km road network was analyzed by dividing it into
21
segments of different lengths with data detail for each length varied from low to high.
High detail consisted of dividing the road into segments based on azimuth and grade with
buffer length and slopes measured along a curvilinear path. Low detail divided the road
into segments only at road grade reversals, and buffer length and slope measured using a
straight-line path to the stream channel. Intermediate detail segments utilized a
combination of high and low detail data analysis methods. Comparison between
simulations for the three levels of detail showed very little difference when predicting
road-generated sediment. On the other hand, when predicting sediment reaching the
stream channel, significant differences were observed between the high level of detail
compared to the intermediate and low levels. They concluded that for sediment
production from the traveled way, low levels of detail will provide adequate information
for reasonable results. If the desired output is sediment delivery, it is suggested that a
high level of detail be used and the road network divided based on buffer geometry rather
than the geometry of the road traveled way.
Sometimes it may seem practical to use Global Positioning System (GPS) as a
method for data input when modeling using WEPP in conjunction with GIS-based road
erosion models. Brooks et al. (2003) used WEPP to model erosion from a large road
network, where GPS was used to survey the road and road attributes, and GIS used to
analyze and manipulate the data. The area of study was a watershed that ranged from
277 to 2706 m in elevation, encompassing 3040 km? and included 1017 km of roads. The
roads were divided into 6955 segments and surveyed using GPS and with the use of a
data dictionary. The following attributes of road data were collected: high points, delivery
points, insloped design, outsloped design, crowned design with one ditch, crowned
22
design with two ditches, road width, road cover type, the presence of ruts, and the road
gradient. They concluded that using GPS as a form of data input through a GIS can be
very useful, although, on a road network of this magnitude the field data collection does
prove to be very time consuming. Further, results indicated that accurate predictions are
possible on a scale this large, but heavily dependent on how well the true system is
represented by the input parameters.
Summary
Research or literature on applying the WEPP or GeoWEPP model to ORV trails
has not been conducted, or found, at the time of this review. The use of GPS for spatial
data input and GIS for data analyses and manipulation, as was conducted by Brooks et al.
(2003), appears to be the best way to calibrate the WEPP model to predict sediment yield
and delivery for a watershed containing 3.2 km of ORV trails. There is limited research,
especially in the Southeast, on calibrating and testing the WEPP model using individual
storm event field data. Using the conclusions by Rhee et al. (2004) to predict sediment
delivery, a high level of detail should be used in order to achieve reliable results.
Through the review of literature it appears that limited information is available on
maintenance techniques and sediment yield and delivery predictions for ORV trails exist.
The National Off-Highway Vehicle Conservation Council (NOHVCC) outlines park
guidelines for ORV?s, but does not delineate any Best Management Practices (2002). A
set of Best Management Practices (BMPs) does not currently exist for ORV?s and ORV
trail systems. BMPs need to be created and used as guidelines during trail design and
maintenance. There is also limited research on the use of soil amendments, such as
Envirotac II, as a stabilizer for ORV trails. It is apparent that research is required to
23
quantify erosion from ORV trails, calibrating and testing the WEPP model using field
data, and testing soil amendments for their ability to reduce erosion and maintenance.
24
RESEARCH PROCEDURES
Introduction
This project was a joint study between the USDA Forest Service, Talladega
Ranger District in Talladega, Alabama and the Biosystems Engineering Department of
Auburn University in Auburn, Alabama. The project was funded by the Auburn
University Environmental Institute. The overall goal of this research project was to
establish some Best Management Practices for ORV trails, in particular the Kentuck
ORV trail system. The project sought to quantify total suspended sediment loads of
streams influenced by the ORV trails, conduct soil erosion prediction modeling to
identify the results of different management practices and perform an initial assessment
on the use of soil amendments to reduce trail maintenance. The project provides data that
can be utilized by trail designers to reduce sediment yield and deposition based on field
data and a calibrated erosion prediction model.
Site Description
The study site is located on the Kentuck ORV trail in Talladega County, Alabama
in the Talladega National Forest as shown in Figure 3.1. The Kentuck ORV trail is a
recreation area for use of All-Terrain Vehicles (ATVs) and motorcycles. The trail system
has a length of about 48 km and an average width of 3.0 m. The trails are broken down
25
into 4 loops of varying lengths. The inner loop, known as the Blue Trail, is about 3.3 km
long and is located within the watershed that this study focused on. The Blue Trail, along
with the study site, is shown in Figure 3.2. The trails are all unpaved with exposed soil in
most areas and some small sections with a gravel running surface. Maintenance practices
used on the trails are primarily broad-based dips, water bars, and water turn-outs with
timber bridges and culverts used at stream crossings. The stream monitored was a
second-order unnamed tributary to Silver Run Creek. The size
Figure 3.1 ? Site Location
26
of the watershed, with the pour point at the crossing, is about 119 ha with an average
flow of 2.5 l/s. The average flow was calculated from the flowrate data that was
collected. Flowrate data included flow for storm events as well as base flow. The soil
description at the site, according to the NRCS Soil Interpretation Record, fit into the
Fruithurst Chewacla series with 50% being Fruithurst and 30% being Chewacla. The
Fruithurst soil consists of well drained upland soils with a 12 cm thick dark yellowish
brown loam surface layer, and a subsoil red clay loam with a depth of about 86 cm.
Average slopes are between 6 and 35%.The Chewacla soil is a poorly drained soil found
on the flood plains with a brown surface layer about 20 cm thick. Under the surface layer
there is a yellowish brown silt loam and loam. Average slopes are between 0 and 2%.
Figure 3.2 ? Blue Trail with and Study Site
27
The trail approach to the stream crossing is broken into three sections, the trail
section, forest buffer section, and the section on the opposite side of the stream. On both
sides of the stream, the trail sections consist of slopes between 0-2 % for a distance of 31
m. On the west side of the stream, there is a water bar at 31 m from the stream channel
that diverts flow into a forest buffer that is 60 m long with 0-2 % slopes. On the east side
of the trail, there is also a water bar at 31 m diverting the flow through a 20 m forest
buffer with a 2 % slope. Above the water bar on the east side of the trail, four more water
bars are present at varying distances from the stream. The trail section on the east side,
above the water bar is 88 m long with slopes ranging from 2 to 18%.
Equipment Description
Stream Water Sampling
Stream water sampling was conducted in a fully automated fashion. Two ISCO
6700 automated water samplers were used at two separate points in the stream. The first
ISCO 6700 was located 16 m upstream from the bridge crossing and the other was placed
60 m downstream from the crossing. The downstream distance to the sampler was large
because there was an intrusion of sediment from a trail turnout just below the crossing.
This sampler was placed in order to capture this trail runoff.
The ISCO 6700 contains 24 one liter bottles that are filled during storm events
(see Figure 3.3). ISCO, or other commercial products, are available at the ISCO website,
http://www.isco.com. Both ISCO 6700?s were connected to an ISCO 674 Tipping
Bucket rain gauge for the purpose of triggering the sampler to start data collection. When
storm intensity reached 0.26 cm/hr, both samplers would turn on and commence
28
sampling. Every 15 minutes, one 250 ml sample was taken for a 24 hour period. Each
one liter bottle represented one hour of the storm event. The intake hose for each sampler
was mounted in the stream roughly 7 to 8 cm from the stream bottom and secured with
steel rods. To try and reduce sample contamination, the sampler would purge with air
before and after each sample was taken. Also, to collect more accurate samples, the
ISCO Water Sampler was calibrated using the intake hose length and the hydraulic head
that the sampler had to pump. To lengthen the time the sampler could stay in the field,
12-V automotive and marine batteries were used rather than the supplied ISCO batteries.
The data, which included time of trigger and time each sample was collected, was
downloaded using an ISCO 524 Rapid Transfer Device for each sampler.
Stream Flow Rate
Stream flow rate was monitored at the upstream site during this study. Flow rates
were collected using a Starflow Ultrasonic Doppler Instrument 6526C. The Starflow was
Figure 3.3 ? 24 1 liter ISCO Water Bottles
29
bolted to a steel plate and then mounted in the stream bottom using steel rods to secure it.
For use in a natural stream, the cross-sectional area must be programmed into the
instrument, in millimeters, so that it can conduct all calculations and output a final stream
flow rate. The Starflow measures water depth and an average water velocity in order to
compute flow rate. The stream cross-section was surveyed using a Topcon 720 Total
Station. Figure 3.4 illustrates the stream profile programmed into the Starflow
instrument. The Starflow was also powered using a 12-V battery and the downloaded
data was collected via a RS 232 serial cable and a laptop computer.
Stream Profile
-1400
-1200
-1000
-800
-600
-400
-200
0
0 1000 2000 3000 4000 5000 6000
Width (mm)
Dep
t
h
(
m
m
)
Rainfall Measurement
Rainfall was measured using an ISCO 674 Tipping Bucket rain gauge. This was
the same rain gauge that was used to trigger the ISCO 6700 water samplers. The rain
Figure 3.4 ? Stream Profile
30
gauge measured both intensity and total rainfall. The sampler was placed in the largest
canopy opening available and connected to the ISCO 6700 water sampler. The data was
stored directly on the ISCO 6700 water samplers and was downloaded simultaneously
with the water sample data. Rainfall amounts were compared to those of a nearby
weather station to determine accuracy. The cumulative rainfall data collected with the
ISCO rain gauge was typically within 20 % of that measured 5.5 km away in Anniston,
AL.
Traffic Measurement
Traffic measurements were conducted in a fashion that allowed for the separation
of vehicle types. A waterproof video camera with LED lights for night data collection
was mounted to a tree next to the trail 25 m past the bridge crossing. The video camera
was wired into a 12-V Video Camera Recorder (VCR) with both video and sound inputs.
A motion sensor, mounted adjacent to the bridge crossing, was connected to the VCR and
used to trigger the video camera. Each time the motion sensor was ?tripped?, the VCR
would record a 5 second video clip with a time and date stamp. This allowed for the
computation of monthly averages and the separation of ATV?s and motorcycles. Figure
3.5 demonstrates the initial video capture setup. The initial setup was only able to record
for a 3 to 4 day period before the 12-V marine battery was drained, so it was modified in
order to allow for an increased traffic data collection time. Two solar panels were
mounted above the video house, wired into a 12-V converter and connected to two 12-V
marine batteries that were wired in parallel. Depending on traffic levels, video data
collection time was extended to about 10 days. Figure 3.6 is a view from the traffic
31
monitoring station of the bridge approach and departure. Traffic volumes were measured
in number of passes per day.
Figure 3.5 ? Initial Video Housing Setup
Figure 3.6 ? View from Traffic
Monitoring Station
32
Laboratory Analysis
All laboratory analyses were conducted at the Biosystems Engineering Wet Lab at
the Auburn University Swine Research Unit. Collected water samples were analyzed for
sediment concentration by measuring Total Suspended Sediment (TSS). TSS
measurements were conducted according to standard 209C, described in the APHA
Standard Methods book (Franson, 1985). Glass fiber filters were washed using 60 ml of
distilled water, dried for one hour at temperatures between 103-105 ?C, and weighed
twice. Filters were weighed using a Denver Instruments A-200D analytical balance
accurate to 0.1 mg with a taring range of 0-200 g and a repeatability of 0.1 mg (Denver
Instruments web page). One liter sample bottles were shaken thoroughly to evenly
distribute sediment throughout. Next, 200 mL of each water sample were pulled by
vacuum through the glass fiber filter and dried. Before weighing, the filters were allowed
to cool to the balance temperature in a desiccator and then weighed as before. Sediment
concentration was calculated using the following formula:
mg total suspended sediment / L
= (A ? B) x 1000
sample volume, mL
where:
A = weight of filter + dried residue, mg, and
B = weight of filter, mg.
In some instances, the sample volumes were decreased because high sediment
concentrations would clog the filter, not allowing the remaining sample to pass through.
33
Soil Amendments
Soil amendments were tested by carefully selecting four plots. Two of the plots
were to be control plots, while the other two were treated with the soil amendment known
as Envirotac. Curved sections of trail tend to be more readily disturbed, so four trail
sections consisting of curves were chosen and paired. Prior to treatment it was
determined that the paired curves had similar ruts, same soils, and the same traffic
volumes. Data analysis on the plots was conducted by measuring cross-sectional profiles
to determine sediment loss or deposition.
Control Plots
Two control plots were used for comparison purposes with the two treated plots.
The plots were first drained because of puddling due to the formation of large ruts. With
the aid of the Forest Service, the plots were reshaped and bladed using a small dozer and
left in what is considered ?ideal? conditions, that is a smooth running surface free of ruts,
outsloped in order to shed water from the trail prism, and minimize the presence of rock
formations for rider safety. Figure 3.7 shows control plot White-A in ideal conditions.
The plot was named ?White-A? because it was the first plot measured and it was on the
section of trail called the White Trail. Steel rods were placed in concrete along the side
of the trail so that multiple trail profiles could be collected for cross-sectional area
computations. With trails closed, the plots were allowed to sit for 48 hours before the
first set of trail profile measurements were collected. Profile measurements were
collected using a string pulled tight between the steel posts. A leveled meter stick was
used to take vertical measurements from the trail surface to the string at 30 cm intervals
34
along the cross-section. Trail profile measurements were collected between mid-October,
2004 and mid-January 2005.
Treated Plots
Two plots, identified as problem areas, were used for the application of treatment.
Initial preparation of the plots was the same as occurred on the control plots. The
puddles were drained and the plots then reshaped. After reshaping, the plots were
scarified with a set of pull-behind discs to a depth of about 10 cm. With the plots
scarified, Envirotac II was applied at the manufacturer?s recommended dilution rate of
four parts water to one part chemical. The application of the chemical was conducted in
two ways. First, after dilution, the solution was streamed through a PTO pump and into 5
nozzles with 0.3 cm orifices. The large nozzle orifice was used because of the thick
consistency of the solution. The nozzles became clogged, forcing the use of a second
application technique which was to spray the solution on the plot directly from the PTO
Figure 3.7 ? Control plot White-A in ideal condition
35
pump. Figure 3.8 shows the method in which the solution was applied to the plots. The
solution was applied evenly throughout the entire plot and allowed to soak in thoroughly.
Figure 3.9 demonstrates one plot with the Envirotac II application. Once the Envirotac II
had been absorbed by the soil, the plots were compacted using a ?Sheep?s Foot? trench
compacter. The compacter was a Wacker 36-in wide, 3000-lb radio controlled machine
with the option of using vibration. Due to the soil type that was being compacted,
vibration was not used. Figure 3.10 shows the plot with the compacter near the end of the
compaction process. The manufacturer recommended 24 hours for the Envirotac II to
dry, however, the plots were allowed to ?set-up? for 48 hours before treated trail sections
were opened to traffic. Data collected on the plot consisted of profile measurements
which were collected on the treated plots in the same manner as on the control plots.
Figure 3.8 ? Application of Envirotac II
36
Figure 3.9 ? Plot soaked in
Envirotac II
Figure 3.10 ? Compaction of treated plot
37
Data Analysis
Data analysis consisted of determining total and average soil loss for each cross-
sectional profile and the complete plot itself. The initial and final cross-sectional profile
of each plot was graphed on the same chart. The area between the curves was calculated
in order to determine the amount of soil loss or deposition. By plotting the initial and
final cross-sectional measurement, it was determined visually where soil was being
detached and deposited.
Erosion Prediction Modeling
Erosion prediction modeling was conducted using the Watershed Erosion
Prediction Project (WEPP) as the driver along with three different interfaces. The WEPP
driver uses ANSI FORTRAN 77 as the source code and is capable of running on any
computer that utilizes MS DOS 5.0+. The three interfaces that WEPP utilizes allow for a
user-friendly data and parameter input as well as various outputs for easy interpretation.
The first interface was the ArcView 3.2a GIS interface, known as GeoWEPP, used to
delineate the watershed and various subcatchments. The second interface used was web-
based, known as Rock: Clime, and was a weather generator model interface. The final
interface used was the more popular windows interface known as WEPP:Windows. The
three interfaces were used to calibrate and run the model to predict erosion for various
management practices.
38
Watershed and Subcatchment Delineation
The ArcView 3.2a GIS interface, GeoWEPP, was used to delineate the watershed
and subcatchments within the watershed. Digital Elevation Models (DEM) and
topographic images were downloaded in ASCII format from the NRCS Data Gateway
web page for use in GeoWEPP. The DEM were used in GeoWEPP to determine the
topography of the area of interest. GeoWEPP allows the user to adjust the critical source
area and minimum channel length, in order to accurately identify the number of channels
in which the flow is allowed to accumulate. In order to create accurate stream channel
delineation, the critical source area and minimum channel length were adjusted to 10 ha
and 200 m, respectively. Using GeoWEPP, flow accumulation and direction were
identified and checked by ground truthing. Using collected GPS data projected into the
UTM 1983 datum, the pour point for the watershed was identified and used to delineate
the watershed. With the pour point identified, GeoWEPP delineated the watershed and 8
subcatchments within. Figure 3.11 identifies the 119 ha watershed with the 8
subcatchments with the 3.3 km, Blue Trail, as the top layer in the image. The watershed
and subcatchments were saved for later modification using the WEPP: Windows
interface.
39
Weather Generator
WEPP uses a stochastic weather generator called CLIGEN. It is accessible
through WEPP: Windows or the internet interface known as ROCK: CLIME. The
ROCK: CLIME interface was used because of its ease of modification. The ROCK:
CLIME weather station of Anniston, AL was used because of its proximity to the trail
system. The Anniston weather station is about 5.5 km from the trail system. Using
PRISM (Parameter-Regression on Independent Slopes Model), the climate parameters
were modified for the exact latitude, longitude, and elevation of the Kentuck Trail
System. These parameters were used to generate one year of simulated weather and then
saved and downloaded. The generated weather was next modified to match the actual
Figure 3.11 ? Watershed and subcatchments
40
precipitation amounts and dates recorded during field collection. This was accomplished
by zeroing out precipitation amount, precipitation duration, time to peak intensity, and
peak intensity. These four values were calculated using field data for every storm event
and input into the generated weather file for its respective date. The weather data is
generated on a daily basis over a one year period. With the weather file modified, it was
ready to be imported into WEPP: Windows. A second weather file was also created
using same process as above, except this file was generated to simulate 30 years of
weather. This file was saved and not modified.
Subcatchment Modification and Model Calibration
Once the subcatchments were delineated with the aide of GeoWEPP and the
weather modified using ROCK: CLIME, a new project was created with WEPP:
Windows in order to conduct final modifications and model calibration. Using a
clinometer and meter tape, slope length and percents were recorded for sections of the
trail that intersected streams. These values were recorded and used to modify the model
of subcatchments that included stream/trail intersects. Since all trail approaches to the
stream included waterbars, it was assumed that the last waterbar before the stream
diverted 100% of the flow into the forest. Therefore, the forest acted as a buffer for flow
of water and sediment traveling along the trail above the final waterbar. WEPP
represents terrain in a three layer fashion. The top layer is used to input the current
management practice. The second layer is used to input the slope length and steepness.
The third layer is used to input the soil types that are found on the particular hillslope.
By inputting breaks in any of the layers, the user can create different overland flow
41
elements (OFE) to better represent the hillslope. Breaks were used within the model to
separate trail sections and soils from forest buffer sections and soils. Also, on sections of
trail that were directed straight down the slope, it was assumed that all flow remained on
trail, either in ruts or bare ditches. These sections of trail were modeled as insloped roads
with a bare ditch, or outsloped road with the presence of ruts. Both parameters yield
similar results because they both conduct flow down the trail rather than shedding water
from the road prism. The final section of trail below the last water bar was also
considered and it was modeled as an insloped trail with all sediment loss being delivered
directly into the stream. The section of trail on the other side of the stream was modeled
in this same manner. Figures 3.12 and 3.13 represent the insloped trail leading directly
into the stream and the insloped trail with the forest buffer, respectively. These figures
demonstrate the data input screen that WEPP uses. Figure 3.12 shows a hillslope
representing a 1 m wide section of the hillslope on which the trail is located. Figure 3.13
is similar except that it includes a 20 m section of forest buffer at the hill bottom. The
different color at the bottom of the hillslope represents the forest buffer with its
associated soils.
For subcatchments containing trails but not crossing any stream channels, the
distance between the trail and the stream was measured using ArcView. These sections
of trail were modeled as outsloping with sediment flowing across the road prism and the
distance between the trail and the stream was considered to be a forest buffer. Figure 3.14
represents this.
42
Figure 3.13 ? Insloped Trail with 20 m Forest Buffer
Figure 3.12 ? Insloped Trail Leading into Stream
43
Once all the subcatchments were modified with the proper slope parameters, the
soil parameters had to be adjusted. To ensure proper soil parameters, field samples were
collected and analyzed by the USDA Forest Service and the Auburn Soils Testing
Laboratory for percent sand, silt, and clay. These values were then compared to data
collected by both the USDA Forest Service and the NRCS. From the WEPP literature, it
was determined that the four most important and sensitive soil parameters used to
determine sediment yield and deposition were the following: interrill erodibility, rill
erodibility, critical shear, and effective hydraulic conductivity. Because of the
importance of these values, the soil parameters determined from the field data were used
along with the following formulas taken from the WEPP Application Help:
Interrill Erodibility (kg*s/m^4)
Ki = 2728000 + 192100 * VFS
Rill Erodibility (s/m)
Kr = 0.00197 + 0.0003 * VFS + 0.03863 * e(-1.84) * ORGMAT
Figure 3.14 ? Outsloped Trail with Forest Buffer
44
Critical Shear (N/m?)
?c = 2.67+.065*Clay ? 0.058 * VFS
Effective Hydraulic Conductivity (mm/h)
Kb = -0.265 + 0.0086 * Sand ^(1.8) + 11.46 * CEC^(-0.75)
where:
VFS = percent very fine sand in the surface soil
ORGMAT = percent organic matter in the surface soil
Clay = percent clay in the surface soil
Sand = percent sand in the surface soil
CEC = Cation Exchange Change
The interrill erodibility, rill erodibility, critical shear, and effective hydraulic conductivity
were input as model parameters and multiple simulations were conducted. The model
was set to output event-by-event storm data so as to make a direct comparison with the
event-by-event field data. The effective hydraulic conductivity was reduced because the
calculated value represented an undisturbed soil and, due to the compaction on the trail
surface, the actual hydraulic conductivity was much lower. The reduced value that was
used was similar to the default WEPP value for a loam road surface. After each
simulation, the storm-by-storm output was recorded and compared to the field data.
Numerous simulations were carried out in which the four parameters defined above were
systematically changed and model outputs were compared to the measured sediment
yield values. When the highest correlation between measured and modeled values was
achieved it was assumed that the model calibration was complete.
45
Management Simulations
Using the non-calibrated model, multiple simulations were conducted to simulate
various management activities. Simulations were conducted on waterbar spacing,
minimum forest buffer lengths, and acceptable slope steepness. The simulations were
conducted using an unmodified 30 year weather file. Simulations were also conducted to
demonstrate the importance of proper maintenance including functioning waterbars,
reducing rut formation, and forested buffers.
Statistical Analysis
Statistical analysis was conducted to test the goodness-of-fit between the
measured and predicted event-by-event sediment losses (Spruill et. al., 2000). The Nash-
Sutcliffe coefficient, R?, was used to measure the goodness-of-fit. The equation used was
the following:
R? = 1 - ? ( Qm ? Qp )?
? ( Qm ? Qavg )?
where:
Qm = measured soil loss (kg)
Qp = predicted soil loss (kg)
Qavg = average soil loss (kg)
By comparing R? values for each simulation, individual predicted storm events were used
and compared to the measured of the same storm events. This allowed for accurate
model calibration and reasonable erosion prediction results for varying management
activities.
46
RESULTS AND DISCUSSION
Stream Water Sampling
Stream water sampling data collection was conducted between December of 2003
and July of 2004. The data collection time was separated into three periods. The first
period was during the winter months when the ORV trails were closed, the second period
was in the early spring when trail maintenance was conducted, and the third period was
during spring and summer when the trails were opened and ATV and off-road motorcycle
trafficking allowed. Data were collected during these three periods and the results are
separated accordingly. Three example storm events with similar cumulative rainfall will
be discussed in the next three sections along with maximum sediment producing storm
events and total sediment production during each period.
Trails Closed Period
Data collected in this period was from January through February 2004. During
this period, the ORV trails were closed, therefore there was no trafficking taking place.
Four storm events were monitored during this period with cumulative rainfall for each
storm ranging from 1.12 cm to 4.88 cm. Data for each storm event is separated into Site
A and Site B, representing the sampler upstream from the bridge crossing and the
sampler downstream from the bridge crossing, respectively. Figures 4.1 through 4.4
exemplify a single storm event that took place February 12, 2004. These figures
47
represent cumulative rainfall, stream flowrate, total suspended sediment (TSS), and
sediment loading. In Appendix A, plots for all storm events during the trail closed period
are represented.
0
0.25
0.5
0.75
1
1.25
1.5
1.75
2
2.25
01234567891011213141516171819202122324
Time Since Start of Sampling (hrs)
C
u
m
m
u
lat
i
ve R
a
i
n
f
a
ll (cm
)
Figure 4.1 ? Cumulative Rainfall for February 12, 2004 storm event
48
0.00
0.50
1.00
1.50
2.00
2.50
3.00
3.50
4.00
4.50
1 2 3 4 5 6 7 8 9 101 12131415161718192021222324
Time (hrs)
Flowr
a
t
e
(
l/s
)
0
20
40
60
80
100
120
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time (hrs)
S
u
s
p
e
n
de
d S
e
d
i
m
e
nt
(
m
g/
l)
Site A Site B
Figure 4.3 ? TSS for February 12, 2004 storm event
Figure 4.2 ? Hydrograph for February 12, 2004 storm event
49
0.000
0.200
0.400
0.600
0.800
1.000
1.200
1 2 3 4 5 6 7 8 9 101112131415161718192021222324
Time (hrs)
S
e
dim
e
n
t
Loa
d (
k
g/
hr
)
Site A Site B
The cumulative rainfall for this storm event was 2.0 cm. The maximum TSS level
measured at the upstream sampler was 95 mg/l while the maximum measured at the
downstream sampler was 108 mg/l for this storm event. These peaks were only reported
for a one hour period before a significant drop was noticed. Sediment load was
calculated by multiplying TSS values with flowrate values. The maximum sediment
loads calculated for this storm event were 0.96 kg/hr and 1.09 kg/ hr for Site A and Site
B, respectively. By integrating under the curve for each sediment load and then taking
the difference between Site A and Site B, total sediment introduced at the crossing was
calculated. From Figures 4.3 and 4.4, it is noted that at the 5
th
and 6
th
hour, the upstream
concentration and sediment load is higher than that of the downstream site. It is difficult
to determine what caused this situation to occur. Possibilities for this include
Figure 4.4 ? Sediment Load for February 12, 2004 storm event
50
measurement error, channel overflow causing deposition, or bank sloughing just
upstream of the Site A sampler. Table 4.1 summarizes sediment loads for the February
12, 2004 storm event and for all four storm events that occurred during the period when
the trails were closed.
Storm Event
12-Feb-04 Entire Closed Period
Location Storm 4 - storm events
(kg) (kg)
Site A 4.8 177.6
Site B 4.9 286.9
Sediment Introduced 0.1 109.4
% of Total 2.8 38.1
The storm of February 12, 2004 did not contribute a significant sediment load to stream
channel. This is shown in Table 4.1. The percent of total value represents the amount of
sediment introduced into the stream at the crossing compared to the total sediment in the
stream. The maximum TSS and sediment load during this period occurred on February 6,
2004 with a cumulative rainfall of 4.9 cm. The TSS level peaked at 1715 mg/l resulting
in 108.9 kg of sediment introduced at the crossing. This explains why the total sediment
load value for the closed period in Table 4.1 is rather large. This storm is shown in
Appendix A, but not here because comparisons were made only between storms with
similar cumulative rainfall.
Table 4.1 ? Sediment Load summary for storm event on
February 12, 2004 and the trail closed period (Jan ? Feb, 2004)
51
Trail Maintenance Period
The trail maintenance period occurred during the month of March 2004. During
this period, the USDA Forest Service conducted their standard annual trail maintenance
in preparation for the opening of the trail on April 1, 2004. Four storm events were
recorded with cumulative rainfall during maintenance ranging from 0.51 cm to 2.032 cm.
Figures 4.5 through 4.8 demonstrate a single storm event that occurred on March 6, 2004.
The figures represent cumulative rainfall, stream flowrate, TSS, and sediment loading.
The cumulative rainfall for this storm event was 2.03 cm with a peak intensity of 2.24
cm/hr. Other storm events recorded during this period can be found in Appendix B.
52
0
0.25
0.5
0.75
1
1.25
1.5
1.75
2
2.25
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time Since Start of Sampling (hrs)
C
u
mm
u
l
at
i
ve Rai
n
f
a
l
l
(
c
m)
0
2
4
6
8
10
12
14
16
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time (hrs)
Fl
ow
r
a
te
(l
/s
)
Figure 4.5 ? Cumulative Rainfall for March 6, 2004 Storm Event
Figure 4.6 ? Hydrograph for March 6, 2004 Storm Event
53
0
50
100
150
200
250
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time (hrs)
S
u
s
p
e
n
de
d S
e
d
i
m
e
nt
(
m
g/
l)
Site A Site B
0
2
4
6
8
10
12
1234567891011213141516171819202122324
Time (hrs)
S
e
di
m
e
n
t
Lo
a
d
(
k
g/
hr
)
Site A Site B
Figure 4.7 ? TSS for March 6, 2004 Storm Event
Figure 4.8 ? Sediment Load for March 6, 2004 Storm Event
54
The maximum TSS level measured for this storm event was 170 mg/l and 200
mg/l for the upstream sampler and the downstream sampler, respectively. Conducting the
similar computation as before, the sediment load introduced at the crossing for this storm
event was 4.01 kg. Table 4.2 summarizes the storm events collected during the
maintenance period.
Storm Event
6-Mar-04 Maintenance Period
Location Storm 4 - storm events
(kg) (kg)
Site A 26.72 30.19
Site B 30.73 34.87
Sediment Introduced 4.01 4.68
% of Total 13.05 13.42
The storm event of March 6 recorded both the largest cumulative rainfall and
largest sediment introduction of any storm during the maintenance period. The March 6
storm event was rather small, but yet a relatively large sediment load was introduced into
the stream at the crossing. The reason for this could have been because during
maintenance, significant disturbance is caused to the trail surface. Maintenance is
conducted using a Caterpillar D3C dozer and, while the trail condition is improved, the
surface disturbance allows for increased soil detachment resulting in higher sediment
losses. Since the March 6 storm event was the first to occur after maintenance
commenced, recorded sediment loads were highest because of the very recent
disturbance. Once the loose, disturbed surface soil is eroded, more energy is required to
move the remaining soil, so the next few storms during the maintenance period do not
have the same impact as the March 6
th
storm had.
Table 4.2 ? Sediment Load summary for storm event on March
6, 2004 and trail maintenance period
55
Trail Open Period
The trail opening day was April 1, 2004. The trail open period data collection
took place from opening day through July 2004. During this period, traffic was allowed
on the trails at all times and under any condition. Instrumentation was installed for data
collection designed to determine traffic volumes. Eight storm events were recorded for
the period while the trail was open and the cumulative rainfall for these events ranged
from 0.97 cm to 3.61 cm. Figure 4.9 though 4.12 were recorded from a storm event that
took place on April 30, 2004. As before, these figures represent cumulative rainfall,
stream flowrate, TSS, and sediment loading. The cumulative rainfall for the April 30,
2004 storm event was 2.24 cm with a peak intensity of 2.64 cm/hr. Other storm events
recorded during this period can be found in Appendix C.
56
0
0.5
1
1.5
2
2.5
0 1 2 3 4 5 6 7 8 9 101112131415161718192021222324
Time Since Start of Sampling (hrs)
C
u
m
m
u
la
tiv
e
R
a
i
n
fa
ll (
c
m
)
0.0
1.0
2.0
3.0
4.0
5.0
6.0
1234567891011213141516171819202122324
Time (hrs)
Fl
ow
r
a
t
e
(
l
/
s
)
Figure 4.9 ? Cumulative Rainfall for April 30, 2004 Storm Event
Figure 4.10 ? Hydrograph for April 30, 2004 Storm Event
57
0
100
200
300
400
500
600
0 5 10 15 20 25 30
Time (hrs)
S
u
s
p
e
nde
d S
e
di
m
e
nt
(
m
g/
l
)
Site A Site B
0.0
1.0
2.0
3.0
4.0
5.0
6.0
1234567891011213141516171819202122324
Time (hrs)
S
e
di
m
e
nt
Lo
a
d
(
k
g/
hr)
Site A
Site B
Figure 4.11 ? TSS for April 30, 2004 Storm Event
Figure 4.12 ? Sediment Load for April 30, 2004 Storm Event
58
The maximum TSS level for the upstream and downstream sampler recorded for
this storm event was 270 mg/l and 280 mg/l, respectively. The sediment load introduced
at the bridge crossing during this storm event was 2.2 kg. Figure 4.11 has a peak during
the first hour of data collection that may have been caused by leaf and sediment
accumulation around the hose intake at the time the first sample was collected. Table 4.3
summarizes the sediment load from this storm event and the sum of sediment loads from
all storm events that occurred while the trail was open. The sediment load recorded for
the April 30, 2004 storm event was the highest recorded while the trail was open.
However, on May 16, 2004 a storm event occurred with a cumulative rainfall and peak
intensity of 3.61 cm and 5.99 cm/hr, respectively, both values were larger than that of the
April 30, 2004 storm event. Plots for this storm event are displayed in Appendix C.9
through C.12. The hypothesized reason for this storm event having a lower sediment
yield is that during the first two weeks of May not a single rain event was recorded. This
would cause the soils to be drier so the water storage would be lowered, resulting in a
higher infiltration volume before runoff would commence. On April 27
th
and 28
th
, two
rain events occurred, causing water storage to be almost at peak, so runoff would occur
more readily during the recorded April 30 storm event.
Storm Event
30-Apr-04 Open Period
Location Ex. Storm 8 - storm events
(kg) (kg)
Site A 18.21 53.47
Site B 20.40 60.35
Sediment Introduced 2.19 6.88
% of Total 10.74 11.40
Table 4.3 ? Sediment load summary for April 30, 2004 storm
event and the trail open period
59
0
20
40
60
80
100
120
140
160
4/16/2004 4/18/2004 4/20/2004 4/22/2004 4/24/2004 4/26/2004 4/28/2004 4/30/2004
Time (days)
# o
f
Pass
es
2-wheeler 4-wheeler Total
The traffic volumes measured on the trail systems were very high and is
demonstrated in Figure 4.13 by showing traffic levels during the month of April. Traffic
volumes were separated into 2-wheel and 4-wheel vehicles. The weekends of April 17,
2004 and April 25, 2004 reported traffic totals of 141 and 71 passes, respectively. Table
4.4 shows the traffic totals and averages for the data collection period. The data in Table
4.4 shows that in 87 days of traffic volume quantification, a total of 2200 passes were
counted resulting in an average of 25.3 passes per day. These traffic totals and averages
were much higher than expected. However, the 4-wheeler average exceeding that of the
2-wheelers was expected. The reported total of 2200 does not represent separate riders;
rather it is 2200 passes in front of the traffic monitoring station. It should be noted that
Figure 4.13 ? Traffic Volumes for April 2004
60
Total Days Traffic Totals Total
2-wheeler 4-wheeler Other
87 506 1586 108 2200
Average 5.8 18.2 1.2 25.3
traffic volumes were not collected for the entire period that the trails were open due to
equipment malfunction. Equipment downtime ranged from 5 days to 2 weeks per month.
Also, the category of ?Other? includes vehicles other than 2-wheelers and 4-wheelers or
vehicles that could not be distinguished due to equipment malfunction.
Overall Sediment Loading
Further analysis between storm events was conducted to determine correlation
between cumulative rainfall, stream flowrate, TSS, and sediment load. All storm events
were separated into three categories based on rainfall return periods. The first category
was for storm events with a return interval of one month or less. This was the equivalent
of a cumulative rainfall of 1.5 cm or less, and it included seven storms. The second
category was storm events with a return interval between one month and one year.
Rainfall for this category was between 1.5 cm and 3.3 cm. It included seven storms as
well. The last category included storm events with a return interval longer than one year.
Within each category, the TSS levels were averaged and the difference taken and the
flowrates were averaged as well. The difference in sediment concentration and the
flowrates were plotted for each category as Figures 4.14 and 4.15, respectively.
Table 4.4 ? Daily traffic totals and averages for trail open period
61
-50
0
50
100
150
200
250
300
350
400
450
0 5 10 15 20 25 30
Time (hrs)
To
ta
l S
u
s
p
e
nde
d
S
e
di
m
e
nt (m
g/l
)
< 1.5 cm 1.5 - 3.0 cm > 3.0 cm
0
2
4
6
8
10
12
0 5 10 15 20 25 30
Time (hrs)
Fl
owr
a
t
e
(
l/s
)
< 1.5 cm 1.5 - 3.0 cm > 3.0 cm
Figure 4.14 ? Average Difference in Sediment Concentrations for
Categorized Storm Events
Figure 4.15 ? Average Flowrate for Categorized Storm Events
62
From Figure 4.14, it is noted that the average difference in sediment
concentrations has little effect until the cumulative rainfall is in excess of 3.3 cm. Little
difference is noticed between the two lower categories in this figure. Figure 4.15, on the
other hand, demonstrates that cumulative rainfall significantly affects stream flowrates.
A very large difference is noticed when the cumulative rainfall exceeds 3.3 cm. Since
variation in TSS levels is minimal, as the cumulative rainfall increases, the flowrates
increase, causing the sediment load to increase. Table 4.5 sums the total sediment
introduced at the bridge crossing and the cumulative rainfall for each category.
The total sediment for storm events less than 1.5 cm total rainfall is 2.7 kg of sediment
which is less than the other categories and as expected. The other two categories
followed as expected. The total sediment for storms with greater than 3.3 cm of rainfall
were significantly higher. The reason for this is that with increased rainfall, there is a
Cumulative Rainfall
Categories
< 1.5 cm 1.5 - 3.3 cm > 3.3 cm
Total Sediment Introduced at Crossing
(kg) 2.73 7.18 110.99
Total Cumulative Rainfall (cm) 7.14 13.87 3.61
significant increase in flowrate which results in much higher TSS values. Further
analysis was conducted and total sediment loads were compared for the three periods,
trails closed, trail maintenance, and trails opened. As before, the values are reported in
kg of sediment. This comparison is shown in Table 4.6.
Table 4.5 ? Sediment Load for Cumulative Rainfall Categories
63
Cumulative Sediment Percent
Trail
Condition Date Rainfall Load of Total
(cm) (kg) (%)
Trail Closed 1/25/2004 2.2 0.13 14
2/6/2004 4.9 108.98 39
2/12/2004 2.0 0.14 3
2/25/2004 1.1 0.11 13
Trail 3/6/2004 2.0 4.01 13
Maintenance 3/16/2004 0.5 0.08 37
3/20/2004 0.9 0.42 25
3/29/2004 2.0 0.17 8
Trail Open 4/11/2004 1.3 0.53 31
4/26/2004 1.8 0.36 43
4/30/2004 2.2 2.18 11
5/16/2004 3.6 2.02 22
6/16/2004 1.3 0.83 3
6/22/2004 1.0 0.60 70
7/2/2004 1.6 0.19 10
12/16/2003 1.0 0.17 28
During this study, storm events that had a cumulative rainfall over 2 cm
contributed elevated sediment loads. It was expected that storms occurring during the
maintenance period would produce the highest level of erosion due to the increased
disturbance that occurred. However, the lowest sediment load was recorded during this
period with only 0.08 kg being contributed during a 0.5 cm storm event. The largest
contribution of sediment into the stream came during the period when the trails where
closed. A 4.9 cm storm event contributed 108.9 kg of sediment at the stream crossing.
The main contributing factor for this increased sediment load was the quantity of rainfall.
The relationship between cumulative rainfall and total sediment load was not a
linear relationship. Figure 4.16 represents cumulative rainfall vs. sediment load. It can
be noted that as the rainfall increases, the sediment load increases as well but at a slower
rate until a certain rainfall is reached. This data suggests that ORV trails do not
Table 4.6 ? Sediment Load for Each Period
64
contribute significant sediment loads during small storm events. However, there is a
potential for large sediment contributions from storm events with a one year or higher
return interval. In Appendix E there is a table with the rainfall and sediment load values.
0.00
20.00
40.00
60.00
80.00
100.00
0.00 1.00 2.00 3.00 4.00 5.00 6.00
Rainfall (cm)
S
e
di
m
e
nt
L
o
a
d
(k
g)
Modeling
The web, ArcView, and windows were used to setup the WEPP model for
calibration and simulation. The web interface, ROCK:CLIME, was used to create a
climate file that was later modified in order to match the climate reported during data
collection. The ArcView interface, GeoWEPP, was used to delineate both the watershed
of interest and the subcatchments within the watershed. The windows interface of WEPP
was used to modify the subcatchments to match the actual trail conditions. With the
Figure 4.16 ? Relationship between Cumulative Rainfall and Sediment
Load
65
setup complete, the model was run, calibration conducted, and the various management
practices simulated for the determination of BMP?s.
Weather Generation
As mentioned before, the weather file to be used with the WEPP model was
created using the web interface. The file was then modified to match the storm events
that were recorded during data collection. A total of 16 storm events were recorded
during the data collection period. During the eight month data collection period, not all
storm events were captured. Figure 4.17 shows all the storm events that occurred as well
along with the storm events that were monitored. Modifications were made to the values
0.00
1.00
2.00
3.00
4.00
5.00
6.00
7.00
12/
1/
03
12/
15/
03
12/
29/
03
1/
12/
04
1/
26/
04
2/
9/
04
2/
23/
04
3/
8/
04
3/
22/
04
4/
5/
04
4/
19/
04
5/
3/
04
5/
17/
04
5/
31/
04
6/
14/
04
6/
28/
04
7/
12/
04
7/
26/
04
Date
Cu
mu
la
t
i
ve
D
a
i
l
y
R
a
in
f
a
ll
(
c
m)
Rainfall with
Sedimen Data
Rainfall without
Sediment Data
Figure 4.17 ? All Storms and Storms with Sediment Data
66
of precipitation, precipitation duration, time to peak intensity, and peak intensity by
verifying values collected by the ISCO 6700 water samplers. Weather information that is
used by WEPP, but not modified, were radiation, wind velocity, wind direction, and dew
point temperature. Table 4.7 shows the portion of the weather file that was modified to
include the storm events that were collected.
Day Month Precip Duration
Time to
Peak
Int. at
Peak
(mm) (h) (h) (mm/h)
25 1 22.35 0.50 0.25 79.25
6 2 48.77 5.50 0.75 38.61
12 2 20.32 4.75 0.25 38.61
25 2 11.76 21.00 0.25 12.20
6 3 20.32 4.75 0.50 22.35
16 3 5.08 0.50 0.25 10.16
20 3 9.14 1.25 0.50 21.34
29 3 20.07 6.25 0.25 10.16
11 4 13.46 2.75 0.25 15.24
26 4 17.53 9.75 0.25 12.19
30 4 22.35 5.50 0.50 29.46
16 5 36.07 3.25 0.50 59.94
16 6 12.70 1.00 0.50 36.58
22 6 10.16 1.75 0.25 15.24
2 7 16.00 1.50 0.50 41.66
16 12 9.65 2.75 0.25 10.16
Parameter Calculation and Model Calibration
Before the model simulations were run, erodibility, shear, and hydraulic
conductivity were calculated using the soils data and the equations outlined in the
methodology. The soil on the trail surface is classified as a clay loam soil with 27 %
sand, 28 % clay, and 44 % silt. Organic matter was 0 and the cation exchange capacity
(CEC) was 12. The CEC was taken from NRCS soils data. Soil parameters for the
Table 4.7 ? Portion of modified weather file
67
WEPP model were calculated using these values. These values are shown in Table 4.8.
Slope length, buffer length, and slope gradient were also input into the model.
Parameter Value
Interrill Erodibility (kg*s/m^4) 8.00E+06
Rill Erodibility (s/m) 0.01
Critical Shear (N/m?) 2.91
Effective Hydraulic
Conductivity (mm/h) 4.86
Once all the values were inputted into the model, initial attempts to calibrate the
model were conducted. It was realized that due to the lack of large storm events in the
data set, proper model calibration was very difficult. Calibration of the model was
conducted as best possible while understanding the limitations of the data set. After each
simulation, the event-by-event summary for each storm was recorded, compared to the
measured value, and the Nash-Sutcliffe R? calculated to determine model performance.
The R? calculated from the initial simulation was 0.90. In order to try and increase model
performance, a maintenance rotation occurring on March 1
st
was added as a model
parameter. Maintenance consisted of smoothing the trail with a blade that affected 100%
of the trail surface and disturbance to a depth of 20 cm. The other value that was
adjusted was the effective hydraulic conductivity. The calculated value for effective
hydraulic conductivity, as shown in Table 4.8, was 4.86 mm/h. Since the trail surface
receives a significant amount of traffic, it was assumed that the hydraulic conductivity
would be lower. So, this value was lowered to 2.5 mm/hr. With the maintenance rotation
and the lowered effective hydraulic conductivity, the new R? value was 0.63. The event-
by-event summary was reanalyzed and it was evident that the model was over-predicting
Table 4.8 ? Calculated soil parameters
68
the amount of sediment being introduced at the crossing. The same process was repeated
using various values for the effective hydraulic conductivity. With the effective
hydraulic conductivity set at 3 mm/hr, the resulting R? was 0.92. After achieving an R?
value of 0.92, it was assumed that the model was calibrated as best as could be expected
from a limited data set. Figure 4.18 shows the measured values vs. the predicted values
for data used to calibrate the model.
From Figure 4.18, it was evident that the model was matching the large storm
events accurately, while the smaller storm events were not accurately matched. The large
storm event of February 6, 2004 was left out and an attempt was made to recalibrate the
model. The same calibration process as before was followed. With the effective
hydraulic conductivity lowered to 0.18 mm/hr, an R? of 0.20 was calculated. This was
the highest R? value that could be achieved. Figure 4.19 shows the measured vs.
predicted values for the calibration of the smaller storms.
R^2 = 0.92
0
20
40
60
80
100
0 20 40 60 80 100
Predicted Sediment Load (kg)
M
easu
red
Sed
i
men
t
L
o
a
d
(
k
g
)
Measured vs. Predicted 1:1
Figure 4.18 ? All Measured vs. Predicted Sediment Load Values
69
R^2 = 0.20
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
00.511.522.533.544.5
Measured Sediment Load (kg)
P
r
e
d
i
c
t
e
d S
e
di
m
e
n
t
Lo
a
d
(
k
g
)
Measured vs. Predicted Series2
The effective hydraulic conductivity was the only valued varied during the
calibration. The effective hydraulic conductivity had a very high sensitivity, and varying
it by as little as 1 mm/hr caused significant changes in the model output. Typical value
ranges for effective hydraulic conductivity along with the other soil parameters are listed
in Table 4.9. The soils values calculated and used during these simulations are all within
the typical values reported for surrounding soil types.
Figure 4.19 ?Measured vs. Predicted Sediment Load Values for
smaller storm events
70
Soil Parameter Minimum Maximum
Interrill Erodibility 1.50E+06 9.80E+06
(kg*s/m^4)
Rill Erodibility 0.0002 0.0227
(s/m)
Critical Shear 0.2 4.68
(N/m?)
Effective Hydraulic
Conductivity 0.1 19.35
(mm/hr)
From the calibration attempts, it was observed that the WEPP model tended to
over-predict average, day-to-day storm events. During both calibration efforts, the two
storm events that proved to be most accurately represented were the two that had a
cumulative rainfall higher than 3.3 cm. From this, it is concluded that the WEPP model
is capable of accurately simulating large storm events, those in excess of 3.3 cm of
rainfall, and it tends to over-predict smaller storm events that are witnessed on a daily
basis. Due to difficulties encountered while attempting to calibrate the model, it was
concluded that proper model calibration would not be possible due to limited number of
large storm events in the data set.
Development of Best Management Practices for ORV Trails
To develop best management practices, the non-calibrated model was used. The
use on a non-calibrated model was justified by utilizing the soil parameters calculated
before. The effective hydraulic conductivity used was 0.18 mm/hr, which is the value
recommended by WEPP for low-volume roads. The weather utilized was a CLIGEN
generated 30 year climate file. The WEPP model was modified to represent trail
conditions under the following management practices: varying distance between water
Table 4.9 ? Typical soil parameter ranges
71
bars, varying gradient, and varying forest buffer lengths. The assumptions that were
made during the simulation were regarding water bars. It was assumed that all water bars
were designed with turn-outs, all surface flow was diverted off the trail surface, and they
were kept in working order. During simulation the output value used to determine the
effectiveness of the varying management practices was the average annual sediment
yield. The average annual sediment yield is the amount of sediment that is transported
through the buffer and deposited into the stream channel. According to Schwabb et. al.
(1993), sediment yields from agricultural fields with corn averaged 16 metric tons per
hectare over a six year period. From this, the target average annual sediment yield used
for these simulation was set to 11 metric tons per hectare (t/ha). If sediment yield for the
varying management practices was below this level, they were recommended as a BMP.
Water Bar Spacing and Slope Gradient
Water bars used along with turn-outs allow for water to be diverted from the trail.
As the distance between the water bars decreases, the volume of water flowing on the
trail decreases, leading to a reduction in erosion. As the slope increases, the velocity of
water increases, so the distance between water bars should be decreased (Brinker, 1995).
Table 4.10 represents the recommended distance between water bars as the slope
increases. The buffer consisted of a 20 year old forest through which the trail crossed a
20 m wide buffer. The recommended water bar distances in Table 4.10 are for sections
of trail that are within 20 m of a stream.
72
Slope
Steepness Distance
(%) (m)
4 14
6 11
8 10
10 9
12 8
14 7
18 6
20 6
Brinker (1995) suggested using the following as a rule of thumb to calculate water
bar spacing on low-volume roads: (in feet) = (400 / slope %) + 100. All the values in
Table 4.10 were less than the values calculated using the equation above. Simulations
were conducted using WEPP in order to acquire the values in Table 4.10. The two
sections of trail adjacent to the stream were modeled and the predicted sediment load for
each were summed and held constant. The upper section of trail, that included the water
bars, was simulated for varying gradient and water bar spacing. The predicted values
from the upper section were summed with the two lower sections and these values had to
be under 11 t/ha. If the values were over 11 t/ha, the slope gradient and water bar spacing
was changed and simulations conducted until the values were under the acceptable limit.
Buffer Length
For all the above simulations, the length of the forested buffer was 20 m. If the
buffer length is increased, then the values in Table 4.10 are still valid. If the length of the
buffer is decreased, then the Table 4.10 is not valid. Alabama Best Management
Table 4.10 ? Recommended water
bar spacing with a 20 m forest
buffer
73
Practices suggest a minimum buffer length of 10.67 m. Table 4.11 contains the
recommended water bar spacing for varying slopes in the case where the manager
chooses to use a forest buffer of 10.67 m. The distances for these water bars are for
sloped trail sections that are within 10 m of the stream.
The reason for the spacing between water bars decreasing as the buffer length
decreases is because with a smaller buffer, there is less area for sediment deposition, so
higher amounts of sediment reach the stream channel. Because of the very close water
bar spacing that exists when the slope grade increases and buffer length decreases, it is
recommended that minimum buffer length and maximum slope gradient of 20 m and
12%, respectively, be used when designing or modifying trails. In effort to minimize
impacts on riparian areas and minimize sediment loading, the section of trail that crosses
the stream should be perpendicular to the stream with slopes between 0 and 2%.
The effectiveness of a forest buffer is directly related to the width of the buffer
itself. Figure 4.20 demonstrates sediment loss and deposition in relation to its location
along the hillside. In the figure, the green line signifies erosion when it is below the
hillslope profile and deposition when it is above. The location in which deposition
Slope
Steepness Distance
(%) (m)
4 8
6 7
8 6
10 5
12 5
14 4
Table 4.11 ? Recommended water
bar spacing with a 10.67 m forest
74
commences is at the position where overland flow encounters the buffer. To demonstrate
the importance of a forested buffer, a simulation was conducted with a 5 m buffer on a
10% slope with a water bar spacing of 17 m. The average annual sediment yield for this
slope was 63.79 t/ha. This same slope with a forested buffer of 20 m produced an
average annual sediment yield of 16.78 t/ha. Simulations were also conducted in which
the water bar spacing and slope were held constant at 14 m and 10%, respectively, and
the buffer width increased while recording the average annual sediment load. Figure 4.21
represents average annual sediment load vs. buffer length. From the figure, it is noted
that as the forest buffer width increases, there is exponential decrease in average annual
sediment load.
Figure 4.20 ? Annual Average Soil Loss and Yield with 10.7 m forest buffer
75
0
10
20
30
40
50
60
70
0 5 10 15 20 25 30 35 40 45
Buffer Length (m)
A
v
er
a
g
e
A
n
n
u
al
S
e
di
m
e
n
t
Lo
ad
(
t
/
h
a
)
Soil Amendment Assessment
Data collection on trail condition, specifically the formation of ruts, was collected
on four plots between October 2004 and January 2005. Two plots were shaped and
treated with the soil amendment Envirotac II, and the other two were only shaped and
used as controls. Measurements were collected every two weeks and consisted of cross-
section profiles taken at three to four locations along the trail sections. The first and last
measurements were plotted together on a chart in order to visually represent soil loss or
deposition due to the formation of ruts. Through a series of triangle calculations, the
differences between the initial and final trail shapes were determined so that net soil
losses or depositions could be reported. Net soil losses or depositions are reported as
cross-sectional area differences in square centimeters for each profile. In areas where
Figure 4.21 ? Annual Average Soil Loss vs. Forest Buffer Length
76
turn-outs were present, profile measurements were extended to determine if deposition
was occurring in the turn-out. Figures 4.22 and 4.23 represent cross-section profiles for
the upper trail and lower trail sections, respectively, of Control Plot A. Similarly, Figures
4.24 and 4.25 represent upper and lower trail sections for Treatment Plot A. Figures for
Control and Treatment Plot B are shown in Appendix E.
77
White Control A
Cross-Section 1
Running Surface
Turn-out
-80
-60
-40
-20
0
20
100 400
Horizontal (cm)
Ve
r
t
i
c
a
l
(
c
m
)
10/18/2004 1 1/13/2005 1
White Control A
Cross-Section 4
Turn-out
Running Surface
Max Soil
Deposit
Max Rut
Depth
-80
-60
-40
-20
0
20
165 465
Horizontal (cm)
V
e
r
t
i
cal
(
c
m
)
10/18/2004 4 1/13/2005 4
Figure 4.22 ? Cross-section Profiles of Upper Section of Control Plot A
Figure 4.23 ? Cross-section Profiles of Lower Section of Control Plot A
78
White Treatment A
Cross-section 1
EmbankmentRunning Surface
-100
-80
-60
-40
-20
0
20
100 300
Hoizontal (cm)
V
e
rtic
a
l
(c
m
)
10/18/2004 3 1/13/2005 3
White Treatment A
Cross-Section-3
Turn-out
Running Surface
Max Rut Depth
Max Soil Deposit
-80
-60
-40
-20
0
20
100 400
Horizontal (cm)
V
e
r
t
ic
a
l
(c
m
)
10/18/2004 1 1/13/2005 1
Figure 4.24 ? Cross-section Profiles of Upper Section of Treatment Plot A
Figure 4.25 ? Cross-section Profiles of Lower Section of Treatment Plot A
79
The formation of ruts on the control plots was very evident. Figures 4.22 and 4.23
represent cross-section profiles along the trail for the control plot. The profile in Figure
4.22 is from the upper section of the slope where the turn-out begins. An evenly
distributed soil loss of 588 cm? is noted across the entire profile. Figure 4.23 represents
the profile for the lower section of the control plot where soil losses are very apparent in
the form of ruts. The maximum rut depth measured for this profile was 18 cm resulting
in an average soil loss of 49 cm?. Outside the running surface, in the turn-out, soil
deposition occurred resulting in a deposition height of 7 cm. The average soil loss for
this entire Control Plot A trail section was calculated to be 84 cm?.
Rutting on the treated plots was as evident as on the control plots. The profiles
for the treated plots are shown in Figures 4.24 and 4.25. Figure 4.24 represents the upper
section of the profile before the presence of the turnout. As seen in the control plot, the
upper section of the profile had a relatively even soil loss across the profile and was
measured to be 42 cm?. In the lower section of the treated plot, Figure 4.25, the
formation of one rut occurred on the running surface and had a measured depth of 11 cm.
Both control plots and treated plots had soil deposition occurring at the lower end
of the plots. In the section of the turn-out for the treated plot, a large amount of
deposition occurred with a maximum measured height of 12 cm. Due to the large
deposition measured in the turn-out, this profile reported an average soil deposition of 27
cm? rather than a soil loss, as with the other profiles. This is explained by the
understanding of how a turn-out is designed to function. As water travels down the trail
it gains energy and detaches sediment from the upper sections and transports it to the
lower sections causing deposition as the water loses energy. The turn-out widens the
80
flow path, therefore reducing the speed, which reduces the energy and results in
deposition. Table 4.12 summarizes the net and average soil loss or deposition for each
cross-section of each plot. Negative values represent soil loss and positive values
represent soil deposition in square centimeters.
Control A Treatment A Control B Treatment B
Net Cross-Section 1 -588 321 -54 -989
Net Cross-Section 2 -514 -221 139 -151
Net Cross-Section 3 -1560 -295 575 -482
Net Cross-Section 4 -1349 n.a. -415 n.a
Net Soil Loss/Depostion -4011 -196 244 -1623
Avg. Soil Loss/Deposition -11 -1 1 -5
Both treated plots reported a net and average soil loss. On Control Plot A, both
the net and average soil losses were much greater for Treatment Plot A. On Control and
Treatment Plots B, the opposite occurred. Control Plot B reported a net deposition rather
than a soil loss like Treatment Plot B reported. It is difficult to draw conclusions based
solely on the numbers reported above because of the variability between plots, so a visual
comparison between the treatment and control plots was conducted. The treated plots did
not perform as well as expected. The formation of ruts between the control plots and the
treated plots was similar. After three months of trafficking, both treated plots and control
plots were equally disturbed and in need of maintenance. This may be due to factors
regarding application technique, time and season of application, and curing time after
application. It appears that further research should be conducted regarding soil
amendments for use as a soil stabilizer on ORV trails.
Table 4.12? Net and average soil loss or deposition for each section.
81
SUMMARY AND CONCLUSIONS
Introduction
The goal of this project was to establish some Best Management Practices for
Off-Road Vehicles trails. A stream crossing of the Kentuck ORV trail system located in
the Talladega National Forest was monitored for approximately eight months containing
three different periods. The three periods were as follows: No Traffic Period,
Maintenance Period, and Trafficking Period. Water samples were taken during 16 storm
events that occurred throughout the three periods. Along with the water samples, stream
flow rate data, rainfall data, and traffic levels were monitored as well. Water samples
were analyzed in a laboratory to determine total suspended sediment that was used to
calculate sediment loads for individual storm events. Using the sediment load data and
rainfall data, an attempt to calibrate the Water Erosion Prediction Project (WEPP) model
was conducted. The WEPP model was then used to simulate various management
practices in order to develop recommendations for BMP?s. During the trafficking period,
four sample plots were constructed and Envirotac II, a soil amendment was applied to
assess it?s effectiveness in reducing the formation of ruts.
Objective 1
During the 16 storm events, a total of 121 kg of sediment was introduced to the
stream at the crossing. Storm events with a low cumulative rainfall contribute very little
82
sediment into the stream channel. However, storm events with increased intensity and an
elevated cumulative rainfall, can potentially contribute significant sediment loads to the
stream channel. Significant sediment loads result from storm events with a one year
return interval or longer. The only storm event collected that had a return period higher
than one year had a sediment load 109 kg and a cumulative rainfall of 4.9 cm. Since this
is the only storm event with a significant sediment load, a long-term study should be
conducted to analyze effects of longer return interval storms. Traffic volumes were also
monitored in this study and it was determined that an average of 25 riders per day rode
the trail system during the trafficking period.
Objective 2
The attempt to calibrate the Hillslope version of the WEPP model proved to be
difficult. The rainfall data was used effectively to create a climate file for the data
collection period. Along with the rainfall data, soils and slope data were used as input
parameters during calibration. During calibration predicted vs. measured values were
observed and a Nash-Sutcliffe R? calculated to determine performance. Since the data set
only contained one large storm event, calibration of the model could not be conducted
properly. However, using the limited data, the calibration attempt was continued and an
R? value of 0.92 was achieved using the Hillslope version of the WEPP model by
adjusting the effective hydraulic conductivity. Model performance appeared to be
accurate when calibrated to predict larger storm events. The model tended to over-
predict smaller storm events, so it was recommended that the model be used only to
predict sediment yield resulting from larger storm events.
83
Objective 3
Using the non-calibrated model, various management practices were simulated in
order to recommend BMP?s. Simulations on water bar spacing, trail gradient and
minimum buffer lengths were conducted using an allowable average annual sediment
yield of 11 t/ha. It is recommended that forested buffers always be used with a minimum
width of 20 m on each side of the stream. It is also recommended that, when possible,
slope gradient be kept below 12 %. On slopes below 12 %, a maximum water bar spacing
of 10 m is suggested and on slopes steeper than 12 %, a maximum water bar spacing of 6
m is suggested. These theoretical values are sections of trail near stream channels and for
clay loam soils which exist in the Talladega National Forest in Alabama. The values
were determined using the Hillslope version of the WEPP model. The Watershed version
of the WEPP model was determined to not be suitable for modeling erosion from ORV
trails in an effort to develop BMP?s. There are two reasons why the Watershed model
was considered to be inappropriate. First, when sections of the ORV trail are
perpendicular to the stream (i.e. at the crossing) it cannot be accurately accounted for in
the model on the watershed scale. Second, the Watershed version does not model
sediment transport in perennial streams. Due to these two factors, the best method to
model ORV trails for recommending BMP?s is to use the Hillslope version of the WEPP
model.
Objective 4
The use of the soil amendment Envirotac II was assessed to determine its
effectives in reducing the formation of ruts after trafficking had occurred. Maximum rut
84
depths on the control plots A and B were 17.8 cm and 16.2, respectively. The maximum
rut depths recorded for the plots A and B, treated with Envirotac II, were 11.43 cm and
17.78 cm, respectively. It is concluded that the plots treated with Envirotac II did not
significantly aid in reducing the formation of ruts. Better results may have been achieved
if treatments would have been applied during the warmer and drier seasons of the year
and overstory vegetation removed to aid in the hardening of the soil surface.
General Conclusion
It is concluded that ORV trails have the potential to produce levels of sediment
that may reduce water quality impairing fish habitat and shortening the life of reservoirs.
However, through the use of proper BMP?s, such as water bar spacing, slope grade, and
buffer lengths, theoretical sediment yields can be reduced so as to minimize the
degradation of water quality.
Future Research
Future research goals should consist of additional water sampling of stream
crossings influenced by ORV trails during periods of traffic, maintenance, and no traffic.
Also, by monitoring various types of stream crossings on ORV trails may help to better
understand the impact that is being created and may prove to be an important area of
research. For comparison purposes, water samples should be collected on a watershed
not affected by ORV trails or roads. Further data collection in these areas for multiple
years could, in turn, be used to better calibrate the hillslope version of the WEPP model.
Research should also be conducted to assess the effectiveness of Envirotac II at reducing
85
rut formation by studying various application techniques. This could be achieved by
conducting a study involving the effectiveness of Envirotac II under varying
environmental conditions during which temperature, soil, moisture content, lighting, and
product concentration are changed. This could lead to a better understanding on product
applicability along with product limitations.
86
BILIOGRAPHY
Baffaut, C., Nearing, M.A., Govers, G. 1998. Statistical distributions of soil loss from
runoff plots and WEPP model simulations. Soil Science Society of America Journal.
62(3): 756-763.
Blackland Research Center. SWAT Fact Sheet. 2005.
http://www.brc.tamus.edu/swat/factsheet.html. Last visited June 7, 2005.
Brinker, R.W. 1995. Forest roads and construction of associated water diversion devices.
Alabama Cooperative Extension System. ANR-916.
Brooks, E.S. GPS-assisted road surveys and GIS-based road erosion modeling using the
WEPP model. Presented at the 2003 Pacific Northwest Region meeting sponsored by
ASAE. Paper No. PNW-03-109.
Croke, J.C., Hairsine, P.B. Management of road runoff: a design approach. Presented at
the Soil Erosion Research for the 21
st
Century, Proc. Int. Symposium. January 2001.
ASAE: 701P0007.
Davidson, E., Fox, M. 1974. Effects of off-road motorcycle activity on Mojave Desert
vegetation and soil. Madrono 22:381-390.
Davis, J.N. 2004. Assessment of chemical soil amendments for erosion control on off-
road vehicle trail. Auburn, AL. MS thesis, Auburn University
Denver Instruments Balance web page,
http://www.denverinstrumentusa.com/media/pdf/literature/apex.pdf. Last visited June 9,
2005.
Eav, L. Unmanaged recreation. Key Messages and Talking Points from the Four Threats
to Leadership Information Packet, USDA Forest Service, Office of Communication ?
Washington D.C.
Eckert, Jr., R.E., Wood, M.K., Blackburn, W.H., Peterson, F.F. Impacts of off-road
vehicles on infiltration rates and sediment production of two desert soils. Journal of
Range Management. 32(5), September 394-397.
87
Egan, A.F. 1999. Forest roads: where soil and water don?t mix. Journal of Forestry. 97:
8, 18-21.
Elliot, W.J., Foltz, R.B., Luce, C.H. 1999. Modeling low-volume road erosion.
Transportation Research Record. 1652 V. 2: 244-249.
Elliot, W.J., Foltz, R.B., Robichaud, P.R. 1994. A tool for estimating disturbed forest site
sediment production. Interior Cedar-Hemlock-White Pine Forests: Ecology and
Management. Washington State University Cooperative Extension 233-236.
Elliot, W.J., Hall, D.E., Graves, S.R. 1999. Predicting sedimentation from forest roads.
Journal of Forestry. August: 23-29.
Elliot, W.J., Luce, C.H., Robichaud, P.R. Predicting sedimentation from timber harvest
areas with the WEPP model. Presented at the Sixth Federal Interagency Sedimentation
Conference, March 10-14, 1996. IX: 46-53.
Envirotac II: soil stabilizer. http://envirotac.us/new/Description.shtml. Last visited April
13, 2004.
Flanagan, D.C., Renschler, C.S., Cochrane, T.A. Application of the WEPP model with
digital geographic information. Presented at the 4
th
International Conference on
Integrating GIS and Environmental Modeling, September 2-8, 2000. GIS/EM4 No. 149.
Fogg, G.E. 2002. Park Guidelines for OHV?s. National Off-Highway Vehicle
Conservation Council.
Franson, M.H. 1985. Standard Methods for the Examination of Water and Wastewater.
Port City Press, Maryland.
Laflen, J.M., Elliot, W.J., Simanton, J.R., Holzhey, C.S., Kohl, K.D. 1991. WEPP soil
erodibility experiments for rangeland and cropland soils. Journal of Soil and Water
Conservation. January-February: 39-44.
Laflen, J.M., Lane, L.J., Foster, G.R. 1991. WEPP a new generation of erosion
prediction technology. Journal of Soil and Water Conservation. January-February: 34-
38.
Liban, C.B. 1998. Non-point source pollution. National Forum. 78:3. 7-9
Lull, H.W. 1959. Soil Compaction on forest and range lands. U.S. Department of
Agriculture, Forest Service. Misc. Pub. No. 768. Washington D.C. 33 p.
Miller, G.A., Azad, S. 2000. Influence of soil type on stabilization with cement kiln dust.
Construction and Building Materials. 14. 89-97.
88
Morfin, S., Elliot, B., Foltz, R., Miller, S. Predicting effects of climate, soil and
topography on road erosion with WEPP. Presented at the 1996 ASAE Annual
International meeting, Paper No. 965016. ASAE, 2950 Niles Road, St. Joseph, MI
49085-9659 USA.
NRCS HecRas Training.
http://www.ndcsmc.nrcs.usda.gov/Training/BASICHECRAS.html Last visited June 7,
2005.
Patric, J.H. 1978. Harvesting effects on soil and water in the in eastern hardwood forests.
Southern Journal of Applied Forestry. 2: 3, 66-73.
Porbaha, A., Pradham, T.B.S., Yamane, N. 2000. Time effect of shear strength and
permeability of fly ash. Journal of Energy Engineering. April:15-31.
Rhee. H., Fridley, J.L., Foltz, R.B. 2004. Modeling erosion from unpaved forest roads at
various levels of geometric detail using the WEPP model. Transactions of the ASAE.
47(3): 961-968.
Sawyer, T. 2002. Military construction: High-tech tools and hard, hard work at FOB
Rhino. McGraw-Hill Construction ENR.
http://www.enr.construction.com/features/technologyeconst/archives/020225c.asp#. Last
visited April 13, 2004.
Schwab, G.O., Fangmeier, D.D., Elliot, W.J., Frevert, R.K. 1993. Soil and Water
Conservation Engineering. John Wiley and Sons, Inc. New York.
Soil Ecology and Research Group. 2002. San Clemente Island ATV trail system-
installation of erosion control demonstrations.
http://www.serg.sdsu.edu/SERG/restorationproj/channel_islands/erosion/project%20repo
rt.html Last visited June 8, 2005.
Sojka, R.E., Lentz, R.D., Trout, T.J., Ross, C.W., Bjorneberg, D.L., Aase, J.K. 1998.
Polyacrylamide effects on infiltration in irrigated agriculture. Journal of Soil and Water
Conservation. 53:4. 325-333.
Sparrow, S.D., Wooding, F.J., Whiting, E.H. 1978. Effects of off-road vehicle traffic on
soils and vegetation in the Denali Highway region of Alaska. Journal of Soil and Water
Conservation. January-February: 20-27.
Spruill, C.A., Workman, S.R., Taraba, J.L. 2000. Simulation of daily and monthly stream
discharge from small watersheds using the SWAT model. Transactions of the ASAE.
43(6): 1431-1349.
89
Strom, J.M., Wilkins, M. 1990. Environmental assessment for the management of off-
road vehicles in the Talladega National Forest. Shoal Creek and Talladega Ranger
Districts. National Forests in Alabama.
Tysdal. L.M., Elliot, W.J., Luce, C.H., Black, T.A. 1999. Modeling erosion from
insloping low-volume roads with WEPP watershed model. Transportation Research
Record. Vol. 2, No. 1652: 250-256.
Washington State Department of Transportation. Polyacrylamide (PAM) application for
soil erosion control. Environmental Affairs Office. Water Quality Program.
Weaver, T., Dale, D. 1978. Trampling effects of hikers, motorcycles and horses in
meadows and forests. The Journal of Applied Ecology. 15(2): 451-457.
Wilshire, H.G., Nakata, J.K. 1976. Off-road vehicle effects on California?s Mojave
Desert. California Geol. 29:123-130.
Wu, L. 2001. Polyacrylamide (PAM) ? effective erosion fighter and infiltration enhancer
but not a conserver of water. http://esce.ucr.edu/soilwater/spring_2001.htm. Last visited
April 19, 2004
90
APPENDIX A
TRAIL CLOSED PERIOD
91
0
0.25
0.5
0.75
1
1.25
1.5
1.75
2
2.25
2.5
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time Since Start of Sampling (hrs)
C
u
mm
u
l
at
i
ve Rai
n
f
a
l
l
(
c
m)
0.00
0.20
0.40
0.60
0.80
1.00
1.20
1.40
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time (hrs)
Fl
ow
r
a
t
e
(
l
/s
)
Figure A.1 ? Cumulative Rainfall for January 25, 2004 Storm Event
Figure A.2 ? Hydrograph for January 25, 2004 Storm Event
92
0
20
40
60
80
100
120
140
160
1234567891011213141516171819202122324
Time (hrs)
Tot
a
l
Sus
p
e
nde
d Se
dim
e
nt
(
m
g/l)
Site A Site B
0
0
0
0
0
0
0
0
0
1234567891011213141516171819202122324
Time (hrs)
S
e
di
m
e
nt
Lo
a
d
(
k
g
/
hr
)
Site A Site B
Figure A.4 ? Sediment Load for January 25, 2004 Storm Event
Figure A.3 ? TSS for January 25, 2004 Storm Event
93
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
5.5
01234567891011213141516171819202122324
Time Since Start of Sampling (hrs)
Cumm
ula
t
i
ve R
a
inf
a
ll (cm)
0
5
10
15
20
25
1234567891011213141516171819202122324
Time (hrs)
Flow
r
a
t
e
(
l/s
)
Figure A.5 ? Cumulative Rainfall for February 6, 2004 Storm Event
Figure A.6 ? Hydrograph for February 6, 2004 Storm Event
94
0
200
400
600
800
1000
1200
1400
1600
1800
2000
1234567891011213141516171819202122324
Time (hrs)
T
o
t
a
l S
u
s
p
e
nde
d
S
e
di
m
e
n
t
(
m
g/
l
)
Site A Site B
0
10
20
30
40
50
60
70
80
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time (hrs)
S
e
d
i
m
e
nt
Loa
d
(
k
g/
hr
)
Site A Site B
Figure A.7 ? TSS for February 6, 2004 Storm Event
Figure A.8 ? Sediment Load for February 6, 2004 Storm Event
95
0
0.2
0.4
0.6
0.8
1
1.2
01234567891011213141516171819202122324
Time Since Start of Sampling (hrs)
C
u
m
m
u
la
tiv
e
R
a
i
n
fa
ll (
c
m
)
0.0
0.5
1.0
1.5
2.0
2.5
3.0
1234567891011213141516171819202122324
Time (hrs)
Flow
r
a
t
e
(
l
/s
)
Figure A.9 ? Cumulative Rainfall for February 25, 2004 Storm Event
Figure A.10 ? Hydrograph for February 25, 2004 Storm Event
96
0
5
10
15
20
25
1 2 3 4 5 6 7 8 9 101112131415161718192021222324
Time (hrs)
Tot
al
Suspended S
e
d
i
ment
(mg/
l
)
Site A Site B
0.0
0.0
0.0
0.1
0.1
0.1
0.1
0.1
0.2
1 2 3 4 5 6 7 8 9 101 12131415161718192021222324
Time (hrs)
S
e
di
m
e
n
t
Lo
a
d
(
k
g/
hr
)
Site A Site B
Figure A.11 ? TSS for February 25, 2004 Storm Event
Figure A.12 ? Sediment Load for February 25, 2004 Storm Event
97
APPENDIX B
MAINTENANCE PERIOD
98
0
0.1
0.2
0.3
0.4
0.5
0.6
0 1 2 3 4 5 6 7 8 9 101112131415161718192021222324
Time Since Start of Sampling (hrs)
Cu
mm
u
l
at
i
ve Rai
n
f
a
l
l
(
c
m
)
0.0
0.1
0.1
0.2
0.2
0.3
0.3
0.4
0.4
0.5
0.5
1 2 3 4 5 6 7 8 9 101112131415161718192021222324
Time (hrs)
F
l
ow
r
a
te
(l/
s
)
Figure B.1 ? Cumulative Rainfall for March 16, 2004 Storm Event
Figure B.2? Hydrograph for March 16, 2004 Storm Event
99
0
10
20
30
40
50
60
70
80
1 2 3 4 5 6 7 8 9 101112131415161718192021222324
Time (hrs)
Tot
a
l
S
u
s
p
e
nde
d S
e
dim
e
nt
(
m
g/
l)
Site A
Site B
0.0
0.0
0.0
0.1
0.1
0.1
0.1
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time (hrs)
S
e
di
m
e
n
t
Lo
a
d
(
k
g/
h
r
)
Site A Site B
Figure B.3? TSS for March 16, 2004 Storm Event
Figure B.4? Sediment Load for March 16, 2004 Storm Event
100
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time Since Start of Sampling (hrs)
C
u
mmula
t
i
v
e
R
a
inf
a
ll
(
c
m)
0.0
0.5
1.0
1.5
2.0
2.5
1 2 3 4 5 6 7 8 9 101112131415161718192021222324
Time (hrs)
F
l
ow
r
a
te
(l/
s
)
Figure B.5? Cumulative Rainfall for March 20, 2004 Storm Event
Figure B.6? Hydrograph for March 20, 2004 Storm Event
101
0
20
40
60
80
100
120
140
160
180
1 2 3 4 5 6 7 8 9 101112131415161718192021222324
Time (hrs)
T
o
t
a
l
S
u
s
p
e
nde
d
S
e
di
m
e
nt
(
m
g/
l
)
Site A Site B
0.0
0.1
0.2
0.3
0.4
0.5
0.6
1 2 3 4 5 6 7 8 9 101112131415161718192021222324
Time (hrs)
S
e
di
m
e
nt
L
o
a
d
(
k
g/
hr
)
Site A Site B
Figure B.7? TSS for March 20, 2004 Storm Event
Figure B.8? Sediment Load for March 20, 2004 Storm Event
102
0
0.5
1
1.5
2
2.5
01234567891011213141516171819202122324
Time Since Start of Sampling (hrs)
C
u
m
m
ul
a
t
ive
R
a
in
f
a
ll
(
c
m
)
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
1 2 3 4 5 6 7 8 9 101112131415161718192021222324
Time (hrs)
F
l
ow
r
a
te
(l/
s
)
Figure B.9 ? Cumulative Rainfall for March 29, 2004 Storm Event
Figure B.10 ? Hydrograph for March 29, 2004 Storm Event
103
0
5
10
15
20
25
30
35
1234567891011213141516171819202122324
Time (hrs)
Tot
a
l
S
u
s
p
e
nde
d S
e
dim
e
nt
(
m
g/l
)
Site A Site B
0.0
0.1
0.1
0.2
0.2
0.3
0.3
0.4
1 2 3 4 5 6 7 8 9 101112131415161718192021222324
Time (hrs)
S
e
d
i
m
e
nt
Loa
d
(
k
g/
hr
)
Site A Site B
Figure B.11 ? TSS for March 29, 2004 Storm Event
Figure B.12 ? Sediment Load for March 29, 2004 Storm Event
104
APPENDIX C
TRAIL OPEN PERIOD
105
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time Since Start of Sampling (hrs)
C
u
m
m
u
l
at
i
ve Rai
n
f
a
l
l
(
c
m
)
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1234567891011213141516171819202122324
Time (hrs)
Fl
ow
r
a
t
e
(
l
/
s
)
Figure C.1 ? Cumulative Rainfall for April 11, 2004 Storm Event
Figure C.2 ? Hydrograph for April 11, 2004 Storm Event
106
0
20
40
60
80
100
120
140
160
180
200
1 2 3 4 5 6 7 8 9 101112131415161718192021222324
Time (hrs)
Tot
a
l
S
u
s
p
e
n
d
e
d
S
e
di
m
e
nt
(
m
g
/
l
)
Site A Site B
0.0
0.1
0.1
0.2
0.2
0.3
1234567891011213141516171819202122324
Time (hrs)
S
e
di
m
e
n
t
Lo
a
d
(
k
g/
h
r
)
Site A Site B
Figure C.3 ? TSS for April 11, 2004 Storm Event
Figure C.4 ? Sediment Load for April 11, 2004 Storm Event
107
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time Since Start of Sampling (hrs)
Cu
m
m
u
l
at
i
ve Rai
n
f
a
l
l
(
c
m)
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 1 2 3 4 5 6 7 8 9 101112131415161718192021222324
Time (hrs)
Fl
ow
r
a
t
e
(
l
/
s
)
Figure C.5 ? Cumulative Rainfall for April 26, 2004 Storm Event
Figure C.6 ? Hydrograph for April 26, 2004 Storm Event
108
0
20
40
60
80
100
120
140
160
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time (hrs)
Tot
a
l
S
u
s
p
e
n
de
d S
e
dim
e
nt
(
m
g
/
l)
Site A Site B
0.0
0.0
0.0
0.1
0.1
0.1
0.1
0.1
1 2 3 4 5 6 7 8 9 101112131415161718192021222324
Time (hrs)
S
e
di
m
e
n
t
Lo
a
d
(
k
g/
hr
)
Site A Site B
Figure C.7 ? TSS for April 26, 2004 Storm Event
Figure C.8 ? Sediment Load for April 26, 2004 Storm Event
109
0
0.5
1
1.5
2
2.5
3
3.5
4
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time Since Start of Sampling (hrs)
C
u
m
m
u
la
tiv
e
R
a
i
n
fa
ll (
c
m
)
0.0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
1234567891011213141516171819202122324
Time (hrs)
Fl
ow
r
a
t
e
(
l
/
s
)
Figure C.9 ? Cumulative Rainfall for May 16, 2004 Storm Event
Figure C.10 ? Hydrograph for May 16, 2004 Storm Event
110
0
50
100
150
200
250
300
350
400
450
500
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time (hrs)
Tot
a
l
S
u
s
p
e
nde
d S
e
d
i
m
e
nt
(
m
g/l
)
Site A Site B
0.0
2.0
4.0
6.0
8.0
10.0
12.0
14.0
16.0
1 2 3 4 5 6 7 8 9 101112131415161718192021222324
Time (hrs)
S
e
dim
e
nt
Loa
d
(
k
g/hr
)
Site A Site B
Figure C.11 ? TSS for May 16, 2004 Storm Event
Figure C.12 ? Sediment Load for May 16, 2004 Storm Event
111
0
0.2
0.4
0.6
0.8
1
1.2
1.4
01234567891011213141516171819202122324
Time Since Start of Sampling (hrs)
C
u
mm
u
l
at
i
ve Rai
n
f
a
l
l
(
i
n
)
0
1
1
2
2
3
3
4
4
5
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time (hrs)
Fl
ow
r
a
t
e
(
l
/
s
)
Figure C.13 ? Cumulative Rainfall for June 16, 2004 Storm Event
Figure C.14 ? Hydrograph for June 16, 2004 Storm Event
112
0
100
200
300
400
500
600
700
1234567891011213141516171819202122324
Time (hrs)
T
o
t
a
l
S
u
s
p
e
nde
d
S
e
di
m
e
nt
(
m
g/
l
)
Site A Site B
0
1
2
3
4
5
6
7
8
9
1 2 3 4 5 6 7 8 9 101112131415161718192021222324
Time (hrs)
S
e
di
m
e
nt
L
o
a
d
(
k
g/
hr
)
Site A Site B
Figure C.15 ? TSS for June 16, 2004 Storm Event
Figure C.16 ? Sediment Load for June 16, 2004 Storm Event
113
0
0.2
0.4
0.6
0.8
1
1.2
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time Since Start of Sampling (hrs)
C
u
m
m
u
lat
i
v
e
R
a
i
n
fa
ll (
c
m
)
0
0
0
0
0
1
1
1
1
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time (hrs)
Fl
ow
r
a
t
e
(
l
/
s
)
Figure C.17 ? Cumulative Rainfall for June 22, 2004 Storm Event
Figure C.18 ? Hydrograph for June 22, 2004 Storm Event
114
0
50
100
150
200
250
300
1 2 3 4 5 6 7 8 9 101112131415161718192021222324
Time (hrs)
Tot
a
l
S
u
s
p
e
nde
d S
e
dim
e
nt
(
m
g/l
)
Site A Site B
0.0
0.1
0.1
0.2
0.2
0.3
0.3
0.4
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time (hrs)
Se
d
i
m
e
nt Lo
a
d
(k
g/
hr)
Site A Site B
Figure C.19 ? TSS for June 22, 2004 Storm Event
Figure C.20 ? Sediment Load for June 22, 2004 Storm Event
115
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
01234567891011213141516171819202122324
Time Since Start of Sampling (hrs)
C
u
m
m
ul
a
t
iv
e R
a
in
f
a
ll
(
c
m
)
0
0
0
0
0
1
1
1
1 2 3 4 5 6 7 8 9 101112131415161718192021222324
Time (hrs)
Fl
ow
r
a
t
e
(
l
/s
)
Figure C.21 ? Cumulative Rainfall for July 2, 2004 Storm Event
Figure C.22 ? Hydrograph for July 2, 2004 Storm Event
116
0
100
200
300
400
500
600
700
1234567891011213141516171819202122324
Time (hrs)
To
t
a
l S
u
s
p
e
nde
d
Se
dim
e
nt
(
m
g/l)
Site A Site B
0.0
0.1
0.1
0.2
0.2
0.3
0.3
0.4
0.4
0.5
1 2 3 4 5 6 7 8 9 1011121314151617181920212 2324
Time (hrs)
S
e
di
m
e
nt
L
o
a
d
(
k
g/
hr
)
Site A Site B
Figure C.23 ? TSS for July 2, 2004 Storm Event
Figure C.24 ? Sediment Load for July 2, 2004 Storm Event
117
0
0.2
0.4
0.6
0.8
1
1.2
01234567891011213141516171819202122324
Time Since Start of Sampling (hrs)
Cu
m
m
u
l
at
i
ve Rai
n
f
a
l
l
(
c
m)
0
0
0
0
0
1
1
1
1
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time (hrs)
Fl
ow
r
a
t
e
(
l
/
s
)
Figure C.25 ? Cumulative Rainfall for December 16, 2003 Storm Event
Figure C.26 ? Hydrograph for December 16, 2003 Storm Event
118
0
20
40
60
80
100
120
140
1234567891011213141516171819202122324
Time (hrs)
Tot
a
l
S
u
s
p
e
nde
d S
e
dim
e
nt
(
m
g/
l)
Site A Site B
0.0
0.0
0.0
0.0
0.0
0.1
0.1
0.1
1 2 3 4 5 6 7 8 9 101112131415161718192021222324
Time (hrs)
Se
d
i
m
e
nt Lo
a
d
(k
g/
hr)
Site 1 A Site 1 B
Figure C.27 ? TSS for December 16, 2003 Storm Event
Figure C.28 ? Sediment Load for December 16, 2003 Storm Event
119
APPENDIX D
TRAFFICKING DATA
120
Date Traffic Date Traffic
2-wheeler 4-wheeler Other Total 2-wheeler 4-wheeler Other Total
7-Aug 4 12 2 18 20-Dec 5 29 4 38
8-Aug 11 10 3 24 21-Dec 13 29 0 42
9-Aug 52 111 5 168 22-Dec 7 8 0 15
10-Aug 34 44 1 79 23-Dec 2 1 0 3
11-Aug 11 32 2 45 24-Dec 0 8 2 10
12-Aug 2 11 0 13 25-Dec 2 2 0 4
13-Aug 3 16 0 19 26-Dec 7 18 0 25
14-Aug 1 0 0 1 16-Apr 4 3 2 9
15-Aug 2 33 0 35 17-Apr 28 107 6 141
16-Aug 13 90 4 107 18-Apr 20 38 4 62
17-Aug 1 41 0 42 19-Apr 6 5 1 12
13-Oct 4 8 0 12 21-Apr 0 1 0 1
14-Oct 1 5 0 6 22-Apr 1 0 0 1
15-Oct 21 59 1 81 23-Apr 6 14 0 20
16-Oct 11 33 11 55 24-Apr 10 14 2 26
17-Oct 10 8 0 18 25-Apr 10 60 1 71
18-Oct 0 2 0 2 26-Apr 0 1 0 1
19-Oct 2 5 0 7 28-Apr 1 0 0 1
20-Oct 5 0 0 5 29-Apr 11 5 0 16
21-Oct 5 0 0 5 30-Apr 0 4 0 4
22-Oct 4 34 4 42 1-May 7 31 0 38
23-Oct 0 21 2 23 2-May 4 7 0 11
11-Nov 2 7 20 29 3-May 2 0 0 2
12-Nov 0 0 0 0 4-May 2 3 0 5
13-Nov 1 31 1 33 5-May 2 0 0 2
14-Nov 13 81 3 97 20-May 0 1 0 1
15-Nov 1 45 1 47 21-May 1 17 0 18
23-Nov 9 21 2 32 22-May 3 15 0 18
24-Nov 0 0 0 0 23-May 3 10 0 13
25-Nov 1 5 1 7 18-Jun 0 12 0 12
26-Nov 1 25 1 27 19-Jun 2 16 2 20
27-Nov 0 0 0 0 20-Jun 4 13 4 21
28-Nov 18 26 2 46 21-Jun 0 2 0 2
29-Nov 59 95 3 157 22-Jun 1 2 0 3
30-Nov 18 29 2 49 23-Jun 0 9 0 9
4-Dec 0 2 0 2 25-Jun 0 13 1 14
6-Dec 0 9 0 9 26-Jun 10 51 2 63
7-Dec 6 29 1 36 27-Jun 0 8 0 8
8-Dec 0 3 1 4 28-Jun 0 3 0 3
9-Dec 2 3 0 5 29-Jun 1 9 1 11
12-Dec 0 1 0 1
13-Dec 0 7 0 7
14-Dec 3 11 0 14 Total Days Traffic Totals Total
15-Dec 0 4 0 4 2-wheeler 4-wheeler Other
16-Dec 0 1 0 1 87 506 1586 108 2200
17-Dec 0 3 0 3
19-Dec 0 4 3 7 Average 5.82 18.23 1.24 25.29
121
APPENDIX E
CUMULATIVE RAINFALL AND SEDIMENT LOADING
122
Storm Date Rainfall Sediment Load
(cm) (kg)
12/16/2003 0.97 0.17
1/25/2004 2.24 0.13
2/6/2004 4.88 108.98
2/12/2004 2.03 0.14
2/25/2024 1.12 0.11
3/6/2004 2.03 4.01
3/16/2004 0.51 0.08
3/20/2004 0.91 0.42
3/29/2004 2.01 0.17
4/11/2004 1.35 0.53
4/26/2004 1.75 0.36
4/30/2004 2.24 2.18
5/16/2004 3.61 2.02
6/16/2004 1.27 0.83
6/22/2004 1.02 0.60
7/2/2004 1.60 0.19
123
APPENDIX F
SOIL AMENDMENT ASSESSMENT
124
White Treatment A
Cross-section 1
EmbankmentRunning Surface
-100
-80
-60
-40
-20
0
20
100 300
Hoizontal (cm)
Ve
r
t
i
c
a
l
(
cm
)
10/18/2004 3 1/13/2005 3
White Treatment A
Cross-Section 2
Turn-out
Running Surface
-80
-60
-40
-20
0
20
100 400
Horizontal (cm)
V
e
r
t
i
cal
(
cm
)
10/18/2004 2 1/13/2005 2
White Treatment A
Cross-Section-3
Turn-out
Running Surface
Max Rut Depth
Max Soil Deposit
-80
-60
-40
-20
0
20
100 400
Horizontal (cm)
V
e
r
t
i
cal
(
c
m
)
10/18/2004 1 1/13/2005 1
Figure E.3 ? Cross-section of lower-slope
of Treatment A
Figure E.2 ? Cross-section of mid-slope
of Treatment A
Figure E.1 ? Cross-section of upper slope
of Treatment A
125
White Control A
Cross-Section 1
Running Surface
Turn-out
-80
-60
-40
-20
0
20
100 400
Horizontal (cm)
V
e
r
t
ic
a
l
(
c
m
)
10/18/2004 1 1/13/2005 1
White Control A
Cross-Section 2
Turn-out
Running Surface
-80
-60
-40
-20
0
20
100 400
Horizontal (cm)
Ve
r
t
i
c
a
l
(
cm
)
10/18/2004 2 1/13/2005 2
White Control A
Cross-Section 3
Turn-out
Running Surface
-80
-60
-40
-20
0
20
75 375
Horizontal (cm)
V
e
r
t
i
cal
(
c
m
)
10/18/2004 3 1/13/2005 3
White Control A
Cross-Section 4
Turn-out
Running Surface
Max Soil
Deposit
Max Rut
Depth
-80
-60
-40
-20
0
20
165 465
Horizontal (cm)
V
e
r
t
ic
a
l
(
c
m
)
10/18/2004 4 1/13/2005 4
Figure E.5 ? Cross-section of mid-
slope of Treatment A
Figure E.6 ? Cross-section of mid-
slope of Treatment A
Figure E.7 ? Cross-section of lower-
slope of Treatment A
Figure E.4 ? Cross-section of upper-
slope of Treatment A
126
White Treatment B
Cross-Section 1
Running Surface
Turn-out
-60
-40
-20
0
20
50 350
Horizontal (cm)
V
e
r
t
ic
a
l
(
c
m
)
10/18/2004 1/13/2005
White Treatment B
Cross-Section 2
Turn-out
Running Surface
-80
-60
-40
-20
0
20
50 350
Horizontal (cm)
V
e
r
t
ic
a
l
(
c
m
)
10/18/2004 1/13/2005
White Treatment B
Cross-Section 3
Turn-out
Running Surface
-80
-60
-40
-20
0
20
100 400
Horizontal (cm)
Ver
t
i
ca
l
(
c
m
10/18/2004 1/13/2005
Figure E.10 ? Cross-section of lower-slope
of Treatment A
Figure E.9 ? Cross-section of mid-slope
of Treatment A
Figure E.8 ? Cross-section of upper-slope
of Treatment A
127
White Control B
Cross-Section 1
Turn-out
Running Surface
-80
-60
-40
-20
0
20
50 350
Horizontal (cm)
Ver
t
i
cal
(
c
m
)
10/18/2004 1/13/2005
White Control B
Cross-Section 2
Turn-out
Running Surface
-80
-60
-40
-20
0
20
50 350
Horizontal (cm)
Ve
r
t
i
c
al
(
c
m
)
10/18/2004 1/13/2005
White Control B
Cross-Section 3
Turn-out
Running Surface
-80
-60
-40
-20
0
20
50 350
Horizontal (cm)
V
e
r
t
i
cal
(
c
m
)
10/18/2004 1/13/2005
White Control B
Cross-Section 4
Running Surface
Turn-out
-100
-80
-60
-40
-20
0
20
75 375
Horizontal (cm)
V
e
r
t
ic
a
l
(
c
m
)
10/18/2004
1/13/2005
Si 3
Figure E.11 ? Cross-section of upper-slope
of Treatment A
Figure E.14 ? Cross-section of
lower- slope of Treatment A
Figure E.13 ? Cross-section of mid-slope
of Treatment A
Figure E.11 ? Cross-section of mid-slope
of Treatment A