Development of a buffer zone evaluation model /procedure

[From the U.S. Government Printing Office, www.gpo.gov]

FY'91 Task 11

Final Product
VA Coastd Resources Mgt. Twgram

DEVELOPMENT OF A BUFFER

ZONE EVALUATION

MODEL/PROCEDURE

Theo A. Dillaha

November 1992

Department of Agricultural Engineering
Virginia Polytechnic Institute and State University
w
Blacksburg, VA 24061-0303

FINAL REPORT
@DEVELOPMENT OF A BUFFER ZONE
EVALUATION MODEL/PROCEDURE

Theo A. Dillaha
Associate Professor
Department of Agricultural Engineering
Virginia Polytechnic Institute and State University
Blacksburg, VA 24061-0303
Phone: (703) 231-6813

This project was funded, in part, by the Virginia Council on the
Environment's Coastal Resources Management Program through Grant
#NA170ZO359-01 of the National Oceanic and Atmospheric
Administration, Of f ice of Ocean and Coastal Resource Management,
under the Coastal Zone Management Act of 1972 as ammended.

This project was funded, in part, by the Virginia Department of
Conservation and Recreation, Division of Soil and water
Conservation through Contract #C199-50311-92-03.

V November 1992

TABLE OF CONTENTS

Page

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 1

Literature Review . . . . . . . . . . . . . . . . . . . . . . . 3
Sediment Transport in Vegetative Buffers . . . . . . . . . . 4
Nutrient Transport in Vegetative Buffers . . i . . . . . . . 5
Vegetative Buffer Research in Virginia . . . . . . . . . . . 7
Vegetative Buffer Models and Design Procedures . . . . . . . 9
Kentucky Filter Strip Model . . . . . . . . . . . . . . . 9
CREAMS Model . . . . . . . . . . . . . . . . . . . . . . 10
GRAPH Model . . . . . . . . . . . . . . . . . . . . . . 10
Phillips Model . . . . . . . . . . . . . . . . . . . . . 11
Phillips Hydraulic Model . . . . . . . . . . . . . . 11
Phillips Detention Model . . . . . . . . . . . . . . 13
Other Design Approaches . . . . . . . . . . . . . . . . 14
State and Federal Vegetative Buffer Programs . . . . . . . 14
Chesapeake Bay Preservation Act . . . . . . . . . . . . 14
Conservation Reserve Program . . . . . . . . . . . . . . 15
National Vegetative Filter Strip Conservation
Practice Standard . . . . . . . . 16
Sediment and Nutrient Loadings due to E;od'in*g*Banis* . . . 17
Literature Review Summary . . . . . . . . . . . . . . . . 18

Model Development . . . . . . . . . . . . . . . . . . . . . . 20
Phillips Hydraulic Model . . . . . . . . . . . . . . . . . 20
Phillips Detention Model . . . . . . . . . . . . . . . . . 21
Shoreline Erosion Model . . . . . . . . . . . . . . . . . 22
Shoreline Erosion . . . . . . . . . . . . . . . . . . . 22
Upland Erosion . . . . . . . . . . . . . . . . . . . . . 23
Impact of Shoreline Stabilization . . . . . . . . . . . 24

Buffer Evaluation Procedure . . . . . . . . . . . . . . . . . 26

Data for Buffer Evaluation Procedure . . . . . . * * * 28
Data Requirements for the Hydraulic and Detenti@n*M@dels . 28
Saturated Hydraulic Conductivity, K . . . . . . . . . . 28
Buffer Length, L . . . . . . . . . . . . . * 0 . 0 9 0 . 29
Upslope Slope-Length Contributing Runoff to the
Buffer, L . . . . . . . . . . . . . . . . . . . . . . 29
sin 6, s . . . . . . . . . . . . . . . . . . . . . . . 29
Manning's Roughness Coefficient, n . . . . . . . . . . . 30
Fraction of Surface Runoff Crossing Buffer as Sheet
Flow, C . . . . . . . . . . . . . . . . . . . . . . . 31
Soil Moisture Storage Capacity, M . . . . . . . . . . . 32
Vegetative Uptake or Net Productivity Factor, V . . . . 32

Page

Data Requirements for Shoreline Stabilization Model . . . 33
Bank or Shoreline Erosion Rates . . . . . . . . . . . . 33
Bank Height . . . . . . . . . . . . . . . . . . . . . . 33
Bulk Density of Bank Soil . . . . . . . . . . . . . . . 34
Concentration of Nitrogen and Phosphorus in the
Bank Material . . ................................ . . . . . 34
Sediment Loading to the Buffer From the Upland
Contributing Area . . . . . . . . . . . . . . . . . . 34
Nitrogen and Phosphorus Loading to the Buffer . . . . . 34

Practical Application . . . . . . . . . . . . . . . . . . . . 35
Example Problem I . . . . . . . . . . . . . . . . . . . . 35
Problem Statement . . . . . . 35
Data . . . . . . . 35
Solution . . . . . . . . . 35
Example Problem II . . . . . ..................................37
Problem Statement . . . . .................................37
Data . . . . . . . . . . . . . . . . . . . . . . . . . . 37
Solution . ......................................................37
Sediment and Nutrient Losses Due to Shoreline Erosion 38
Sediment and Nutrient Transport Through the Buffer 38

Summary and Conclusions . . . . . . . . . . . . . 40

References . . . . . . . . . . . . . . . . . . . 41

LIST OF TABLES

page

Table 1. Estimation of hydraulic conductivity values
from Soil Survey permeability estimates . . . . 28

Table 11. Hydraulic conductivity as a function of soil
texture . . . . . . . . . . . . . . . . . . . . 29

Table III. Estimates of Sin 0 for various buffer
slopes in percent and degrees . . . . . . . . . 30

Table IV. Estimates of Mannings roughness coefficient
as a function of landuse and condition for
shallow sheet flow conditions (Engman, 1986) 31

Table V. Net primary productivity (adapted from:
Leith, 1975) . . . . . . . . . . . . . . . . . . 33

EXECUTIVE SUMMARY

A procedure is presented for evaluating the impacts of proposed
vegetative buffer modifications on buffer effectiveness. The
procedure is based on the hydraulic and detention models developed
by Phillips for evaluating buffer effectiveness. Phillips's
original models were modified to correct several limitations
encountered. The modified models consider the effects of
concentrated flow and vegetative uptake on buffer performance.

The proposed model is relative simple in concept and application
and is suitable for use by planners. All of the data required by
the model can be collected on site or can be estimated from the
literature. Labratory analysis of soil and bank samples,,however,
will greatly improve model reliability with respect to nutrient
losses. In areas with shoreline erosion, the procedure also allows
the benefits of shoreline control to be considered.

DEVELOPMENT OF A BUFFER ZONE EVALUATION
MODEL/PROCEDURE

Theo A. Dillaha
Associate Professor
Department of Agricultural Engineering
Virginia Polytechnic Institute and State University
Blacksburg, VA 24061-0303

INTRODUCTION

The original purpose of this project was to develop a qualitative
technique for evaluating vegetative buffer effectiveness with
respect to sediment and nutrient removal. The technique was
intended for use by planners in evaluating the relative
effectiveness of various buffer zone modification schemes. After
f ield testing of the proposed procedure, it was agreed that the
project objectives would be expanded to account for the effects of
vegetative buffer modification due to shoreline stabilization
practices. This was necessary because installation of shoreline
stabilization systems may reduce vegetative buffer effectiveness by
reducing vegetative buffer length, increasing vegetative buffer
slope, and disturbing buffer vegetation. However, the benefits of
shoreline stabilization for reduced sediment and nutrient losses
due to control of shoreline erosion can more than compensate for
reduced vegetative buffer effectiveness in most cases.

The resulting assessment process allows impact predictions to be
made for site specific conditions of individual vegetative buffer
alterations as well as relative effectiveness comparisons between
vegetative buffers. The vegetative buffer zone evaluation
procedure is designed to use soil survey data and other parameters
including slope and surface roughness. The site evaluation methods
will consider both surface runoff and subsurface flow within the
buffer.

Specific objectives of the project were to:

1. Select a vegetative buffer assessment methodology best suited
for soils and shoreline conditions in Virginia.

2. Modify the method as required to improve the method's
suitability for Virginia conditions.

3. Determine the availability of physical parameters necessary to
apply the selected buffer zone evaluation methodology. Define
ranges of parameters suitable for application in Virginia's
coastal zone.

4. Test the proposed methodology on specific sites in Virginia.

5. Prepare a f inal report and guide detailing the development,
use and limitations of the proposed methodology.

LITERATURE REVIEW

Vegetative buffers (also referred to as vegetative filter strips,
grass filter strips, buffer strips, vegetative buffers, riparian
buffer zones, filter strips, etc.) are bands of planted or
indigenous vegetation, situated between pollutant source areas and
receiving waters. They are presumed to remove sediment and
chemicals f rom runof f and ground water interacting with the buf f er.
Pollutant removal in vegetative buffers is accomplished by a
variety of physical, chemical, and biological processes. These
processes are poorly understood and there is considerable
uncertainty as to the effectiveness of vegetative buffers in
removing pollutants from surface runoff and ground water.
Currently, there are no standards or widely accepted methods for
evaluating vegetative buffer effectiveness. Consequently, it is
difficult if not impossible to determine how effective vegetative
buffers are in protecting water quality.

Numerous short-term studies have found that vegetative buffers are
initially very effective in removing sediment and sediment-bound
pollutants f rom. surf ace runof f under shallow sheet f low conditions.
The long-term (more than one-year) effectiveness of vegetative
buffer for pollutant removal, however, has not been investigated
very extensively. Riparian buffer zone design procedures, proposed
over the past 10-years, have been research oriented and based on
short-term experimental studies. These studies did not consider
the long-term effects of sediment and nutrient accumulation in
vegetative buffers. These studies and design methods also
generally ignore the effects of concentrated flow conditions on
vegetative buffer performance. This is unfortunate, as most flow
in the real world will enter vegetative buffers as concentrated
flow rather than the shallow sheet flow used in model development
(Dillaha et al., 1989). Consequently, those equations that do
exist, generally overestimate vegetative buffer effectiveness with
respect to sediment and nutrient removal because they do not
consider the effects of concentrated flow and the accumulation of
sediment and nutrients in vegetative buffers over time.

The major pollutant removal mechanisms associated with vegetative
buffers involve changes in flow hydraulics that enhance the
opportunity for the infiltration of runoff and pollutants into the
soil profile, deposition of total suspended solids (TSS),
filtration of suspended sediment by vegetation, adsorption on soil
and plant surfaces, and absorption of soluble pollutants by plants.
For these mechanisms to be effective, it is essential that runoff
pass slowly through the vegetative buffer to provide sufficient
contact time for the removal mechanisms to function.

Infiltration is one of the most significant removal mechanisms
affecting vegetative buffer, performance. Infiltration is important
since many pollutants associated with surface runoff enter the soil
profile in the buffer area with infiltrating water. Once in the

soil profile, many pollutants, particularly N and P, are removed by
a combination of physical, chemical, and biological processes.
Infiltration is also important because it decreases the amount of
.surface runoff, thus reducing the ability of runoff to transport
pollutants. Since infiltration is one of the more easily
quantifiable mechanisms affecting buffer performance, many
vegetative buffers have been designed to allow all runoff from
design storms to infiltrate into the buffer (Midwest Plan Service,
1985). This approach results in large land requirements because it
ignores other removal mechanisms.

vegetative buffers also purify runoff through the process of
deposition. Because vegetative buffers usually offer high
resistance to shallow overland flow, they decrease the velocity of
overland flow immediately upslope and within the buffer, causing
significant reductions in sediment transport capacity. If the
transport capacity is less than the incoming suspended solids load,
then the excess suspended solids may be deposited and trapped
within the buffer. Sediment-bound pollutants will also be removed
during the deposition process.

The filtration of solid particles by vegetation during overland
flow and the absorption and adsorption processes are not as well
understood as the infiltration and deposition processes.
Filtration is probably most significant for the larger soil
particles, aggregates, and organic particles while adsorption is
thought to be a significant factor with respect to the removal of
dissolved pollutants. The major questions concerning adsorption
and absorption involve their long-term effectiveness as nutrients
accumulate in the buffer (Lee et al., 1989; Dillaha et al., 1989).

Sediment Transport in Vegetative Buffers

Historically, the design of vegetative buffers has been based
almost entirely upon local custom. Wilson (1967) presented the
results of a sediment trapping study in grass buffers which gave
optimum distances required to trap sand, silt, and clay in flood
waters on flat slopes. He concluded that grass buffer length,
sediment load, flow rate, slope, grass height and density, and
degree of vegetative submergence all affect sediment removal.
Neibling and Alberts (1979) used a rainfall simulator on
experimental field plots with a slope of 7% to show that 0.6, 1.2,
2.4, and 4.9 m long grass buffers all reduced total sediment
discharge by over 90% from a 6.1 m long bare soil area. Discharge
rates for the clay size fraction were reduced by 37, 78, 82, and
83%, for the 0.6, 1.2, 2.4, and 4.9 m grass buffers, respectively.
Significant deposition of solids was observed to occur just upslope
of the leading edge of the grass buffer and 91% of the incoming
sediment load was removed within the first 0.6 m of the grass
buffer. Sediment discharge.of clay sized particles (<0.002 mm) was
reduced 37% by the 0.6 m strip. No equations were presented to
estimate the influence of parameters on sediment yield.

The most comprehensive research to date on sediment transport in
vegetative buffers has been conducted by a group of researchers at
the University of Kentucky (Barfield et al., 1977; 1979; Kao and
Barfield, 1978; Tollner et al., 1976; 1978; 1982; Hayes et al.,
1979a,b; 1982). Tollner et al. (1976) presented design equations
derived from experimental studies relating the fraction of sediment
trapped in simulated vegetative media to the mean flow velocity,
flow depth, particle fall velocity, filter length, and the spacing
hydraulic radius (a parameter similar to the hydraulic radius in
open channel flow that is used to account for the effect of media
spacing on flow hydraulics). The Kentucky researchers reported
high trapping efficiencies as long as the vegetative media was not
submerged, but trapping efficiency decreased dramatically at higher
runoff rates that inundated the media. The Kentucky researchers,
like Neibling and Alberts (1979), observed that much of the
sediment deposited just upslope of the filter and within the first
meter of the filter, until the upper portions of the filter were
buried in sediment. Subsequent flow of sediment into the filter
resulted in the advance of a wedge-shaped deposit of sediment down
through the filter. The Kentucky resea 'rchers did not consider
nutrient trapping or the long-term effectiveness of vegetative
Ilk buffers.

Hayes and Hairston (1983) used field data to evaluate the Kentucky
model for multiple storm events in Mississippi. Eroded material
from fallow cropland subject to natural rainfall was used as a
sediment source. 'Kentucky 311 (Festuca arundinacea) tall fescue
0 trimmed to 10 cm was used and the model predictions agreed well
with the measured sediment discharge values.

Kao et al. (1975) proposed a vegetative buffer arrangement in which
grass strips were alternated with strips of bare ground to solve,
the problems associated with sediment inundation of the filter and
0 the killing of vegetation. Kao indicated that with the proper
vegetative buffer area to source area ratio, most of the trapped
sediment would be retained in the bare area just upslope of the
buffer. The trapping of sediment upslope of the buffer maintained
high buffer efficiencies and enabled periodic removal of deposited
sediment without damaging the buffer. Kao's results were based
0 upon laboratory studies with artificial media and were not tested
in the field.

Nutrient Transport in Vegetative Buffers

Nutrient movement through vegetative buffers has been investigated.
0 by several researchers but no comprehensive design methods have
been presented. Doyle et al. (1977) applied dairy manure to 7 x 5
m fescue plots on a Chester silt loam (fine-loamy, mixed, thermic,
Typic Hapludult) soil with a slope of 10%. Dissolved nutrient
concentrations were measured after passing through 0.5, 1.5, and
4.0 m of fescue buffer strips. Dissolved P was reduced by 9, 8, and
0 62% after passage through 0.5, 1.5, and 4.0 m buffers,

respectively. Nitrate (NOA losses decreased by 0, 57, and 68%,
respectively, but ammonia (NH3) concentrations increased with
increasing filter length presumably due to the release of NH3 from
decomposing organic N, which was trapped in the filter previously.

Young et al. (1980) used a rainfall simulator to study the ability
of vegetative buffers to control pollution from feedlot runoff.
Field plots were constructed on a 4% slope with the upper 13.7 m in
an active feedlot and the lower 27.4 m planted in either corn (Zea
mays), oats (Auena sativa), orchardgrass, (Dactylis glomerata) or
a sorghum-sudangrass (Sorghum vulgare-Sorghum sudanensis) mixture.
Water was applied to the plots to simulate a 25-year, 24-hour
duration storm. Total runoff, sediment, P and N were reduced by
81, 66, 88, and 87%, respectively, by the orchardgrass and by 61,
82, 81, and 84%, respectively, with the sorghum- sudangrass
mixture. The authors concluded that vegetative buffers were a
promising treatment alternative.

Thompson et al. (1978) studied the effectiveness of orchardgrass
buffers on a sandy loam soil in reducing nutrient loss from the
application of dairy manure to frozen or snow-covered orchardgrass
plots. Fresh dairy manure was applied to 24 m orchardgrass plots
and runoff quality determined after traveling through 12 and 30 m
of additional orchardgrass during natural runoff events. Total P,
total Kjeldahl nitrogen (TKN), and N losses were reduced by an
average of 55, 46, 41, and 45%, respectively, after passing through
12 m of filter. A 36 m filter resulted in P, N03, TKN, and N
reductions of 61, 62, 57, and 69%, respectively. Nutrient
concentrations in the runoff from the 36 m filters approached that
from control plots to which no manure had been added. Bingham et
al. (1978) applied poultry manure to 13 m long fescue grass plots
on an eroded Cecil clay loam (clayey, kaolinitic, thermic Typic
Hapludult) with 6-8% slopes. Buffer length/waste area length
ratios of about 1.0 were reported to reduce pollutant loads to near
background concentrations. Total P, TKN, N03, and total-N were
reduced 25, 6, 28, and 28%, respectively.

Edwards et al. (1983) monitored storm runoff for 3 years from a
paved feedlot. Storm runoff was measured and sampled as it left
the feedlot, after passing through a shallow concrete settling
basin, and after passing through two consecutive 30.5 m, long fescue
buffers. Runoff, TSS, P, and N were reduced by -2, 50, 49, and 48%,
respectively, after passing through the first buffer and by an
additional -6, 45, 52, and 49%, respectively, after passing through
the second buffer. Total runoff from the buffers was greater than
the incoming runoff because rainfall rates during runoff events
exceeded the infiltration capacity of the buffers. This rainfall
excess coupled with the added area of the buffers resulted in
increased runoff. Removal efficiencies would have been higher if
the settling basin located. upslope of the buffer had not removed
54, 41,, and 35% of the TSS, P, and N, respectively. Most of these
solids and nutrients would have been removed in the buffers because

they were either settleable solids or nutrients bound to settleable
solids.

Patterson et al. (1977) applied liquid dairy waste through a gated
pipe to a fescue buffer on Hosmer silt loam (fine-silty, mixed,
mesic Fragiudalf) on a 3.4% slope. After applying dairy waste to
the 35 m vegetative buffer for one year, pollutant reductions
averaged 42, 38, 7, and 71% f or BOD., NH3. P04, and TSS,
respectively. Nitrate loss from the filter was greater than NO,
loading to the buffer, presumably due to mineralization of organic
N and nitrification of NH3 that had been trapped in the buffer
previously. Paterson et al. (1977) also noted problems with
maintaining a good grass cover on the buffer area. They
recommended that several buffer areas should be use and rotated on
a weekly basis to maintain good grass cover.

Magette et al. (1989) used a rainfall simulator on field plots to
study the effectiveness of 4.6 and 9.2 m long grass buffers in
removing nutrients and sediment from agricultural runoff. Nutrient
removal appeared to decrease as the number of runoff events
increased. Gross sediment, N, and P losses in surface runoff were
reduced by 52, -15, and 6%, respectively, by the 4.6 m buffers, and
75, 35, and 20%, respectively, by the 9.2 m buffers. The buffers
were reported to be much more effective in removing sediment than
nutrients. Buffer effectiveness was also reported to decrease with
time and with decreasing buffer to source area ratio.

Vegetative Buffer Research in Virginia

Researchers at Virginia Tech (Dillaha et al., 1986; 1987; 1988;
1989; and Lee et al., 1989) used a rainfall simulator to evaluate
the effectiveness of grass buffers for the removal of sediment, N,
and P from cropland runoff. Simulated rainfall was applied to
nine experimental field plots on an eroded Groseclose silt loam
soil (clayey, mixed, mesic Typic Hapludalt) with a 5.5 by 18.3 m
bare cropland source area and either a 0, 4.6, or 9.1 m long grass
buffer (5.5 m wide) located at the lower end of each plot.
Fertilizer was applied to the,plots at rates of 222 kg/ha of liquid
N and 112 kg/ha of P,O, and K2*0. Water samples were collected from
the base of each plot and analyzed for sediment and nutrient
content. One set of plots was constructed so that flow through the
filters was concentrated rather than shallow and uniform. The 9.1
and 4.6 m grass buffers with shallow uniform flow removed 87 and
75% of the incoming suspended solids, 69 and 57% of the incoming P,
and 72 and 61% of the incoming N, respectively. Dissolved
nutrients in the buffer effluent were sometimes greater than the
incoming dissolved nutrient load, presumably due to lower removal
efficiencies for dissolved nutrients and the release of nutrients
previously trapped in the buffers. Plots with concentrated flow
were much less effective than the shallow uniform flow plots, with
percentage reductions in sediment and nutrient loadings averaging
23 to 37% less for sediment, 46 to 53% less for N, and 43 to 46%

less for P.

The effectiveness of existing vegetative buffers in the
Commonwealth of Virginia was qualitatively evaluated by visiting
and observing vegetative buffers on 18 farms in Virginia (Dillaha
et al., 1986). Buffers were evaluated by talking with landowners
and soil conservationists and walking the length of the buffers to
evaluate potential problems. All the vegetative buffers surveyed
were composed of grasses and other low growing vegetation and were
used in combination with cropland. Almost all the buffers were
installed for water quality improvement in conjunction with
Virginia's Chesapeake Bay Program. Buffers were rarely used before
1983 on cropland in Virginia because they were not a recognized
conservation practice eligible for state or federal cost-sharing
money. Buffer performance was generally judged to fall into two
categories depending upon the topography of the site. In hilly
areas, buffers were judged to be ineffective for removing sediment
and nutrients from surface runoff because drainage usually
concentrated in natural drainageways within the fields before
reaching the buffers. Flow across the buffers during larger runoff
producing storms (the most significant in terms of water quality)
was therefore primarily concentrated and the buffers were locally
inundated and ineffective. This assessment was confirmed by the
fact that little sediment was observed to have accumulated in the
majority of the buffers observed. Buffers in these areas, while
not effective for trapping sediment and nutrients, were judged to
be beneficial because they provided effective cover in areas
immediately adjacent to streams that are often susceptible to
severe localized channel and gully erosion. They also provide a
narrow buffer between cropland and streams that may reduce the
aerial drift of fertilizers and pesticides to streams during
application.

In flatter areas, such as in the Coastal Plain, buffers appeared to
be more effective. Slopes were more uniform, and significant
portions of stormwater runoff entered the buffers as shallow
uniform flow. This observation was supported by the presence of
significant sediment accumulations in many of the Coastal Plain
buffers surveyed. Several one to three year old buffers were
observed that had trapped so much sediment that they were higher
than the fields they were protecting. In these cases, runoff
tended to flow parallel to the buffers until a low point was
reached where it flowed across as concentrated flow. In this
situation, the buffers acted more like a terrace than a vegetative
buffer. Flow parallel to the buffers also was observed on several
farms where moldboard plowing was practiced. When soil was
turn-plowed away from the buffers, a shallow ditch was formed
parallel to the field. If this ditch was not removed later by
careful disking, runoff again concentrated and flowed parallel to
the buffer until it reached a low point and crossed as channel
flow. Conclusions drawn from the plot studies and on-site
assessments of vegetative buffer effectiveness included (Dillaha et

al. 1989)

1. Vegetative buffers are effective for the removal of sediment
and other suspended solids from cropland runoff only if flow
is shallow and uniform and if the buffers have not been
previously inundated with sediment.
2. The effectiveness of vegetative buffers for sediment removal
appears to decrease with time as sediment accumulates within
buffers. This may or may not be a problem in "real world"
buffers because vegetation may be able to grow through
sediment accumulations.
3. Total N and P in runoff are not removed by vegetative buffers
as effectively as sediment. Presumably, much of the N and P
in cropland runoff is dissolved or associated with very fine
sediment which vegetative buffers can not remove efficiently.
4. Shorter vegetative buffers (<10 m) are not effective in
removing dissolved N and P from agricultural runoff.
Dissolved P and N losses from the experimental vegetative
buffer plots studied were often higher than the inflow,
presumably due to the release of P and N trapped in the grass
buffers previously.
IN 5. Buffer strips characterized by concentrated or deeper channel
type flow are much less effective for sediment, N, and P
removal than vegetative buffers with shallow uniform flow.
Buffers with concentrated flow were 40 to 60%, 70 to 95%, and-
61 to 70% less effective with respect to sediment, P, and N
removal than uniform flow plots.
6. Most on-farm vegetative buffers observed during the Virginia
Tech study were judged to be ineffective for sediment and
nutrient removal. The majority,of flow entering the grass
portion of the buffers was judged to be concentrated because
runoff tended to accumulate in natural drainageways long'
before reaching the vegetative buffers. This was more of a
problem in hilly areas and less of a problem in flatter areas
such as the Coastal Plain.

The Virginia Tech researchers concluded that the effectiveness of
the experimental vegetative buffers should not be used as a direct
indicator of real world vegetative buffer effectiveness because of
the concentrated flow problems previously discussed. Concentrated
flow effects under real agricultural conditions were estimated to
be orders of magnitude greater than those encountered during the"
experimental field studies (Dillaha et al., 1989).

Vegetative Buffer Models and Design Procedures

Kentucky Filter Strip Model: Barfield et al. (1979) developed a
steady state model, the Kentucky filter strip model, for
determining the sediment filtration -capacity of grass media as' a
function of flow, sediment, load, particle size, flow duration,
slope, and media density. Outflow concentrations were primarily a
function of slope and media spacing for a given flow condition.

The Kentucky filter strip model was extended for unsteady flow and
non-homogeneous sediment by Hayes et al. (1979a) . A graphical
solution of the Kentucky model was described by Hayes et al.
(1982). However, the complexity of the procedure makes solution of
the equations difficult unless the sediment is well graded.
Methods for determining the values of the hydraulic parameters
required by the Kentucky model for real grasses were presented.
Using three different types of grasses, model predictions were
reported to be in close agreement with laboratory data (Hayes et
al., 1978).

A simplified procedure derived from the Kentucky filter strip model
was developed for the SCS to estimate the long-term effectiveness
of grass filter strips (Hayes and Dillaha, 1992; Dillaha and Hayes,
1992). The procedure estimates the trapping efficiency of grass
buffers with respect to sediment but does not consider other types
of contaminants or buffer vegetation. The model is fairly simple
to use but requires the use of the WEPP model (Lane and Nearing,
1989), which is not yet available to the public, to estimate
sediment and surface runoff loadings to the buffer.

CREAMS Model: Agricultural Research Service researchers (Flanagan
et al., 1986; Williams and Nicks, 1988) have attempted to evaluate
the effectiveness of vegetative buffers for erosion control using
the CREAMS model (Knisel, 1980). Williams and Nicks (1988) applied
CREAMS to a 1.6 ha watershed in Oklahoma. Filter strip
effectiveness was found to be dependent on strip length, Manning's
n, slope, and slope shape. The authors concluded that CREAMS can
be-a useful tool for evaluating vegetative buffer effectiveness in
reducing sediment yield. This model, like others mentioned
previously, cannot consider the long-term effectiveness of
vegetative buffers because it has no way of accounting for sediment
accumulations within the vegetative buffer. Consequently, CREAMS
would be expected to overestimate long-term sediment trapping.
CREAMS also is severely limited by its sediment transport model
that tends to overestimate sediment transport. The model also
cannot account for concentrated flow effects, and Manning's n is
the only factor used to simulate the effects of vegetative buffer
vegetation. CREAMS does have nutrient transport submodels, but
their use with vegetative buffers has not been reported. In
summary, CREAMS was not developed to describe vegetative buffers
and use of the model for vegetative buffer design is highly
questionable since it does not simulate the principal physical
processes affecting transport in vegetative buffers.

GRAPH Model: Lee et al. (1989) developed an event-based model,
GRAPH (GRAss PHosphorus), to simulate P transport in vegetative
buffers by incorporating chemical transport submodels into the
grass filter strip model in SEDIMOT II (Wilson et al., .1984; Warner
et al., 1984), a stormwater.and sediment transport model developed
for strip mine reclamation. The grass filter model in SEDIMOT II
was derived from the Kentucky filter strip model, GRASSF, developed

by Hayes (1979). GRAPH considers the effects of advection
processes, infiltration, biological uptake, P desorption from the
land surface to runoff, adsorption of dissolved P to suspended
solids in runoff, and the effects of changes in sediment size
distribution on P transport. Required data for the model includes:
rainfall intensity and duration, an inflow hydrograph, a sediment
graph, sediment size distribution, vegetative buffer dimensions and
hydraulic characteristics, inflow graphs for dissolved P, P
desorption and adsorption reaction coefficients for soil and plant
matter, and the P content of each soil particle size class. GRAPH
simulates time varying infiltration, surface runoff, sediment
yield, particle size distribution, and dissolved and sediment-bound
P discharge along with P and sediment trapping efficiencies in
vegetative buffers. GRAPH was verified with data from vegetative
buffer field plots. Model predictions and observed P transport 'in
grass buffers compared favorably.

Phillips Model: Phillips (1989a,b) presented a theoretical method*
for evaluating the relative effectiveness of buffer zones in
removing sediment, sediment adsorbed chemicals, and dissolved
chemicals from surface and subsurface flow. The method does not
make absolute predictions of buffer effectiveness but rather
estimates the effectiveness of a given buffer relative to a
reference buffer. Philips's method predicts the relative
effectiveness of buffers for water quality improvement using two
models, the hydraulic model and the detention model. Neither' of
the Phillips models have been validated or tested with field data
because they attempt to characterize the long-term effectiveness of
buffers and no long-term data on buffers has been collected with
which to verify these or any other long-term models.

Phillips Hydraulic Model: The Phillips hydraulic model was
developed to describe the transport of sediment and sediment-bound
chemicals through vegetative buffer zones. The model assumes that
sediment transport through the filter is a function of the energy
of overland flow and is based on the Bagnold stream power equation.
The detention model equation as originally proposed by Phillips is:
0.4( Sb)-I.3(.1
Bb =( KII)(_@b) b)0.6
BZ K, Lz SZ n.
where: K = saturated hydraulic conductivity of the buffer soils
L = buffer length
s = sin 6, where 6 is the slope angle relative to the
horizontal
n = Manning roughness coefficient
b = subscript denoting the buffer of interest,-and
r = subscript denoting the reference buffer

The hydraulic model is derived as follows. For a given mass of
water, Bagnold's stream power equation (Bagnold, 1977) can be used
to estimate the sediment transport capacity. That is, the time rate

of energy expendature per unit weight of flowing water is:
PgAL VS -VS (2]
pgALr
where: p = density of water
g = gravitational constant
A = cross-sectional area of flow
Lr = length of the flow reach
V = mean flow velocity
s = slope of the hydraulic grade line (-sin 0)

For steady state flow conditions, the flow rate per unit width can
be expressed as:
q= VA = Vy (3]

where y is the steady-state f low depth. Equation [ 3 ] can be
rearranged to:
V= q [4]
Y

The average flow velocity can be expressed with Manning's equation
as:
V= -1 R2/3S1/2 [)
n

where R is the hydraulic radius of the flow. For shallow sheet
flow conditions, R = y. Multiplying both sides of Equation (5] by
the area per unit width gives the flow rate per unit width:
q= VA= 1 y2/3 S 1/2A [6]
n

For steady, shallow sheet flow conditions, A = y and Equation [6]
can be rearranged to:
nq 3/5 [7]
S1/2
Equations [2], [4] and [7] can then be combined to:

Pu= 0.4S1. 3n -0. 6 [8]
Phillips incorporated buffer length, L, into Equation (8] by
assuming q = Li, where i is the steady-state excess rainfall rate.
Equation (8] can then be expressed as:

PU=LO 0.4I0.4s1.3 n-0.6 [9]
Then, denoting the saturated hydraulic conductivity, K, as an
indicator of the infiltration capacity of the soil, a general index
of buffer effectiveness relative to a reference buffer is obtained:

Bb = Kb(Lb) 0.4( ib)o.4( -1.3(0.6 (10]

Br Kr Lr ir Sr nr.
Assuming that rainfall intensity is the same on both the buffer of
interest and the reference buffer, ib= ir, Equation (10 reduces to
Phillips, hydraulic model (Equation [1]).

Phillips Detention Model: The Phillips' detention model estimates
the relative effectiveness of buffers in removing dissolved
substances from surface and subsurface flow through vegetative
buffer. Contaminate removal in the buffer is defined as a function
of the total contact time of both surface and subsurface flow in
the buffer. The model is derived from Darcy's law and the Manning
equation.

where M is the soil moisture storage capacity and the other terms
are defined as before. The soil moisture storage capacity is
defined as the available soil moisture content (soil moisture
content at field capacity minus soil moisture content at wilting
point) times the lessor of the seasonable high water table depth or
the depth to a confining soil layer.

Phillips derived the detention model as follows. The total
contact time due to surface runoff through the buffer can be
expressed as:

T= L [12]
V

where T, is the detention time due to surface runoff in the filter.
Combining Equations [2], [8], and [12] gives:

Ts=ns-0.3q O.4 L (13]
where q. is the surface discharge.

For subsurface flow, Phillips used Darcys law to estimate the
velocity of subsurface flow:
V=Ks [14]

where K = saturated hydraulic conductivity. Detention time due to
subsurface flow,, Tq., was then estimated as:
Tg=KsL [15]
Considering both surface and subsurface throughflow, Phillips
defined an index of detention, T*, for a given flow of:

T*=T,*T,=[nO-6Ls-0.3 (qlq,,) -1-1]*[.KsL(qq1q)J [16]
where q. is the subsurface discharge component of flow. Phillips
assumed that q, q.,, and % were a function of K. Phillips further
assumed that "the portion of discharge which travels overland or in
subsurface flow is a fucntion of the infiltration capacity, which
is assumed to be a fucntion of K. For the overland flow component,
detention time varies as the -0. 4 power of q.. For a given
stormwater mass, since K is an index or surrogate of q,, T=f (K-0*4).
Since the portion of the discharge traveling on the surface is an
inverse function of K, the sign is reversed and T=f(K")." The
relative abilities of buffers to hold infiltrated water was given
by MIM, where M is the soil moisture storage capacity obtained by
multiplying the available soil moisture capacity (field capacity
minus wilting point) by the depth to the water table or a confining
soil layer. Equation (16] can then be expresses as:
b b)2(L -0.7( M
T* .6 b)0.4( Sb) b)
[17]
T*.r nr Lr K, S, M.-
which is functionally identical to Equation (11].

other Design Approaches: Procedures for the design of vegetative
buffers with respect to organics removal have been presented by
Norman et al. (1978) and Young et al. (1982). However, these
procedures were based primarily on infiltration or limited organics
removal data. Regression type design equations for P reduction
were presented by Young et al. (1982), but details of their
development were not presented and they have not been verified.

State and Federal Vegetative Buffer Programs

Chesapeake Bay Preservation Act: Regulations have been proposed in
Virginia as part of the Chesapeake Bay Program to require
vegetative buffers along all water bodies in designated Resource
Protection Areas (Chesapeake Bay Preservation Area Designation and
Management Regulations, Chesapeake Bay Local Assistance Board).
The Resource Protection Areas are defined as "sensitive lands at or
near the shoreline that have intrinsic water quality value ... and
are sensitive to impacts which may cause significant degradation to
the quality of state waters or loss of aquatic habitat." This
definition includes all tidal and nontidal wetlands, tidal
shorelines, and all tributary streams within Virginia's Chesapeake
Bay drainage basin.

Along all tidal waters, a 100 ft (30.5 m) vegetative buffer zone is
required, and a 50 ft (15.2 m) buffer is required along nontidal
waters. The buffer length is measured from the mean high water
level of nonvegetated wetlands and from the wetland for vegetated
wetlands. If agricultural lands are adjacent to waters, then "the
buffer zone area shall maintain as a minimum best management

practice, a 25 ft (7.6 m) wide vegetative buffer measured landward
from the mean high water level of tidal waters or tributary
streams, or from the landward edge of any wetlands." Buffers will
not be required for agricultural drainage ditches if the adjacent
land has best management practices in place in accordance with a
conservation plan approved by the local Soil and Water Conservation
District. The regulation specifies that the vegetative buffer
shall be composed of either trees with a dense ground cover, grass,
or an approved legume cover which can be managed to prevent
concentrated flows from breaching the vegetative buffer. The
vegetative buffer must be maintained until the landowner has
implemented an approved BMP program which provides water quality
protection at least the equivalent of that provided by the
vegetative buffer. The regulation specifies that for the purposes
of the Act, 100 ft buffers remove 75 and 40% of the incoming
sediment and nutrients, respectively.

The proposed regulations are a step in the right direction, but it
is highly unlikely that they will result in 75% sediment and 40%
nutrient reductions. In many areas, little if any runoff will
flow across the buffers and in other areas most runoff will cross
the vegetative buffer as concentrated flow. This will be a major
problem with the proposed regulations because vegetative buffers
will only be required along perennial streams depicted on USGS
topographic quadrangle maps. Consequently, most surface runoff
will collect in ephemeral drainageways before reaching. the
vegetative buffers and cross as concentrated flow. The state of
Maryland has similar buffer zone requirements along shorelines and
major tributaries but does not require vegetative buffers
explicitly for removing pollutants from surface runoff.

Conservation Reserve Program: The use of constructed vegetative
buffers in the United States has increased significantly in the
past few years, because vegetative buffers were an approved USDA
cost-share practice under the Conservation Reserve Program (CRP) of
the Food Security Act of 1995. The CRP was established to
encourage farmers to take highly erodible land out of crop
production and convert the land to permanent (10 year) cover. This
program was designed to reduce soil erosion, improve water quality
and wildlife habitat, as well as eliminate production of excess
commodities. Farmers participating in the CRP receive an annual
rental payment for land enrolled in the program. As originally
implemented, only land classified as highly erodible was eligible
for participation in the CRP. In 1988, the CRP was modified to
include vegetative buffers because of their potential environmental
benefits. The requirement that the land be highly erodible was
eliminated for vegetative buffers. Requirements for vegetative
buffers under the CRP include:

1. The land must be adjacent and parallel to a stream, river,
lake, estuary, or wetland greater than 2 ha (5 acres) in area.
2. The land must have been planted in an agricultural commodity

in at least two years from 1981 through 1985.
3. The land must still be suitable for crop production.
4. The land with a vegetative buffer must be capable of reducing
sediment delivery to adjacent water bodies.
5. The land must be planted to permanent grasses, trees, or
shrubs.
6. The vegetative buffer must be a minimum of 20 m (66 ft) in
length and no more than 30 m (99 ft) in length.
7. The vegetative buffer may not be grazed or harvested during
the 10 years of the contract.

National Vegetative Filter Strip Conservation Practice Standard:
The U.S. Soil Conservation Service is currently in the process of
updating the national conservation practice standard for vegetative
buffers to overcome some.of the limitations of vegetative buffers.
The proposed standards define vegetative buffers as vegetated areas
which are designed to remove sediment, nutrients, pathogens,
organic materials, pesticides, and other contaminants from surface
runoff by filtration, deposition, infiltration, adsorption,
adsorption, decomposition, and volatilization. The key word here
is "designed". This implies that vegetative buffers are not
suitable for every site and that their length and position will be
a function of local site conditions and hydrology.

An important part of the proposed standard is the statement: "The
practice (vegetative buffer) applies ... in locations above the
occurrence of concentrated flow and above conservation practices
such as terraces or diversions which concentrate flow." The new
standard relaxes previous requirements that vegetative buffers be
located immediately adjacent to streams and instead says that they
should be located where they will be the most effective for
pollutant removal. This may be at the lower boundary of a field,
or it may be within a.field.

The proposed standards suggest that the design of vegetative
buffers and the suitability of a particular site for vegetative
buffers must consider (Dillaha, 1989):

1. Adequacy of soil drainage and depth to water table to ensure
satisfactory vegetative growth and prevent prolonged
saturation of the soil.
2. Provisions for preventing hillside seeps and other continuous
discharge of water through the vegetative buffers.
3. Reduced effectiveness of vegetative buffers under snow or
frozen ground conditions.
4. Vegetative buffer length required to provide the desired
pollutant reduction over the design life of the vegetative
buffer. In other words, pollutant accumulation and subsequent
release from vegetative buffers must be a design constraint.
5. The effects of slope onvegetative buffer effectiveness.
6. Provisions for mowing, to maintain the effectiveness of
vegetative buffers composed of grass and similar vegetation.

7. Effects of grazing on.vegetative buffer.performance.
8. Effects of application of herbicides to vegetative buffers or
adjacent fields for weed control. , If herbicides are applied
to fields, sprayers should be turned off before crossing
vegetative buffers.or using them for turn rows.
9. Vegetative buffers should be installed on the contour as much
as possible to filter runoff before it concentrates in natural
drainageways.
10. Care should be taken during tillage operations to avoid
tilling into vegetative buffers and causing localized flow
problems.
11. Large fields with significant natural drainageways or grassed
waterways are acceptable for vegetative buffers only if the
vegetative buffers are installed on both sides of internal
field drainageways. This will allow pollutants to be trapped
before they can enter the drainageways.
12. Some sites may require limited grading to correct flow
problems within the vegetative buffers caused by gullies or
high areas within or immediately downslope of the vegetative
buffers.
13. Shrub and wildlife strips should not be permitted because they
are relatively ineffective for water quality improvement when
compared to grass and legume vegetative buffers.
14. At sites with significant flow along or parallel to vegetative
buffers, shallow berms or diversions may be needed at 15 to 30
m intervals to intercept runoff and force it to flow through
the vegetative buffer before it can concentrate further.
15. Vegetative buffers should not be installed in areas higher
than the fields they are intended to protect.

Sediment and Nutrient Loadings Due to Eroding Banks

A major difficulty with the Chesapeake Bay Preservation Act as
currently implimented is that the Act makes it difficlt to modify
buffer zones even if the modifications would reduce sediment and
nutrient loadings to the Bay and other water bodies. Consider the
case of eroding banks. Ibison et al. (1990) examined the loss of
sediments and nutrients from eroding tidal shorelines along the
Chesapeake Bay and its tributaries. Eroding banks were reported to
be responsible for 5.2 and 23.6% of the controllable N and P,
respectively, entering Virginia's portion of the Chesapeake Bay.
In a follow up study, Ibison et al. (1992) confirmed the results of
the previous study and reported that the sheer mass of materials
lost through shoreline erosion results in-nutrient loading rates
(on an areal basis) to , the Chesapeake Bay several orders of
magnitude higher than upland loading rates. For example, N and P
losses from shoreline erosion were estimated to be approximately
25,000 kg-N/ha-yr and 15,000 kg-P/ha-yr versus losses of 2 to 80
kg-N/ha-yr and 0.3 to 19 kg-P/ha-yr for cultivated farm land.
Consequently, stabilizing one hectare of eroding bank may reduce
nutrient loadings to the Bay as much as stopping all nutrient loss

from 300 to 800 ha of cultivated farmland.

Many of the actively eroding banks along the Chesapeake Bay and its
tidal tributaries are characterized by high steep banks. The banks
erode at an average rate of 0.2 m/yr (Byrne and Anderson, 1977)
with reported rates as high as 3.3 m/yr (Ibison et al., 1992). To
stabilize these banks, disturbance of the existing riparian zone is
often required for the construction of shoreline structures or to
grade the bank back to a stable slope of 2 to 1 (run to rise) or
50%. This necessitates removal of all vegetation during grading
and then replanting after grading is complete. It may also be
necessary to remove large trees on and in the vicinity of the bank
to prevent mass slumping and loss of soil during large storms that
cause trees to fall down the banks, bringing tons of soil with
them. The Shoreline Programs Bureau recommends that all large
trees be removed from steep slopes and that large trees also be
removed from the zone within two bank heights distance from the top
of steep banks. Removal of trees under these circumstances is
estimated to reduce average annual soil loss from banks by
approximately 10% (Hill, 1992).

Literature Review Sununary

As discussed previously, vegetative buffers as presently
implemented are unlikely to be very effective in removing sediment
and nutrients from surface runoff because they are usually
installed with little consideration of site conditions which affect
their performance. Equations which have been developed for
vegetative buffer design assume that runoff is uniformly
distributed across the width of vegetative buffer as shallow sheet
flow. This will rarely be the case in real world situations as
flows in all but the most uniformly sloping fields tend to
concentrate in internal field drainageways before reaching field
boundaries where vegetative buffers are usually located.
Consequently, significant portions of field runoff will cross the
vegetative buffers as concentrated flow, locally inundating the
vegetative buffers, and greatly reducing vegetative buffer
effectiveness for sediment and nutrient removal. In addition,
almost all vegetative buffer research reported has been of a
short-term nature which did not consider the long-term
effectiveness of vegetative buffers for pollutant reduction. The
design equations and models developed from these studies do not
consider the effects of sediment and nutrient accumulation in
vegetative buffers. Consequently, they will probably over predict
vegetative buffer effectiveness over the long run.

Of all the models reviewed, only the Phillips method was developed
for vegetation other than grasses. The Phillips method, because of
its theoretical basis and simplicity, is also easily modified to
account for important factors such as concentrated flow and
vegetative uptake which were not considered by any of the models

discussed. Consequently, the Phillips method appears to be the
most reasonable model for estimating relative buffer strip
effectiveness and satisfying the objectives of this project.

MODEL DEVELOPMENT

The method developed by Phillips (1989a,b) was selected for
evaluating buffer zone effectiveness. The method was modified to
correct several errors and weaknesses in the model and to better
reflect Virginia conditions. In addition, a procedure is presented
to account for the benefits of shoreline protection structures in
riparian buffer zones.

Phillips Hydraulic Model

Several limitations were encountered with Phillips's hydraulic
model. First, the buffer length term, L, is used to estimate the
flow rate per unit width to the buffer zone. But buffer length
does not give an estimation of unit loading to a buffer. What is
needed to estimate loading is the length of the upslope area
contributing runoff to the buffer plus the length of the buffer.
Consequently, an additional term, L*, must be defined that
represents "the length of the upslope area contributing runoff to
the buffer." The L in the Phillip hydraulic model can therefore be
redefined as L + L*.

A second limitation involves the saturated hydraulic conductivity
term, K, which is used to estimate infiltration in the buffer zone.
Phillips's approach does not consider the effects of buffer zone
length on infiltration, ie. infiltration losses from surface runoff
would be as significant with a one meter length buffer as they
would be for a 100 m buffer. To incorporate the effects of buffer
length on total infiltration, the hydraulic conductivity term can
be multiplied by the buffer length, L.

A third limitation of Phillips's approach is the failure to
consider the effects of concentrated flow. Concentrated flow
effects can be incorporated into the model by incorporating the
ratio of the fraction of flow entering the buffer and the reference
buffer as shallow sheet flow.

The new hydraulic model with the above modifications can now be
expressed as:
Bb =( KbLb)@
L,*+L Sr n
.r

where: X saturated hydraulic conductivity of buffer soils
L buffer zone length
L* length of area upslope of the buffer contributing runoff
to the buffer
S sin 0, where 0 is the slope angle relative to the

horizontal
n = manning roughness coefficient
C = fraction of surface runoff crossing the buffer as sheet
flow
b = subscript denoting buffer of interest, and
r= subscript denoting reference buffer

Phillips Detention Model

In checking the derivation of the Phillips detention model, an
error was found in the derivation. The model is rederived as
follows. Given Equation (12] and Darcys law (Equation [14]),
Equation (15] is actually:
Tg=L (19 Ks
Ks
Equation [16] therefore becomes:
T*=[n 0.6 Ls -0.3 (q,/qq)-0.4]* L (qg/q) [20]
Ks

As with the hydraulic model, if K is substituted for (6qq6q/6q q) in the
subsurface flow portion of Equation [201, the equation reduces to:

T*=nO.6L2 S-1.3 (qs,/q q) -0.4 [21]
This equation is identical to Phillips' except the exponent of the
slope term changes from -0.7 to -1.3. The saturated hydraulic
conductivity can also be substituted for qs/q, but in this case,
the sign of the exponent of K must be reversed because surface
runoff is inversely proportional to K since higher values of K
indicate higher potential for infiltration and therefore reduced
runoff. Equation (21] can therefore be simplified to:

T*=n O.6L2s-1.3KO.4 [22]

The resulting hydraulic model with the addition of the soil
moisture storage capacity term discussed previously is:
Bb=nb0.6Lb2Kb 0.4 Sb -1.3 Mb
Br nr Lr Kr Sr Mr [23]

As with the hydraulic model, a concentrated flow term should also
be added to account for short circuiting due to concentrated flow
effects. An additional term is also added to account for the
relative ability of different types of vegetation and vegetative
growth stages to assimilate nutrients. Since little information is
available on nutrient uptake for the ecosystems of interest, the
net productivity of relevant ecosystems is used as an indicator of
nutrient uptake since they are proportional. This factor is

incorporated into the model as a ratio similar to the way
concentrated flow effects are handled. The resulting equation is:
Bb .( n.)0'6( Lb)2 Cb)( Vb) [24]
Br nr Lr K.- M.- Cr V@
where V is the vegetative uptake or net pr, Am @tivity of the
vegetative buffers being compared. It is essential to consider
both the type of vegetation and the maturity of the vegetative
system when estimating net productivity. For example, a riparian
zone in a rapidly growing early secession stage would be expected
to be able to assimilate and hold more nutrients than a more stable
mature riparian buffer approaching or at climax.

The original Phillips detention model did not consider the effects
of the length of the buffer on the soil moisture capacity.
Intuitively, one would expect the soil moisture capacity to be
proportional to the area of the buffer since a buffer twice as long
as another buffer, all other conditions equal, would be expected to
have twice the soil moisture storage capacity. Consequently, M
should be multiplied by L to account for the area of the buffer *
The vegetative uptake factor, V, is also multiplied by L for
identical reasons. The resulting detention model is:
0 6 4 1 4( b)-I,I( C")( V
B Lb) ( b) b) [25]
B. n. Lr K, s, Mr C, V,

Readers should keep in mind that the the original Phillips models
as well as the proposed model are theoretical and have not been
verified due to a lack of adequate field data for validation.

Shoreline Erosion Model

Shoreline erosion: The proposed shoreline erosion model is simple
but should provide a fairly good estimate of sediment and nutrient
losses per meter of shoreline for both protected. and unprotected
shoreline. The model does not estimate sediment and nutrient
losses from partially protected shorelines. For stabilized
shorelines, shoreline erosion and nutrient losses are assumed to be
zero.

If a shoreline is not protected and actively eroding, the sediment
loss is estimated in a manner similar to that recommended by Ibison
et al. (1992):
YB z HB EB pB [26]
where: Y. = bank sediment loss per unit width (kg/m-yr)
E..= bank or shoreline erosion rate (m/yr)
H, = bank height (m)
PB = bulk density of bank soil (kg/m')

If the nutrient content of the bank is known or can be estimated
then:
Y,I YB NB [27]
1000
YBP YB PB [28]
1000
where: YB,, nitrogen lost in eroding bank sediment (kg/m-yr)
YB, phosphorus lost in eroding bank sediment (kg/m-yr)
X. concentration of nitrogen in bank material (mg/g)
PB concentration of phosphorus in bank material (mg/g)

Upland Erosion: To determine the relative contributions. of
sediment and nutrients from shoreline erosion and disturbed upland
requires an estimate of shoreline and upland contributions. Upland
losses need to be estimated per unit width like the shoreline
losses for comparison purposes. Since the actual trapping
efficiency of the buffer is unknown, the conservative assumption
is made that the buffer trapping efficiency is 100% for the portion
of the runoff entering the buffer as shallow sheet flow. All
sediment and nutrients in concentrated flow are assumed to pass
through the buffer unattenuated. Sediment, nitrogen and phosphorus
movement through the buffer can therefore be represented as:

Yj* YU (1-C) (29)
Y@N YUN ( i - C [30]
G YUP (1-C (31]
where: Y,,* mass of sediment from upland area passing through
buffer (kg/m-yr)
Yu sediment loading to the buffer from the upland
contributing area (kg/m-yr)
C fraction of surface runoff from the field entering the
buffer as sheet flow
Yum* mass of nitrogen from upland area moving through
buffer (kg/m-yr)
Y,,. loading of nitrogen to buffer from upland area (kg/m-
yr)
Y,,* mass of phosphorus from upland area moving through
buffer (kg/m-yr)
Ym loading of phosphorus to buffer from upland area
(kg/m-yr)

The sediment loading to the buffer from the upland contributing
area, Yu, can be estimated from published values for various
landused- or it can be calculated directly using the Universal Soil
Loss Equation (Wischmier and Smith, 1976), the Revised Universal
Soil Loss Equation (Renard et al., 1992), WEPP (Lane and Nearing,

1989), or other similar erosion predictor. The estimate will
probably be given in tons/acre-yr and will need to be converted to
kg/m-yr to be consistent with the shoreline erosion units. If the
soil loss units are in tons/acre-yr, then the estimated sediment
loading to the buffer can be converted to kg/m-yr with the
following equation:
Y,= 2242 A,, E [32]
WB
where: A,, upland area contributing sediment to the buffer,
acres
Eu = sediment loss from upland area contributing sediment to
the buffer as predicted by USLE or other method,
tons/acre-yr
W. = buffer width (long dimension of buffer, dimension
perpendicular to field slope direction), m

Similarly, if the nutrient loadings to the buffer are in lbs/acre-
yr, they can be converted to kg/m-yr with the following equations:
YU'V = C1 AU NU (33]
WB

YUP = C1 AU PU [34]
WB Z

where: NU nitrogen loading to buffer from upland area, M/L'
P, = phosphorus loading to buffer from upland area, M/L1
C1 = units conversion factor
= 0.454 for nutrient loadings (N, or Fu) in lb/acre and
upland area (Au) in acres
= 1.0 for nutrient loadings (Mu or Pu) in kg/ha and
upland area (Au) in hectares

Impact of Shoreline Stabilization: If modifications made to the
buffer due to shoreline stabilization reduce the effectiveness of
the buffer as predicted by Equations [18] and [25], then the
combined effects of buffer effectiveness and shoreline
stabilization must be evaluated to assess the effectiveness of the
system as a whole. The combined effects of buffer modification and
shoreline stabilization can be estimated as follows. To compare
the effectiveness of the buffer and shoreline stabilization, it is
necessary to come up with an estimate of the absolute (not
relative) effectiveness of the proposed buffer. The modif ied
Phillips models do not do this directly, but they can be used in an
indirect way to do so.

First assume that the reference buffer is 100% effective in
removing sediment from shallow sheet flow and totally ineffective

in removing contaminants from concentrated flow. The amount of
sediment and nutrients moving through the reference buffer can then
be estimated using Equations [29], [30] and [31]. The relative
effectiveness of the reference and proposed buffer is calculated
using the modified hydraulic model, Equation [18]. The
effectiveness of the buffer/shoreline stabilization system, E,, can
be estimated as:
Es@ YV* [35]
(Bb / B') ( YB + YU.-

where BbIBI is the relative buffer effectiveness predicted with the
hydraulic model. If E,<l, the buffer/shoreline stabilization system
has a net positive effect. If Es>l, the negative consequences of
buffer alteration outweigh the benefits of shoreline stabilization.

Buffer Evaluation Procedure

The steps to follow in applying the model are as follows:

1. Decide upon an acceptable effectiveness of the buffer relative
to the reference buffer in advance. For example, if it is
decided that the buffer must be at least as effective as the
reference buffer then the ratio of B,/B, for both the hydraulic
and detention models must be greater than or equal to 1.0. If
the given buffer only needs to be 90% as effective as the
reference buffer then the ratio of BbIB, would need to be
greater than or equal to 0.9. The important idea is to decide
on the required ratio before analysis.

2. Collect model data required to define the characteristics of
the reference buffer and the buffer of interest. The
reference buffer can be either a reference buffer whose
characteristics are defined by a local regulatory authority or
it may be the characteristics of the buffer you are
considering modifying before modifications. Parameters that
must be defined for both buffers include:

K = saturated hydraulic conductivity of buffer soils
L = buffer zone length
L* = length of area upslope of the buffer contributing runoff
to the buffer
s = sin 0, where 6 is the slope angle relative to the
horizontal
n = manning roughness coefficient
C = fraction of surface runoff crossing the buffer as sheet
flow (modify for proposed hydraulic modifications in
proposed buffer)
N = soil moisture storage capacity
V = vegetative uptake or net productivity factor

Procedures, sources of information and tables for estimating
these values are given in the following section.

3. Estimate the effectiveness of the proposed buffer for removing
sediment and sediment-bound chemicals using Equation [18], the
hydraulic model. If the proposed buffer is not as effective
as desired, consider increasing the buffer length and changing
the proposed vegetation (higher density vegetation may be used
to increase Manning's n). It may also be possible to reshape
the upland area or install hydraulic structures to reduce
concentrated flow and increase the proportion of sheet flow
entering the buffer. It is unlikely that properties such as
slope and hydraulic conductivity can be changed.

4. Estimate the effectiveness of the proposed buffer relative to
the reference buffer for removing dissolved chemicals using

Equation [25], the detention model, and the data values
collected in step 1. If the proposed buffer is not as
effective as desired, consider increasing the buffer length
and changing the proposed vegetation (higher density
vegetation may be used to increase Manning's n and the
vegetative uptake factor, V). It may-also be possible to
reshape the upland area or install hydraulic structures to
reduce concentrated flow and increase the proportion of sheet
flow entering the buffer. It is unlikely that properties such
as slope and hydraulic conductivity can be changed.

5. If shoreline erosion is not a factor, the evaluation and/or
design is complete. If shoreline erosion is to be considered,
continue with steps 6 through 9.

6. Calculate sediment, nitrogen and phosphorus losses due to
shoreline erosion using Equations (26], [27] and (28],
respectively.

7. Estimate sediment, nitrogen, and phosphorus losses from the
field contributing surface runoff to the buffer using the USLE
or other soil loss estimation technique, estimates of soil
nutrient concentrations, or general estimates of sediment and
nutrient losses from the literature for various landuses.

8. Calcu late sediment, nitrogen and phosphorus transport through
the buffer using Equations [29], [30], and [31].

9. Calculate the overall effectiveness of the buff er/shore line
protection system using Equation (35]. If E,<l then the
combined buffer/shoreline stabilization system is more
beneficial than the original buffer alone. If E,>1, shoreline
stabilization results in increased sediment and nutrient
losses.

DATA FOR BUFFER EVALUATION PROCEDURE

Data Requirements for the Hydraulic and Detention Models

The following values must be estimated to use the modified
hydraulic and detention models:

Saturated hydraulic conductivity of buffer soils, K
Buffer length, L
Length of area upslope of the buffer contributing runoff to the
buffer, L*
Sin 0. where 0 = slope angle relative to the horizontal, s
Manning roughness coefficient, n
Fraction of surface runoff crossing buffer as sheet flow, C
Soil moisture storage capacity, M
Vegetative uptake or net productivity factor, V

These values can be estimated as follows.

Saturated Hydraulic Conductivity, K: can best be estimated from
permeability values given in modern county soil survey reports.
These values are generally presented in the "Physical and
Chemical Properties of the Soils" table. Permeability values
are usually given as a range in inches\hour. For planning
purposes, use the midpoint of the range unless better
information on the permeability value is available. Typical
permeability ranges given in soil surveys are given in Table I.
Hydraulic conductivity can also be estimate if the general soil
texture is known as shown in Table Il. Soil survey estimates
are preferred, however.

Table I. Estimation of hydraulic
conductivity values from Soil
Survey permeability estimates.

Permeability Suggested
Range, Hydraulic
in/hr Conductivity

------------ ------------
0.06-0.2 0.13
0.2-0.6 0.40
0.6-2.0 1.30
0.6-6.0 3.30
6.0-10.0 8.00

Table II. Hydraulic conductivity as a function of soil
texture.

Hydraulic Conductivity, m/day
Soil Texture Range Suggested

Clay soils, surface 0.01-0.2 0.10
Loam soils, surface 0.1-1 0.50
Fine sand 1-5 2.00
Medium sand 5-20 10.00
Coarse sand 20-100 40.00
Clay, sand, gravel mix 0.001-0.1 0.05

Buffer Length, L: is the distance measured along the land surface
between the upland edge of the buffer and the down slope edge
of the buffer. Buffer length is best measured in the field but
it may also be estimated from areal photos. Any units may be
used as long as they are consistent with the units used in
estimating the upslope slope-length contributing runoff to the
buffer, L*.

Upslope Slope-length Contributing Runoff to the Buffer, L*: is the
average length that surface runoff tranverses in the field or
area before reaching the buffer. It is best estimated from on
site field estimates, but it can also be estimated from areal
photographs. The average upslope Contributing slope-length can
also be estimated as:
L* = AU [361
WB

where A. is the area of the field contributing runoff to the
buffer, andWB is the width of the buffer (long dimension of the
buffer).

Sin 0, s: is a measure of the land slope across the buffer where
0 is the slope angle relative to the horizontal. The value of
s is best measured from in buffer surveys but it can also be
estimated from topographic maps. Values for Sin 0, s can be
estimated from estimates of buffer slope in percentage or
degrees using the data given in Table III.

Table III. Estimates of Sin(O) for various buffer slopes in
percent and degrees.

Buffer Buffer S, Buffer Buffer S,
Slope, Slope, Sin(O) Slope, Slope, Sin(O)
percent Degrees percent- Degrees

0.5 0.3 0.0050 40 21.8 0.3714
1 0.6 0.0100 45 24.2 0.4104
2 1.1 0.0200 50 26.6 0.4472
3 1.7 0.0300 60 31.0 0.5145
4 2.3 0.0400 70 35.0 0.5735
5 2.9 0.0499 80 38.6 0.6247
6 3.4 0.0599 90 42.0 0.6690
7 4.0 0.0698 100 45.0 0.7071
8 4.6 0.0797 120 50.2 0.7682
9 5.1 0.0896 140 54.4 0.8137
10 5.7 0.0995 160 58.0 0.8480
12 6.8 0.1191 180 60.9 0.8742
14 8.0 0.1386 200 63.4 0.8944
16 9.1 0.1580 300 71.5 0.9487
18 10.2 0.1772 600 80.5 0.9864
20 11.3 0.1961 1200 85.2 0.9965
25 14.0 0.2425 2400 87.6 0.9991
30 16.7 0.2873 4800 88.8 0.9998
35 19.3 0.3304 9600 89.4 0.9999

Manning's Roughness Coefficient, n: is a measure of the roughness
of the lands surface in the buffer. It is used to estimate how
much surface runoff is retarded (reduction in velocity of
overland flow) as it passes through the buffer. Manning's
roughness coefficient is difficult if not impossible to measure
directly, so it is usually estimated from tabular values that
give the roughness coefficient as a function of land use and
condition. The roughness coefficients used in the buffer
evaluation procedure should be for shallow flow conditions, not
for the more readily available channel flow conditions.
Manning's roughness coefficient for shallow sheet flow
conditions can be estimated from Table IV.,

Table IV. Estimates of Mannings roughness
coefficient as a function of landuse and
condition for shallow sheet flow
conditions (Engman, 1986).

Landuse and Condition Suggested
Value

-------------------------- ----------
Forest (light underbrush) 0.30
Forest (dense undergrowth) 0.40
Bare sand 0.01
Bare clay loam (eroded) 0.02
Fallow - no residue 0.05
Chisel plow <0.25t/ac residue 0.07
Chisel plow 0.25-1t/ac residue 0.18
Chisel plow 1-3 t/ac residue 0.30
Chisel plow >3 t/ac residue 0.40
Disk/harrow <0.25 t/ac residue 0.08
Disk/harrow 0.25-1 t/ac residue 0.16
Disk/harrow 1-3 t/ac residue 0.25
Disk/harrow >3 t/ac residue 0.30
No-till <0.25 t/ac residue 0.04
No-till 0.25-1 t/ac residue 0.07
No-till 1-3 t/ac residue 0.30
Moldboard plow (fall) 0.06
Coulter 0.10
Range (natural) 0.13
Range (clipped) 0.10
Grass (Bluegrass sod) 0.45
Short grass (prairie) 0.15
Dense grass 0.24
Bermuda grass 0.41

Fraction of surface runoff crossing buffer as sheet flow, C: is an
estimate of the amount of surface runoff from the field
contributing runoff to the buffer that enters and presumably
crosses the buffer as shallow sheet flow. it is a dif f icult
and somewhat arbitrary value to estimate. The f raction of
surface runoff crossing the buffer as sheet flow can best be
estimated using the following procedure:

1. Survey the area upslope of the buffer that contributes
surface runoff to the buffer and estimate the area of the
upslope contributing area, A,.
2. Delineate the major drainageways within the upslope
contributing area.
3. Delineate and measure the area, Aj, contributing surface
runoff to each drainageway, i.
4. The fraction of surface runoff crossing the buffer as

shallow sheet flow can then be estimated as:
C =AUE A, (37] A
AU
Soil moisture storage capacity, M: is def ined as the available soil
moisture storage capacity of the buffer soils times the lessor
of the depth to the seasonal high water table or the depth to an
impeding soil layer (i.e. impervious or semi impervious soil
layer. Available soil moisture is defined as the difference
between the soil water content at field capacity and the wilting
point. Estimates of available soil moisture are given in the
"Physical and Chemical Properties of the Soils" table of modern
soil survey reports or in soil property databases (i.e. the
Soils-5 database) as a function of soil type. Soil moisture
storage capacity can therefore be estimated as:

M = 8A D [38]

where OA is the available soil moisture of the buffer soils and
D is the lessor of the depth to the seasonably high water table
or the depth to an impeding soil layer.

Vegetative uptake or net productivity factor, V: is used as an
index of the buffer vegetation's ability to assimilate
nutrients. Since little information was available on the
nutrient assimilative capacity of vegetative ecosystems in
Virginia, it was decided to use net primary productivity as an
indicator of net nutrient uptake. This is reasonable since
nutrient uptake and net primary productivity are highly
correlated. Unfortunately, data on net productivity for
Virginia ecosystems is almost as rare as data on net nutrient
uptake, particularly as a function of sucessional stage. In
instances where the buffer under study and the reference buffer
are composed of the same type and age of vegetation, the
vegetative uptake factor will have no effect on buffer
performance. The values in Table V are gross estimates of net
primary production. The user is encouraged to find better
sources.of information on net primary production or net nutrient
uptake for the region being investigated if possible.

Table V. Net primary productivity (adapted from: Leith, 1975)

Vegetation Net Primary Productivity,
g/m'-yr

Mixed forest, dense under story 1000
Mixed forest, sparce under story 500
woodland, good cover 600
Woodland, sparce cover 400
Grass, good stand 500
Grass, poor stand 300
Woodland, shrubs, grass 800
Cultivated land 650
Swamps and marsh 2000

Data Requirements for Shoreline Stabilization Model

The following additional values must be estimated to evaluate the
benefits of shoreline stabilization:

Bank or shoreline erosion rate (m/yr), E,,
Bank height (m), H,
Bulk density of bank soil (kg/ml), p,,
Concentration of nitrogen in bank material (mg/g), N,
Concentration of phosphorus in bank material (mg/g), P.
Sediment loading to the buffer from the upland contributing area
(kg/m-yr), Y.
Nitrogen loading to buffer from upland area, M,,
Phosphorus loading to buffer from upland area, P,,

These values can be estimated as follows.

Bank or shoreline erosion rates (m/yr), E,,: are very difficult to
measure. Measurement techniques include interpretation of areal
photos over a period of years or actual measurements of
shorelines relative to fixed measurement points. In any case,
years of measurements are required to obtain average annual
erosion rates. Consequently, estimates of shoreline erosion for
planning purposes are best obtained from published reports on
shoreline erosion in Virginia (Byrne and Anderson, 1977; Ibison
et al., 1990; 1992; Hardaway et al., 1992).

Bank height (m), H,: measurements must be made in the field or from
detailed topographic surveys of the study site if they exist.
Bank height should be measured vertically from the base to the
top of the bank.

Bulk density of bank soil (kg /M3) ,p.: can be determined in the
laboratory from undisturbed field samples or from one or more
field techniques. It can also be estimated from soil survey
report descriptions of soil horizons. Soil bulk densities of
coastal soils generally range from 1.3 to 1.8 g/cm3 (multiply
g/CM3 by 1000 to get units of kg /M3 ) . If a soil bank has
distinct horizons with varying bulk densities, the bulk
densities of each horizon should be area weighted to calculate
an average bulk density for the whole profile. if bulk density
information is not available, assume an average value of 1.5
g/CM3 ( 1500 kg /M3 ) .

Concentration of nitrogen, N,,, and phosphorus, P,, in the bank
material (mg/g): must be estimated from laboratory analyses as
suggested by Ibison et al. (1992). To reduce the costs of
laboratory analyses, a single composite soil sample can be
created by combining soil samples from each soil horizon in
proportion to the thickness of each horizon. If laboratory
analysis of soil samples is not possible, then estimates of soil
nutrient levels can be approximated with concentrations reported
by Ibison et al. (1990, 1992) for similar landuses and
locations.

Sediment loading to the buffer from the upland contributing area
(kg/m-yr), Y,: are best determined by direct calculation using
the Universal Soil Loss Equation (Wischmier and Smith, 1976),
the Revised Universal Soil Loss Equation (Renard et al., 1991),
WEPP (Lane and Nearing, 1989), or other similar erosion
predictor. The estimate will probably be given in tons/acre-yr
(except WEPP predictions, which are given in kg/m-yr) and will
need to be converted to kg/m-yr to be consistent with the
shoreline erosion units. If the soil loss units are in
tons/acre-yr, then the estimated sediment loading to the buffer
in kg/m-yr can be determined with Equation [32]. Rough
estimates of sediment loading as a function of landuse can also
be obtained from the literature.

Nitrogen, Nu, and phosphoru5f Pul loading to the buffer: from upland
areas can be estimated in several ways. First, estimates can be
by collecting and analyzing soil samples from the area
contributing runoff to the buffer. The average soil nutrient
concentration can then be multiplied by the estimated soil loss
from the area with appropriate unit conversions to estimate
nutrient loadings to the buffer. Nutrient loadings can also be
estimated using models such as CREAMS (Knisel, 1980). Lastly,
nutrient loadings to the buffer can be estimated from published
unit area nutrient losses for various landuses. For example,
Beaulac and Reckhow (1982, cited by Ibison et al., 1992),
reported nutrient losses from cropland of 1.9 to 71 lbs-N/ac-yr
(2.1 to 80 kg-N/ha-yr) and 0.23 to 17 lbs-P/ac-yr (0.26 to 19
kg-P/ha-yr). Areal nutrient loadings can be converted to a per
unit width basis using Equations (33] and (34].

PRACTICAL APPLICATION

Example Problem I

Problem Statement: For a site with the following conditions,
determine how long a buffer is required to be at least as effective
as the given reference buffer.

Data:

Reference buffer characteristics

Buffer soil: Suffolk sandy loam (K = 1.3 m/day)
Buffer length, L = 100 ft (30.5 m)
Field slope-length, L* 400 ft (122 m)
Buffer slope = 10% (s 0.0995 from Table III)
Buffer vegetation is forest with heavy undergrowth (n = 0.4 from
Table IV)
Fraction of runoff crossing buffer as sheet flow (C = 0.5)
Soil moisture storage capacity (available soil moisture = 0.15
from soil survey and depth to water table = 5 m, M = 0.15*5
0.75)
Vegetative uptake for forest (V = 1000 from Table V)

Study buffer

All conditions are the same except the landowner wants to thin
the trees, without improving ground cover, to improve his view
of the water.

For forest with light under brush, n = 0.3 and V = 500. All
other factors are the same except L and L*.

Solution:

Since the new buffer must be as effective as the reference buffer,
BbIB, must be greater tan or equal to 1. Substituting the known
values into the hydraulic model, Equation [18]:
1. 0=( 1 . 3Lb )( 00)-1.4( 0 , 0995)-1,3( 0 .3)0.6 (0-5 0.0084 Lb
1.3*100 500 0.0995 0.4 0.5)=
Solving the above equation gives L, = 119 ft (36 m). Therefore, the
landowner would be required to have a 118 ft buffer rather than a
100 ft buffer if the landowner wants to thin the forest vegetation
according to the hydraulic model.

In a similar manner, the detention model, Equation [25], gives:

Solving the detention equation gives Lb = 124 ft (38 m). This value

1 =(0.3)0.6( Lb)4 1. 3 0.4 0.0995)-1-3 0.75)(0.5 500 )=4.29*10-9L4
0.4 100 1.3 0.0995 0.75 0.5 1000

is very similar to that recommended by the hydraulic model.

Example Problem II

Problem Statement: Now suppose the landowner in Example Problem I
decides that she wants to protect her property from shoreline
erosion. Given the following information, what would the impact of
shoreline stabilization be on sediment and nutrient losses from her
combined buffer/shoreline protection system?

Data:

Assume that the landowner's property has a steep bank immediately
next to the shore. Characteristics of the bank are as follows:

Bank or shoreline erosion rate (m/yr) , E, = 0. 5 m/yr (1. 64 ft/yr)
Bank height (m) , Hs = 5 m (16.4 ft)
Bulk density of bank soil (kg/m') , PB = 1500 kg /M3
Concentration of nitrogen in bank material (mg/g), N,=0.365 mg/g
Concentration of phosphorus in bank material (mg/g) , PB=O. 25 mg/g
Sediment loading to the buffer from the upland contributing area
(kg/m-yr), Y,, = 338 kg/m-yr for a 305 m wide buffer and a 9.18
acre field contributing runoff to the buffer with a soil loss
rate of 5 tons/ac-yr
Nitrogen loading to buffer from upland area, N,,;:: 10 kg/ha-yr
Phosphorus loading to buffer from upland area, P,= 5 kg/ha-yr

Solution:

To stabilize the shoreline requires that the bank be graded back to
a grade of 2:1 and that all large tree be removed from a distance of
2 bank heights from the top of the new graded slope. This means
that of the original 38 m (124 ft) buffer (from Example Problem I),
10 m will be graded and planted to grass, small trees and shrubs;
the next 10 m will have all the large trees removed, and the
remaining 18 meters will be the thinned trees described in Example
Problem I.

For the 2:1 graded region, the slope will now be 30' or s=0.5. For
the rest of the slope, s=0.0995. Area weighting the slope gives a
mean slope of s=(10*0.5+28*0.0995)/38=0.2049

The weighted Manning's roughness coefficient for the buffer will be
(n=0.24 for the stabilized grass slope, n=0.3 for the forested
area): n=(10*0.24+28*0.3)/38=0.284. The vegetation factor will be
v=500 since both light woods and grass have the same net primary
productivity value.

Buffer effectiveness according to the hydraulic model is:
Bb=(1-3*124)(500)-0.4 0.2048)-l-3( 0.284 01 0 . 5)= 0.395
Br 1.3*100 500 (0.0995 0.4 0.5

Buffer effectiveness according to the detention model is:
Bb (0.284 )0.6( 124 )4( 1. 3 )0.4( 0 . 2049 )-1.3( 0 . 75 )(0.5)( 500)= 0.376
Br 0.4 (100) (1.3) (0.0995) 0. 7 5 0 0. 5 1000
Both the hydraulic and detention models predict greatly reduced
buffer effectiveness due to the grading done to stabilize the
shoreline bank. Both models predict that the buffer will only be
about 40% as effective as the reference buffer.

However, if the benefits of shoreline stabilization are considered,
the overall system may still be beneficial.

Sediment and nutrient losses due to shoreline erosion:

With the information given above, sediment loss due to shoreline
erosion is:
YB = HB EB PB = 5*0.5*1500 = 3750 kg/m-yr

In a similar manner, nutrient losses due to shoreline erosion are:
YBN - YB NB 3750*0.65 = 2.437 kg-N/m-yr
1000 1000
YBP - YB PB 3750*0.25 = 0.937 kg-P/m-_yr
1000 1000

Sediment and nutrient transport through the buffer:

Sediment passing through the buffer can be estimated as:
Yu*=Yu(l-C) = 2242 Au Eu (1-C) = 2242 8.63*5 (1-0.5) = 159 kg/m-yr
WB 304.8
Similarly, nutrients passing through the buffer are:
Yun=Yun(1-,_C)=C1 Au Nuq(1-c) = 13.49*10 (1 -0. 5 ) =0. 057 kg-N/m-yr
WB 305

Yup=Yup (1-_C)= C1, AU Pu (1-C) =1 3. 49 *5 (1-0. 5 ) =0. 029 kp-P /m-yr
WB 305

The overall effectiveness of the buffer/ shoreline stabilization
system for sediment reduction can then be represented by Equation
[35]:
Since Es,<1, shoreline stabilization results in greater sediment loss
reduction than the buffer alo ne. The shoreline stabilization system
results in a 90% decrease in sediment losses. Equation [35] can

ES Y@ 159 =0. 102
(BI, / B,) ( Yr, + YU* 0.395*(3750+159)

also be applied to nutrient losses. The resulting values of ES for
nitrogen and phosphorus are 0.058 and 0.076, respectively. This
means that the shoreline stabilization/buf f er system reduced overall
nitrogen and phosphorus losses by 94 and 92%, respectively.

SUMMARY AND CONCLUSIONS

The proposed model is relative simple in concept and application and
is suitable for use by planners. All of the data required by the
model can be collected on site or can be estimated from the
literature. Labratory analysis of soil and bank samples, however,
will greatly improve model reliability with respect to nutrient
losses. In areas with shoreline erosion, the procedure also allows
the benefits of shoreline control to be considered.

REFERENCES

Bagnold, R.A. 1977. Bed load transport by natural rivers. Water
Resources Research 13(2):383-312.

Barfield, B. J., E. W. Tollner and J. C. Hayes. 1977. Prediction
of sediment transport in grassed media. ASAE Paper No. 77-2023.
Am. Soc. of Agric. Engrs., St. Joseph, Mich.

Barfield, B. J., E. W. Tollner and J. C. Hayes. 1979. Filtration
of sediment by simulated vegetation, I. Steady-state flow with
homogeneous sediment. TRANSACTIONS of the ASAE 22(3):540-545, 548.

Beaulac, M.N. and K.H. Reckhow. 1982. An examination of landuse-
nutrient export relationships. Water Resources Bulletin 18:1013-
1023.

Byrne, R.J. and G.L. Anderson. 1977. shoreline erosion in Tidewater
Virginia: Special Report in Applied Marine Science and ocean
Engineering No. 111. Virginia Institute of Marine Sciences,
Gloucester Point. 102 p.

Dillaha, T.A., J.H. Sherrard, and D. Lee, 1986. Long-term
effectiveness and maintenance of vegetative filter strips.
VPI-VWRRC-BULL 153, Virginia Polytechnic Institute and State
University, Blacksburg, 55 p.

Dillaha, T.A., R. B. Reneau, S. Mostaghimi, V.0. Shanholtz, and W.
L. Magette, 1987. Evaluating nutrient and sediment losses from
agricultural lands: vegetative filter strips. CBP/TRS 4/87. U. S.
Environmental Protection Agency, Chesapeake Bay Liaison Office,
Annapolis, Md., 93 p.

Dillaha, T.A., J.H. Sherrard, D. Lee, S. Mostaghimi, and V.0.
Shanholtz. 1988. Evaluation of vegetative filter strips as a best
management practice for feedlots. Journal of the Water Pollution
Control Federation, 60:1231-1238.

Dillaha, T.A., R.B. Reneau, S. Mostaghimi, and D. Lee. 1989
Vegetative filter strips for agricultural nonpoint source pollution
control. Transactions of the ASAE 32(2):491-496.

Dillahd, T.A. 1989. Personnel notes from the Work Conference on
Filter Strips, March 6-9, 1989. U.S. Soil Conservation Service,
Washington, D.C.

Dillaha, T.A. and J.C. Hayes. 1992. Vegetative filter strips: II.
Application of design procedures. ASAE Paper No. 92-2103. Am. Soc.
Of Agric. Engrs., St. Joseph, Mich. 18 p.

Doyle, R. C., G. C. Stanton and D. C. Wolf. 1977. Effectiveness of
forest and grass buffer filters in improving the water quality of
manure polluted runoff. ASAE Paper No. 77-2501. Am. Soc. Of
Agric. Engrs., St. Joseph, Mich.

Edwards,. W. M., L. B. Owens, and R. K. White. 1983. Managing
runoff from a small, paved beef feedlot. J. Environ. Qual. 12:
281-286.

Engm an, E.T. 1986. Roughness coefficients for routing surface
runoff. Journal of Irrigation and Drainage Engineering, ASCE
112(l):39-53.

Flanagan, D.C., W.H. Neibling, G.R. Foster, and J.P. Hurt. 1986.
Application of CREAMS in filter strip design. ASAE Paper No.
86-2043. Am. Soc. Of Agric. Engrs., St. Joseph, Mich.

Hardaway, C.S., G.R. Thomas, J.B. Glover, J.B. Smithson, M.R.
Berman, and A.K. Kenne. 1992. Virginia Bank Erosion Study. Contract
report to Virginia Department of Conservation and Recreation,
Division of Soil and Water Conservation. Virginia Institute of
Marine Science, Gloucester Point, Vir.

Hayes, J.C., B.J. Barfield, and R.I. Barnhisel. 1978. Evaluation of
grass characteristics related to sediment filtration. ASAE Paper
No. 78-2513. Am. Soc. of Agric. Engrs., St. Joseph, Mich.

Hayes, J.C. 1979. Evaluation of design procedures for vegetal
filtration of sediment in flowing water. strips. A thesis
presented to the University of Kentucky in partial fulfillment of
the requirements for the degree of Doctor of Philosophy. Lexington.

Hayes, J. C., B. J. Barfield, and R. I. Barnhisel. 1979a.
Filtration of sediment by simulated vegetation II. Unsteady flow
with non-homogeneous sediment. TRANSACTIONS of the ASAE
22(5):1063-1067.

Hayes, J. C., B. J. Barfield, and R. I. Barnhisel. 1979b.
Performance of grass filters under laboratory and field conditions.
ASAE Paper No. 79-2530. Am. Soc. of Agric. Engrs., St. Joseph,
Mich.

Hayes, J.C., B.J. Barfield, and R.I. Barnhisel. 1982. The use of
grass filters for sediment control in strip mine drainage: Volume
III. Empirical evaluation of procedures and verification. Report
No. IMMR82/070. Inst. for Mining and Minerals Research, Univ. of
Kentucky, Lexington.

Hayes, J.C. and J.E. Hairston. 1983. Modeling the long-term
effectiveness of vegetative filters on on-site sediment controls.
ASAE Paper No. 83-2081. Am. Soc. of Agric. Engrs., St. Joseph,
Mich.

Hayes, J.C. and T.A. Dillaha. 1992. Vegetative filter strips: I.
Site suitability and procedures. ASAE Paper No. 92-2102. Am. Soc.
Of Agric. Engrs., St. Joseph, Mich. 17 p.

Hill, C.L. 1992. Personal communication. Department of Conservation
and Recreation, Division of Soil and Water Conservation, Shoreline
Programs Bureau, Gloucester Point, VA.

Ibison, N.A., C.W. Frye, J.E. Frye, C.L. Hill and N.H. Burger. 1990.
Sediment and nutrient contributions of selected eroding banks of the
Chesapeake Bay estuarine system. Department of Conservation and
Recreation, Division of Soil and Water Conservation, Shoreline
Programs Bureau, Gloucester Point, VA. 71 p.

Ibison, N.A., J.C. Baumer, C.L. Hill, N.H. Burger and J. E. Frye.
1992. Eroding bank nutrient verification study for the lower
Chesapeake Bay. Department of Conservation and Recreation, Division
of Soil and Water Conservation, Shoreline Programs Bureau,
Gloucester Point, VA. 80 p.

Kao, T.Y., B.J. Barfield, and R. I. Barnhisel, 1975. On-site
sediment filtration using grass strips. National Symposium on urban
Hydrology and Erosion Control, Univ. of Kentucky, Lexington.

Kao, T.Y. and B.J. Barfield. 1978. Predictions of flow hydraulics
of vegetated channels. TRANSACTIONS of the ASAE 21(3):489-494.
Knisel, W.G., ed. 1980. CREAMS: A field-scale model for chemical,
runoff, and erosion from agricultural management systems.
Conservation Research Report No. 26. U.S. Dept. Agric., Washington,
D.C. 640 p.

Knisel, W.G. 1980. CREAMS: A field-scale model for chemicals ' runoff
and erosion from agricultural management systems. Conservation
Research Report No. 26. US Dept. Argic., Washington, DC.

Lane, L.J. and M.A. Nearing. 1989. USDA - Water erosion prediction
project: hillslope profile model documentation. NERSL Report No. 2,
USDA-ARS National Soil Erosion Research Laboratory, West Lafayette,
IN.

Lee, D., T.A. Dillaha, and J.H. Sherrard. 1989. Modeling phosphorus
transport in grass buffer strips. Journal of the Environmental
Engineering Division, ASCE 115(2):408-426.

Leith, H. 1975. Primary production of the major vegetation units of
the world. In: Primary Productivity of the Biosphere, Leith, H. and
R.H. Whittaker, eds., Springer-Verlag, Inc., New York. pp. 203-215.

Magette, W.L., R.B. Brinsfield, R.E. Palmer, and J.D. Wood. 1989.
Nutrient and sediment removal by vegetated filter strips.
TRANSACTIONS of the ASAE 32(2):663-667.

Midwest Plan Service. 1985. Livestock waste facilities handbook.
MWPS-18. Midwest Plan Service, Iowa State Univ., Ames, Iowa.

Neibling, W. H. and E. E. Alberts. 1979. Composition and yield of
soil particles transported through sod strips. ASAE Paper No.
79-2065. Am. Soc. of Agric. Engrs., St. Joseph, Mich.

Norman, D. A., W. M. Edwards and L. B. Owens. 1978. Design
criteria for grass filter areas. ASAE Paper No. 78-2573. Am. Soc.
of Agric. Engrs., St. Joseph, Mich.

Paterson, J. J., J. H. Jones, F. J. Olsen, G. C. McCoy, 1977. Dairy
liquid waste distribution in an overland flow vegetative-soil filter
system. ASAE Paper No. 77-4582. Am. Soc. of Agric. Engrs., St.
Joseph, Mich.

Phillips, J.D. 1989a. Nonpoint sources pollution control
effectiveness of riparian forests along a coastal plain river. J. of
Hydrology 110(1989):221-237.

Phillips, J.D. 1989b. An evaluation of the factors determining the
effectiveness of water quality buffer zones. J. of Hydrology
107(1989):133-145.

Renard, K.G., G.R. Foster, G.A. Weesies and J.P. Porter. 1991..
RUSLE: Revised universal soil loss equation. Journal of Soil and
Water Conservation 46(l):30-33.

Thompson, D. B., T. L. Loudon, and J. B. Gerrish, 1978. winter and
spring runoff from manure application plots. ASAE Paper No.
78-2032. Am. Soc. of Agric. Engrs., St. Joseph, Mich.

Tollner, E. W., B. J. Barfield, C. T. Haan and T. Y. Kao. 1976.
Suspended sediment filtration capacity of simulated vegetation.
TRANSACTIONS of the ASAE 20(4):940-944.

Tollner, E. W., J. C. Hayes, and B. J. Barfield. 1978- The use of
grass filters for sediment control in strip mine drainage, Vol. 1.
Theoretical studies on artificial media. IMMR 35-RRR2-78, Institute
for Mining and Minerals Research, University of Kentucky, Lexington.

Tollner, E. W., B. J. Barfield, and J. C. Hayes. 1982.
Sedimentology of erect vegetal filters. Proceedings of the
Hydraulics Division, ASCE 108(12):1518-1531.

Warner, R.C., B.N. Wilson, B.J. Barfield, D.S. Logsdon, and P.L.
Nebgen. 1984. A hydrology and sedimentology watershed model. Part
II: Users' manual. In Design manual for the SEDIMOT II hydrology
and sedimentology model. Dept. of Agricultural Engineering,
University of Kentucky, Lexington.

Williams, R.D. and A.D. Nicks. 1988. Using CREAMS to simulate
filter strip effectiveness in erosion control. J. of Soil and Water
Conservation 43(l):108-112.

Wilson, L.G. 1967. Sediment removal from flood water by grass
filtration. TRANSACTIONS of the ASAE 19(l):35-37.

Wilson, B.N., B.J. Barfield, I.D. Moore. 1984. A hydrology and
sedimentology watershed model. Part I: Modeling techniques. In
Design manual for the SEDIMOT II hydrology and sedimentology model.
Dept. of Agricultural Engineering, University of Kentucky,
Lexington.

Wischmeier, W.H. and D.D. Smith. 1978. Predicting rainfall erosion
losses - a guide to conservation planning. Agriculture Handbook No.
537. U.S. Department of Agriculture. 58 p.

Young, R. A., T. Huntrods, and W. Anderson. 1980. Effectiveness of
vegetative buffer strips in controlling pollution from feedlot
runoff. J. Environ. Qual. 9:483-487.

Young, R. A., M. A. Otterby, and A. Roos. 1982. An evaluation
system to rate feedlot pollution potential. USDA-ARS, ARm-NC-17.
Washington, DC.

a a 0 0