United States
Environmental Protection
Agency
Robert S. Kerr Environmental
Research Laboratory
Ada OK 74820
Research and Development
EPA-600/S2-83-071 Nov. 1983
&ERA Project Summary
Effect of Animal Grazing on
Water Quality of Nonpoint
Runoff in the Pacific
Northwest
Keith E. Saxton, Lloyd F. Elliott, Robert I. Papendick,
Michael D. Jawson, and David H. Fortier
This study was conducted to evaluate
the effects of cattle grazing on runoff
volumes and rates, erosion and sedi-
mentation, and chemical and bacteria
concentrations and quantities in the
surface runoff discharged from agri-
cultural watersheds in the principle
grazing areas of the Pacific Northwest.
The purpose of this study was to
document these water quantity and
quality variables over a period of 3 years
from a small watershed managed in a
typical fashion for this region under
sustained economic production with-
out obvious resource abuse. The effects
of management alternatives were not
evaluated, but the impact of cattle
presence was defined by excluding
grazing on a smaller, but otherwise
similar, grazed watershed. Although a
wide variety of water quality parameters
were measured on the analyzed sam-
ples (approximately 23 types of deter-
minations), special emphasis was given
to the bacterial analyses because the
animal presence was expected to have
more effect on these than on any other
parameter.
Streamflow water quality was inten-
sively studied for 3 years (1976-1979)
on a grazed [21.5 hectares (ha)] and an
ungrazed check (0.9 ha) watershed near
Potiatch. Idaho. The objective was to
identify water quantity, erosion, and
water quality from a typical summer
grazed watershed in this region of
winter precipitation and to identify
animal impacts by comparison with the
ungrazed watershed. Special emphasis
was placed on bacteriological water
quality measurements and interpre-
tations.
The study period contained a near-
drought year and two more nearly
normal years with significant runoff
that provided good water quality
determinations. Erosion was minimal
on the grazed watershed although
cattle trails were an obvious source.
Chemicals from the grazed and un-
grazed watersheds were of low concen-
trations and quantities and the water
quality was not impaired for most uses.
Indicator bacterial numbers were often
high and were closely related to cattle
presence on the watershed. Unexpect-
ed persistence of indicator bacteria was
found after fall removal of the livestock
and significant numbers were found in
the spring months after the temperature
raised and before grazing began. These
results show that bacterial quality is
related to livestock, but there is consider-
ble doubt that indicator bacterial water
quality standards developed for point
source are appropriate for assessing
nonpoint source bacterial contamination.
More research is needed to identify
appropriate bacterial indicators for
nonpoint runoff. Research is also
needed to determine the effect of
alternative pasture and grazing man-
agement on runoff quantity and quality.
This Project Summary was developed
by EPA's Robert S. Kerr Environmental
Research Laboratory. Ada, OK, to
-------
announce key findings of the research
project that is fully documented in a
separate report of the same title (see
Project Report ordering information at
back).
Introduction
In this age of environmental aware-
ness and increased population density, it
is urgent to define and improve stream
water quality if streams are to continue to
be a source of beauty, recreation, and
water supply. Streamflow which ema-
nate as surface runoff from agriculture
pastureland is suspected of having less
than desired quality because of the
obvious possibility of fecal, chemical, and
sediment contamination. Cattle grazing
occupies a significant portion of the
agricultural landscape in much of the
nation, with cattle densities ranging from
feedlots to many hectares per animal on
the western desert ranges. Management
for protection of soil, vegetation, and.
water resources ranges from (a) total
sacrifice and disregard to (b) normal
protection for continuous economic
protection to (c) near disuse or major
conservation efforts.
The effect of cattle grazing on down-
stream water quality with various manage-
ment sherries is currently very poorly de-
fined. Scientific judgement would indi-
cate that the water quality and flow rates,
erosion and sediment production, chemi-
cals, and bacteria could all be altered by
the presence of livestock on a watershed
and the management practices associ-
ated with this agricultural operation.
However, there is no method of predicting
these effects other than through experi-
mental data obtained by scientific study.
These results, when combined with those
of related studies, will provide inferences
of cause and effects.
The result of not obtaining the envi-
ronmental impact knowledge of agricul-
tural production operations such as cattle
grazing could be potentially devastating
should regulations and controls be man-
dated without scientific basis. Even the
uninitiated recognize the fact that every
operation on agricultural land will have
some environmental and downstream
impact, and changes and demands for
restrictions and controls are inevitable. It
is imperative that scientific facts and
documented alternatives be available so
that society can make intelligent choices
on what controls, if any, are necessary.
This study of agricultural runoff water
quality was one of four similar studies
cooperatively sponsored by the U.S.
Environmental Protection Agency (EPA). The
other three were located in Ohio,
Oklahoma, and Nebraska. Each study
location measured water quality from
grazed pastureland, but each had a
different emphasis depending upon the
staff and facilities available. This study
differed most distinctly from the others by
the fact that the study investigated the
opposing summer grazing-winter runoff
combination.
The report which follows is a summary
description of the study objectives;
facilities and instrumentation used; data
organized by hydrology, sedimentation,
and 'water quality; and interpretations
and conclusions. Although much detail is
omitted, the summaries in the text and
appendices of the full report provide
sufficient information so that the reader
may review the data to substantiate the
conclusions or to develop their own. This
opportunity for reinterpretation is impor-
tant because the implication of these
findings may likely change in the future,
but the carefully documented facts of
these observations will remain perma-
nent.
Conclusions
The three study years encompassed a
variety of precipitation amounts and
event types from near drought to a near
normal snowpack. The mean precipitation
quantity for the study period was slightly
below the long-term average, but in
general, the data are quite representative
for the study region and certainly will
apply throughout a broad region of the
west where winter precipitation and
summer grazing dominate.
Surface runoff, whose quality was the
object of this study, was 199 and 136
millimeters (mm) of the 676 and 516 mm
(near normal) of precipitation for the main
watershed for the two near-normal years,
or about 20 percent of the precipitation.
The quantity from the small, upland check
watershed was only one-third of this
amount. The important feature was that
adequate surface runoff occurred from
both the main and check watersheds and
that the water quality of both areas was
well defined. Because of the difference in
size and topographic setting, significant
natural differences in runoff volume were
expected and no attempt was planned or
conducted to relate runoff quantities to
watershed treatments of livestock graz-
ing.
Erosion and streamf low sediment were
much less from these pasture lands than
from similar tilled agricultural lands.
Average sediment yields were only 382
kilograms per hectare per year (kg/ha/yr)
from the main watershed for the study
period compared with an average of
5,900 kg/ha/yr for a typical nearby tilled
watershed during a study period of 1961 -
65. But the check watershed averaged only
19 kg/ha/yr for the study period, thus it
showed much less erosion and sediment
transport than the main watershed. Part
of this reduced sediment from the check
watershed was obviously associated with
the reduced surface runoff, but beyond
that, the observations showed that cattle
trails and trampling within the main
watershed slopes became small interceptor
ditches from surface flow upslope and
eroded several centimeters (cm) in depth
and width. Several shorter trails within
the stream alluvium leading to water places
eroded even more. And the small stream
banks (most less than 25 cm) had some
deterioration due to trampling during
grazing and watering. No quantitative
assessment was made of these cattle
effects and they appeared to be not highly
significant to the overall water quality,
although this is certainly a quality aspect
that could be altered by management
techniques of controlled cattle traffic.
Total coliforms (TC) in runoff from
grazed and ungrazed areas in the Pacific
Northwest did not correlate with the
presence or absence of animals. Fecal
coliform (FC) and fecal streptococcal (FS)
numbers were elevated i n runoff from the
grazed area when cattle were present
above that when they were absent.
However, even after animals were absent
from the area for several months, FC
numbers were elevated in the runoff
above recommended levels of 200/100
ml (FC)and 1,000/100 ml (TC)for primary
contact. In fact, FC and FS numbers
appeared to increase from less than
100/100 ml to several thousand per 100
ml in the runoff from the grazed area for
several months in the spring during
warm, wet weather after animals were
removed the previous fall. Almost 3 years
of cattle absence were required for FC
numbers in runoff from the check
watershed to be consistently below the
maximum recommended numbers for
primary contact. The data indicate that
the use of conventional FC/FS ratios or
FC numbers in runoff as a measurement
of recent fecal pollution by cattle on a
grazed area is of limited or no value.
FC/FS ratios of 1, which indicate recent
animal fecal pollution, were found after
animals had been removed for several
months.
Nitrogen (N) and phosphorus (P)
deliveries from the watersheds were low-
-generally much lower than from areas
used for other agricultural purposes.
Tqtal-N losses from the grazed area were
3.8 and 3.8 kg/ha during water years
-------
1978 and 1979, respectively, while total-
P losses during the same periods were
1.29 and 0.93 kg/ha, respectively, with
very low losses of ortho-P. More N fell on
the grazed watershed in precipitation
than was lost in the runoff.
Recommendations
This research project provided hydro-
logic, erosion, and water quality infor-
mation from a specific site, with one
management level, and for minimal study
years. While these baseline data are
extremely useful for general recommen-
dations that water quality is quite good
from pastureland of moderate grazing
and management, additional research
must be conducted to determine the ef-
fect of management practices on sedi-
ment, chemical, and bacteriological wa-
ter quality parameters and to establish
bacterial water quality standards for
nonpoint sources. Only through carefully
planned and documented studies with
varying management levels and tech-
niques will it be possible to adequately
understand the effect of management op-
tions on local and downstream water
quality.
Research is desperately needed to
determine the meaning of FC and FS
numbers in runoff from nonpoint sources
such as grazed areas. This study indica-
ted that these organisms multiplied on the
pasture during warm, wet weather
several months after the animals were
removed. Voluminous literature indicates
that pathogenic microorganisms, which
the presence of these organisms indi-
cates, cannot survive for long periods
outside the warm-blooded host. While
the numbers of FC and TC in runoff
exceeded those recommended for pri-
mary contact, the situation on the water-
shed should not present a health hazard.
Fecal coliform and TC numbers in runoff
as defined for point sources do not appear
to be viable water quality criteria for
nonpoint sources.
Chemicals such as N and P in runoff
from those areas were below levels that
would usually be of environmental
concern. Grazing areas in the Pacific
Northwest, managed as the area de-
scribed in this report, should not present
chemical environmental hazards. Future
studies should concentrate on bacterio-
logical processes and interpretations for
potential health and environmental ha-
zards.
Experimental Design
A paired watershed design was select-
ed to study the water quality impact of
animal grazing. The main watershed
with an area of 21.5 ha [53.1 acres (a)]
was summer grazed with management
typical of this region. A check watershed
of 0.9 ha (2.2 a) adjacent to the main
watershed and a part of the original
pasture was fenced to exclude all cattle.
These two watersheds were intensively
instrumented for complete hydrologic,
sediment, and water quality measure-
ments, and were continuously monitored
for the period December 1976 to July
1979.
The watersheds were carefully select-
ed to be generally representative of the
partially forested grazing areas of the
Pacific Northwest. The level of recent and
current management was of particular
interest because management was not to
be varied but was to be typical (i.e. most
representative) such that neither the
worst nor best case was represented. The
areas studied met these criteria.
The pasture was grazed at a rate that
fully utilized the grass production which
produced overgrazing in dryer years but
was not beyond sustained use over a
period of years. Little management for
grazing distribution or to prevent cattle
contact with the stream was done, but
this is typical for this region of open
pasture, summer grazing. The cattle were
usually removed in late October and
returned to the pasture in late May. The
actual watershed management was
carefully recorded and monitored through
cattle numbers, forage production and
utilization, cattle habits and movements,
and dung distribution.
The small check watershed was fenced
at the outset of the study and grazing was
prohibited during the remainder of the
study other than occasional deer and a calf
that broke in for 1 day during the first
summer. No attempt was made to remove
existing manure at the study initiation,
thus the animal effects were a decaying
function from their last presence in
October 1976. To approximately maintain
the grass quantities similar between the
two watersheds and keep the presence of
cattle as the only major difference,
the check area grass was moved, baled,
and removed about midsummer each
year, after most vegetative growth had oc-
curred.
Results and Discussion
Hydrology and Sediment
Daily quantities of precipitation,
streamflow, and sediment were meas-
ured on the main and check watersheds.
Precipitation
For the three study years, precipitation
ranged from near drought during the
1976-77 winter to above normal for
1977-78, and near normal for 1978-79.
The annual values for these years are
shown in Table 1.
Precipitation during the three study
years averaged 65 mm (2.6 in) below the
Potlatch normal of 622 mm (24.5 in). The
first year was a severe regional drought
with very limited snow cover at any time
and only 370 mm (14.6 in) on the
watershed. The two subsequent years
were more nearly normal, but neither of
these years had a large snowpack
accumulation because of mid-winter
thaws. February 1979 had the largest
accumulation of snow depth (about 45
cm) and this melt provided the largest
surface runoff measured during the
study. Discussions with local ranchers
indicated that normal accumulative snow
depths would be slightly greater than any
that were measured.
Streamflow
Observed streamflow from the study
watersheds followed the precipitation
trends very closely (Table 1). The 1976-77
winter had far below expected runoff
amounts, while the two subsequent
study years had more normal amounts.
The study period provided a wide variety
of flow events in different sequences
which make the average results appli-
cable to other situations.
Observations drawn from these events,
daily, monthly, and annual summaries
indicate that the check watershed had
significantly less runoff than the main
watershed. The one exception was the
frozen ground runoff events in 1976-77.
This is not unexpected because the check
watershed was situated on an upland
area with less steep topography and no
wet bottomland. These differences are
always a problem in paired watershed
studies, and especially so when the
topography, geologic setting, and total
areas differ considerably. The surface
runoff of the check area was only 40 and
30 percent of the main watershed for the
1977-78 and 1978-79 seasons. Little or
no runoff was observed from its south
side except in frozen conditions. The
effect of these differences on the water
quality interpretations is likely not so
severe because the check area did have
sufficient runoff these.2 years to provide
significant surface flushing and numer-
ous sampling opportunities to contrast
simultaneously with those from the main
watershed.
-------
Tablet. Annual Streamflow Summary
1977*
Main Check
1978 1979+
Main Check Main Check
Precipitation, mm
370
676
516
Runoff
Volume, m3
Depth, mm
Total Flow
Surface Flow
Peak flow. I/ sec
2.451
11
10
26
308
34
34
4
38.193
178
119
1O3
422
46
46
4
40.623
189
136
145
370
41
41
3
"Measurements Nov. 4. 1976 to Sept. 30, 1977.
^Measurements Oct. 1, 1978, to July 5. 1979.
Sediment
Sediment concentrations and total
quantities were quite low for both
watersheds when compared with any
other agriculturally used areas. The
average concentrations for all 3 years
from the main watershed was 424
milligrams/liter (mg/l) (Table 2), with
the maximum sample having 12,700
mg/l and minimum sample having 1
mg/l (during baseflow). Average sedi-
ment yield was 382 kg/ha. These values
compare with mean concentrations and
yield of 7,400 mg/l and 5,900 kg/ha,
respectively, from wheat land in our
region (Missouri Flat Creek, 70 km2,
1961-1965). Thus, the grassed pasture
area is more than a full magnitude less
than that delivered from a much larger
farmed watershed.
The check watershed had significantly
less sediment discharge of only 19 kg/ha
corresponding with the much less total
streamflow, but the mean concentrations
were also only 10 to 20 percent of those
from the main watershed. The maximum
sampled concentration was 2,105 mg/1.
The much lower concentrations and
yields may partially be influenced by the
topographic setting, but the observations
indicate that much of this difference is
attributed to the presence of cattle on the
main watershed. Many cattle trails were
developed throughout the main water-
shed, and they were especially concen-
Table 2. Annual Sediment Summary
trated in the spring and channel area
where watering occurred. Several trails
in this mid-watershed area intercepted the
surface runoff and became the controlling
channels. They eroded significantly (5 to
10 cm in depth), and this sediment went
directly into the small streams.
The stream immediately above the
main gaging location had small (15 to 30
cm) banks that were largely unprotected by
vegetation. Through watering and grazing,
the cattle added some disturbance to this
area which added to the natural tendency
for this small channel to erode. The
upland channel directly west of the
weather station also had some exposed
banks and a small overfall which eroded
some during the study. Again, there was
some cattle activity in this channel
which undoubtedly aggravated the ero-
sion. Beyond the trails and channels,
there was little visible erosion. These
upland grass areas probably produced
amounts similar to those measured on
the check watershed where there were
no apparent areas exposed from good
grass vegetation.
Manure Distribution
To provide direct measurements of the
manure distribution on the main water-
shed, counts were made three different
dates during 1978 at early summer,
midsummer, and just after cattle were
removed. The watershed was stratified
7977* 1978 /979f 3-yr avg
Main Check Main Check Main Check Main Check
Precipitation, mm 370
Surface Runoff, mm 10
676
516
34
119
46
136
41
517
88
40
Sediment
Yield, kg
Yield, kg/ha
Avg. cone., mg/l
877
41
408
15
16
49
10,489
488
410
25
28
60
13.291
618
455
12
13
32
8.219
382
424
17
19
47
"Measurements Nov. 4, 1976 to Sept. 30, 1977.
^Measurements Oct. 1, 1978 to July 5. 1979.
into three sampling zones of (a) high
cattle use due to loafing or watering, (b)
along drainages, and (c) general grazing.
Eight sample sites were made in zone (a),
10 along the waterways, and 20 randomly
placed throughout the watershed. Each
sample consisted of counting the number
of manure drops in a 50-m2 area and
sampling the weight and size of the
droppings. The high use areas were
difficult to count due to trampling of the
droppings.
The fecal deposits were not evenly
distributed over the watershed, but were
concentrated in high-use areas and, to
some extent, along drainages. The high
drainage density was probably due to the
location of some animal trails in the
pasture. The situation is compounded
when it is considered that two of the high-
use sites were in or immediately adjacent
to drainage bottoms.
On an overall basis, the density of fecal
deposits across the entire watershed was
2,645 deposits/ha, providing a coverage
of 1.45 percent. Therefore, if the distribu-
tion of droppings were more uniform, the
concentrations in the drainages would
have approached that which was present
elsewhere in the nonconcentrated areas.
In the opinion of the researchers, one
must suspect that the fecal bacteria
counts in the runoff are a result of the
distribution of feces much more than the
overall fecal density. This is one subject
which should be seriously considered for
additional research.
Water Chemicals and Oxygen
During the drouth water-year (1977), N
and P losses in runoff from the main and
check watersheds were insignificant. In
water-year 1978, total N loss was only
3.8 and 0.48 kg/ha from the main and
check watersheds, respectively, which
was more than during 1977 but was still
low. Nitrate concentrations were almost
negligible with a total loss of only 0.33
and .08 kg/ha NO-j-N, respectively. N
parameters were higher from the main
watershed than the check, but no param-
eters were of sufficient magnitude to be
of environmental concern. N concentra-
tions were high in runoff when the pond
was drained and in the initial runoff sam-
ples in the fall after the summer grazing
season. These higher concentrations
probably resulted from relatively fresh
manure particles being carried in the run-
off from the pasture and within the main
stream channel. The pond was fenced to
exclude cattle but they did have access to
the main channel below the pond to the
sampling site. Throughout this study, the
periods of greatest nutrient concentra-
-------
tions were associated with small runoff
volumes such as early season events,
summer events, or pond drainage. Thus,
these events contributed very little to the
annual constituent discharge. Average
concentrations of N in the baseflow was
about one-half that of the surface runoff.
While NOs-N losses ip runoff from the
main watershed were slightly greater in
water-year 1979 than in 1978 (1.33
kg/ha vs. 0.33 kg/ha), most N losses in
runoff were less in 1979, thus total N
losses were similar. Again, N losses in
runoff were low and were generally
lower than would be expected in runoff
from areas used for other agricultural
purposes. Total N losses were greater
from the main watershed than the check
watershed. Exchangeable sediment NH/w
was not run during water-year 1979
because it was negligible during water-
year 1978.
P deliveries from both watersheds
were negligible during water-year 1977.
P deliveries in runoff were very low for
water-year 1978 with runoff from the
grazed area delivering only 0.2 kg/ha/yr
ortho-P with a total-P loss of only 1.27
kg/ha/yr. P levels in the runoff did
increase when cattle were present
during the July 4 event. Very small
amounts of P were delivered by baseflow
even though baseflow was about one-
third of the total water lost during the
1978 water year.
In water-year 1979, P deliveries in
runoff and baseflow were low from the
main and check watersheds. Generally,
average ortho-P concentrations in runoff
from the main watershed were higher in
water-year 1979 than in 1978; in both
years ortho-P in runoff decreased as the
length of time animals were off the
pasture increased. Total-P lost in runoff
from the grazed watershed was slightly
greater in water-year 1978 (1.27 kg/ha)
than in water-year 1979 (0.93 kg/ha). If
baseflow was added, the values were
1.39 and 1.05 kg/ha, respectively. The
greater total-P loss in water-year 1978
was likely due to the December 10-15,
1977, event that had a high average
total-P concentration of 3.12 mg/l.
Ortho-P and total-P losses from the check
watershed during water-year 1979 were
0.02 and 0.07 kg/ha, respectively.
The oxygen demand, pH, SC, Cl~, and
cation delivery for the main and check
watersheds for water-years 1977, 1978,
and 1979 are summarized in the Appen-
dix of the report. During the establish-
ment year, water-year 1977, oxygen de-
mand for the main watershed was about
twice that of the check (probably due to
stagnation); pH for the main watershed
Table 3. Water Quality Parameter Concentration Responses to Streamflow
Parameter
Bacteria, oxygen
demands, N and P
nutrients
Type of Runoff
Event
Rainfall
Time of Year
Fall and
spring
Response to Streamflow
Direct response to
Streamflow changes with
successively smaller •
peak concentration
It
Specific conduc-
tance, pH, Cl,
Na. K. Ca, Mg
Snowmelt or
frozen ground
rainfall
Mixed (snowmelt
followed by
rainfall)
Snowmelt or
rainfall
Snowmelt or
rainfall
Winter
Fall, winter,
and spring
Late winter
and/or
spring
Fall, winter.
and spring
Little or inverse
response to Streamflow
Little response during
snowmelt. Direct
response to rainfall
Little response to
Streamflow changes
Inverse response to
Streamflow
appeared slightly more alkaline than the
check, the SC of the main watershed was
somewhat greater than the check water-
shed, as was Cl* and cation delivery. The
oxygen demand of the runoff from the
grazed watershed was about three times
that of the ungrazed watershed during
water-year 1978 while Cl~ cations were
about five times greater. Values for these
parameters increased in runoff from the
main watershed when cattle were
present (July 4 event). Oxygen demand of
runoff in water-year 1979 was greater
from the grazed area than in 1978, but
runoff from the ungrazed area was
similar. Cations were not run during
water-year 1979 because data from the
previous year had shown that SC was
well correlated with the Na, K, Ca, and Mg
present in the runoff.
Indicator Bacteria
TC counts in runoff showed little
relationship to the presence or absence
of cattle as shown by the summarized
results of the main and check watershed.
TC counts probably responded more to
climate than to the cattle. From results of
this study, TC counts appear to be useless
for determining stream pollution.
In general, the FC and FS numbers
decreased as expected throughout the
winter months. Their numbers, however,
unexpectedly increased again in the
spring with warmer weather. Also, the
FC/FS ratios varied more than expected,
ranging from as high as 4/1 on the check
watershed to 1/1 on the main watershed.
FC/FS ratios of about 0.1 to 1.0
indicate cattle fecal contamination.
Runoff contaminated with cattle normally
has a FC/FS ratio of about 0.7 while that
with human had a FC/FS ratio of 4.0 and
higher. Animal FC/FS ratios will increase
with time, but they should remain less
than one. The high FC/FS ratios long
after cattle were removed from the area
would lead one to question the validity of
the FC/FS ratio for nonpoint areas. Fur-
thermore, when animals were introduced
back onto the watershed, the June 4,
1977, readings showed high FC and FS
numbers but a FC/FS ratio of 0.02, which
is indicative of wild animal fecal con-
tamination. During the 3 year study peri-
od, except for a few mice and squirrels,
wildlife populations on the study areas
were almost nil.
The effect of grazing cattle on runoff
water quality from the study watersheds
can be summarized by the following. TC
in runoff had an apparent relationship to
cattle grazing operations, but numbers
were elevated when they should not have
been. FC and FS numbers in runoff from
the grazed watershed were elevated
when cattle were on the pasture. After
cattle were removed, FC and FS numbers
generally declined throughout the fall
and winter months. However, after
several months absence of cattle from
the area, FC and FS numbers in runoff
from the grazed watershed were elevated
and in many cases exceeded suggested
wdfter quality standards. FC and FS
numbers appeared to increase in runoff
from the grazed area in the spring in
rainfall runoff following a period of warm,
dry weather several months after animals
were removed. FC/FS ratios in runoff
varied considerably and indicated recent
cattle fecal pollution for long periods of
time after animals were off the grazed
area. FS numbers in runoff from the
grazed and check areas did not change
appreciably during the study. The check
-------
Table 4. Average Percent Fecal Coverage in the Drainage Sampled Within the Study
Watershed, Nov. 4, 1978
Sampling
Site No. Drainage Runoff Areas
Fecal Coverage
Percent Area
1 Sacrifice (lower) pasture (plus rest
of grazed main watershed)
2 Ungrazed (check) watershed
11 Southeast area just above mam sampling station
12 Northeast area just above main sampling station
13 All of the grazed watershed except the
southeast, northeast and sacrifice areas
14 North pasture area draining into large gulley
channel
15 Northwest pasture area
16 Southwest pasture area
17 Woodland area
18 Logged area (outside of fenced cattle-grazed
area; horses, however, were grazed here)
20*
0
NS§
0.7
2.6
0.9
3.1
2.2
1.8
NS
'Total remainder of watershed averaged 1.4 percent.
§/VS =Not sampled
was ungrazed from the fall of 1976
throughout the study. FC numbers in run-
off from the check were not consistently
below the maximum recommended
numbers for primary contact until the
1979 water-year, although FC numbers
in the check watershed runoff during
water-year 1978 were genera I ly below or
close to the recommended primary con-
tact standard. The data would suggest
that using conventional FC/FS ratios or
FC numbers in runoff as a measurement
of recent fecal pollution by cattle on a
grazed area is of limited value in the Pacif-
ic Northwest.
Parameter Response and
Spatial Distribution
Parameter Response
The type of runoff event affected the
parameter response to streamflow (Table
3). With snowmelt on thawed ground
followed by rain, NOa-N does not follow
the hydrograph initially during snowmelt;
however, when the ground thawed, NOs
concentration increased. Ortho-P did not
follow the hydrograph at any time during
a snowfall-rain on frozen soil event. For
the same event, COD also did not appear
to respond to streamflow until the ground
began to thaw on the second day of the
event. The water quality parameter, in
this case total-P, did not respond to the
streamflow increase caused by snow-
melt, but did respond to the later stream-
flow increase caused by rainfall runoff.
During long and large events, the check
watershed behaved similarly. The high-
est check watershed values often oc-
curred in the first few runoff samples as
the check watershed channel was
flushed since it had no base flow. For
small events, parameter concentrations
for check runoff, therefore, did not corre-
late well with flow rates.
Spatial Distribution
Hand samples were collected from
various locations within the watershed
during events in the 1978 and 1979
water-years to determine the effect of
grazing patterns, manure density and
land cover on water quality parameter
concentrations. Of specific interest was
whether the pollution load was primarily
a result of cattle activity in and im-
mediately adjacent to the stream channel
(the "sacrifice area" between the north
and south pastures) or if the stream water
quality reflected the grazing activity over
the entire watershed.
During the 1978 water-year, samples
were taken December 6, 1977, during a
small snowmelt/rainfall event; December
13, 1977, during the largest event of the
water-year; and on March 15,1977, dur-
ing a late winter medium sized rainfall
event. In water-year 1979 samples were
taken from an event caused primarily by
snowmelt on February 13, 1979, during
the largest event of the water-year and of
the entire study; February 27, 1979, dur-
ing a large rainfall/snowmelt event; April
6, 1979, during an early spring rainfall e-
vent; and on May 4, 1979, at the be-
ginning of a large spring rainfall event.
This last sampling occurred after indica-
tor bacterial numbers had increased
above their late winter values.
Fecal density and distribution data
were collected during the study by the
range management component of the
project (see Section VI of the Range
Management Section in the report). The
amount of a drainage area within the
pastured watershed covered by feces
varied from 20 percent for site 1 (sacrifice
pasture) to 0.7 percent for site 12
(northeast area) at the November 7,
1978, sampling date (Table 4). The
percent manure cover was determined
three times during the summer of 1978.
The values obtained on Nov. 7, 1978,
after the cattle were removed from the
watershed represent the fecal densities
present when these water quality sam-
ples were taken from the 1979 water-
year runoff. FC and FS numbers in runoff
from the various sampling sites within
the watershed did not correlate clearly
with fecal disposition in water-year 1978
or 1979. In fact, no trends are apparent
from the sites sampled. Also, N and P in
the runoff from the various sampling sites
within the pastured watershed did not ap-
pear to agree with manure deposition.
Presumably vegetation and soil cover had
a greater effect on runoff nutrient content
than did manure deposition.
These results would indicate that
indicator bacteria and nutrients in runoff
from the main pasture areas were
generally equal to those at the main
sampling station. The pollution load can
be considered nonpoint source in origin
and not primarily a result of the greater
grazing activity in the sacrifice area.
6
-------
Keith E. Saxton, Lloyd F. Elliott. Robert I. Papendick, and Michael D. Jawson are
with Washington State University. Pullman, WA 99164; DavidH. Fortier is with
USDI. Bureau of Land Management, Coeur D'Alene, ID 83814.
R. Douglas Kreis is the EPA Project Officer (see below).
The complete report, entitled "Effect of Animal Grazing on Water Quality of
Nonpoint Runoff in the Pacific Northwest," (Order No. PB 83-245 225; Cost:
$14.50. subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Robert S. Kerr Environmental Research Laboratory
U.S. Environmental Protection Agency
P.O. Box 1198
Ada. OK 74820
-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use S300
------- |