United States
Environmental Protection
Agency
Corvallis Environmental
Research Laboratory
Corvallis, Oregon 97330
GUIDELINES FOR ASSESSING THE
BENEFITS OF BEST MANAGEMENT PRACTICES
TO STREAM ECOSYSTEMS
CERL - 039
December 1977
-------
GUIDELINES FOR ASSESSING THE
BENEFITS OF BEST MANAGEMENT PRACTICES
TO STREAM ECOSYSTEMS
CERL - 039
December 1977
-------
GUIDELINES FOR ASSESSING THE BENEFITS OF
BEST MANAGEMENT PRACTICES TO STREAM ECOSYSTEMS
by
Jack H. Gakstatter
Thomas E. Maloney
and
Frederick B. Lotspeich
Corvallis Environmental Research Laboratory
U.S. Environmental Protection Agency
Corvallis, Oregon
-------
GUIDELINES FOR ASSESSING THE BENEFITS OF BEST MANAGEMENT
PRACTICES TO STREAM ECOSYSTEMS
Introduction
The Environmental Protection Agency recognizes the biological responses
of water bodies to best management practices (BMP) as an important measure of
their effectiveness and enhancement of stream or lake health and utility.
This is especially true where BMP's focus on control of pollutants from
nonpoint sources (NFS).
Testing, evaluation, and development of ecological assessment methods
for linking pollution controls with benefits to man's use of water is one of
the four objectives of the nonpoint source ecological effects research
program begun in FY 77 at the Con/all is Environmental Research Laboratory
(CERL). The impetus for release of this document at this time is the initia-
tion of the interagency Model Implementation Program (MIP) and the many
requests from the Section 208 sector. The guidelines presented here should
be considered state-of-the-art. As research outputs develop, these guidelines
will be updated and improved.
Recently an interagency agreement was established between the U.S. En-
vironmental Protection Agency and the U.S. Department of Agriculture for the
common interest of developing Model Implementation Programs (MIP) for water
quality management. This effort will involve selecting 3-5 geographic areas
in the United States, implementing best management practices (BMP) in those
areas, and evaluating their effectiveness. The evaluations will include
1
-------
measuring decreases in nonpoint source (NPS) pollutants delivered in runoff
to surface waters and subsequent changes in surface water quality. CERL's
involvement will carry the evaluation one step further and examine the re-
sponse of the biological stream community to reduced levels of NPS
pollutants. The biological evaluations will be limited to one or two MIP,
areas.
Objectives
The purpose of the approach outlined herein is to evaluate the effect of
BMP's by examining biological changes in stream systems and relating these
changes to the observed reduction of NPS pollutants.
The objective of establishing guidelines at this time serves several
purposes. First these guidelines fix the level of effort that must be dedi-
cated to a project if realistic and usable data are to be obtained. Second,
they serve as partial criteria in selecting or determining the adequacy of
the sites and third, they identify the approach that must be followed in
attempting to interpret the effects of nonsteady state, nonpoint source
inputs on stream communities.
These guidelines will also serve as the basis for evaluating extramural
research proposals which may be solicited at some future time to support the
MIP Program.
Approach
It is assumed that the major pollutants to be controlled via BMP's are
sediment, phosphorus, nitrogen and/or degradable organics. Pesticides,
herbicides, and trace metals may also be of interest depending upon the
nature of the study area(s). If the latter pollutants are of specific con-
cern, the scope of the basic biological studies would not change, but the
-------
study plan would be modified to include measures of bioaccumulation which are
relevant to human health and the well-being of the aquatic ecosystem.
There are at least three approaches or experimental designs which could
be used to evaluate the impact of BMP's on stream ecosystems. The design of
choice for a given MIP area will depend on a number of factors including:
(1) size of the MIP area and the number of watersheds amenable to
study therein,
(2) availability of a non-contributing* watershed(s) within or near
the MIP area to serve as a control,
(3) spatial orientation of the non-contributing watershed to the con-
tributing watershed, if both are located in the same drainage
basin,
(4) availability of adequate physical, chemical and biological
baseline data,
(5) time available to gather baseline data if they do not exist.
The first approach or design involves a minimum of three study areas,
but preferably four or five, within a MIP area and is called the "gradient"
approach. The term gradient is used because BMP's will be implemented on
varying percentages of the total area of each contributing watershed. An
example of this design, using five watershed/stream (W/S) systems, is illus-
trated in Table 1. Each of the five watersheds (and associated stream rea-
ches) must be geomorphologically similar and the land uses in the four
contributing watersheds must be the same.
* a non-contributing watershed is covered by a land use which generated mini-
mal quantities of NPS constituents; e.g., forest or ungrazed grassland. A
contributing watershed generates relatively high levels of NPS constituents;
e.g., row crop agriculture or construction activities.
-------
TABLE 1. Gradient Analysis Design
% of watershed area
Watershed/stream covered by a contributing % of contributing land
identification land use ._ use area treated by BMP's
A 100 0
B 100 33
C 100 66
D 100 100
E 0 (non-contributing) 0
The sample design, illustrated in Table 1, can be modified as necessary
to fit a specific area; i.e., the percent of the contributing land use area
treated by BMP's does not have to be exactly as shown. The simplest but
least desirable form of this design includes only three W/S systems: W/S
systems A and D, as shown in Table 1, plus a third W/S system in which 50% of
the contributing land use is treated by BMP's.
The inclusion of W/S E (a non-contributing watershed) is desirable as an
indication of the best stream quality achievable through BMP implementation
in a specific geomorphologic setting. In reality, it may be impossible to
find a suitable non-contributing watershed in an area of intensive land use.
If W/S E is not included in the study, the relative stream quality improve-
ments attributable to BMP implementation can still be determined.
A modified gradient approach can be used to rank the effectiveness of
different BMP's as well as the effect of different intensities of a given
BMP.
The advantages of the design outlined above are that, (1) a minimal
amount of pre-BMP data are required, (2) W/S A (100% covered by a contri-
buting land use; no BMP's) serves as a control for the between year variation
which will occur in stream discharges, pollutant loads, concentrations and
biological responses resulting from climate differences, and (3) the gradient
-------
approach will demonstrate the relationship between the intensity of BMP
implementation in the watersheds and the impact on the stream ecosystem as
illustrated in Figure 1.
The "biological index," represented by the ordinate in Figure 1, may
reflect aquatic species diversity, biomass, hatchability and survival of fish
eggs or a combination of factors indicative of the biological "health" of a
stream system. Curves A, B, and C in Figure 1 are hypothetical examples of
the biological response to different intensities of BMP implementation. A
response corresponding to curve B indicates improvement in biological quality
proportional to the percent of total watershed influenced by BMP's. Curve A
indicates substantial biological improvement with BMP implementation over the
initial 50% of a watershed's area with a diminished rate of improvement
thereafter. Curve C indicates that significant biological improvement only
occurs when runoff from most of a watershed is controlled by BMP's. Although
Figure 1 is a hypothetical example, data of this type obtained from a MIP's
project would provide extremely useful guidance in the future application of
BMP's.
The disadvantages of the gradient approach are that (1) it may be
difficult to find a set of geomorphologically similar systems which fit the
design, and (2) an extensive sampling and data analysis program will be
required.
The second design involves paired watershed/stream systems both similar
in geomorphologic and land use characteristics. Best management practices
would be implemented in one of the watersheds but not the other. Both water-
sheds should be heavily impacted (i.e. nearly 100% coverage) by the selected
-------
(good) 10
(bad)
FIGURE 1
25 50 75
Percent of Contributing Watershed Influenced by BMP's
An example of possible relationships between the biological quality of streams
and the percent of watershed area influenced by BMP's.
100
-------
land use. A desired variation to this design would include a third water-
shed/stream system geomorphologicaly similar to the first two but unimpacted
by cultural activities, (i.e. forested or ungrazed grass land). Advantages
to this design are, (1) that the relative effects of BMP implementation could
be measured without lengthy baseline studies, and (2) the sampling program
would be smaller than that required for the gradient approach. The major
disadvantage is that the approach is "all or nothing" and no information will
be obtained about the effects of BMP implementation over part of a watershed.
The third approach involves an intensive study of one stream draining a
watershed heavily impacted by NPS activities. The nature of the biological
community and associated water quality would have to be established in its
i
"stressed" condition prior to the implementation of BMP's. After BMP imple-
mentation, the same studies would be conducted at the same sites to determine
the extent of water quality improvement and the associated changes in the
biological community. Advantages of this approach are that, (1) a watershed
of this kind would be relatively easy to locate, and (2) the sampling program
would not be as large as that required for the gradient approach. The great-
est disadvantage is that the natural year to year variations caused by
climatic differences may mask any benefits achieved by BMP implementation
unless the magnitude of the variation is well defined by collecting (or
having in hand) several years of baseline data.
Study Duration
If the gradient analysis design is used, a minimum of one year of base-
line data will be required to demonstrate that the biological communities in
the stream draining each of the "contributing" watersheds are, in fact, the
same. Following the baseline study period and BMP implementation, at least
-------
three years of study will be needed to demonstrate the impact of BMP's on
stream biota.
The second (paired watershed/stream systems) approach should also
include one year of baseline data to demonstrate that the physical, chemical
and biological features of the two stream systems are similar. Following BMP
implementation in one of the watersheds, a three year study period would be
required.
The third approach, using a single watershed, requires the longest base-
line study of the three options. Three to five years of baseline data would
have to be collected to establish the variability of the physical, chemical
and biological characteristics of the stream system before BMP implementation
followed by three years of post-BMP study to demonstrate the change.
Measurements
To evaluate the impact of BMP's a number of in-stream physical, chemical
and biological measurements will be necessary. It is assumed that the
majority of the physical and chemical measurements will be required as part
of the MIP whether or not the biological studies are included. These physi-
cal-chemical measurements will be needed to evaluate the water quality change
related to BMP implementation. However, they also will be useful in explain-
ing observed changes in the biological community.
The following measurements will be needed in addition to the biological
data, but should be provided as part of the routine MIP evaluation.
1. Stream discharge rate:
A continuous record of the stream discharge rates is required at
the downsteam end of each study reach.*
*Dependinguponthelength of the study reach, additional sites for
measuring stream discharges as well as other physical and chemical parameters
may be desired. Q
-------
Rationale: discharge wil.l be needed to estimate the total quantity
of various constituents of interest to quantify the physical-chemical
effects of BMP's. The pattern of runoff events, particularly the first
"flush" after an extended dry period, will be of interest regarding its
impact on stream water quality and the biological community. The fre-
quency, duration, and intensity of runoff events will be documented by
the continuous discharge record.
2. Total sediment load/suspended sediments:
Measurement of the total sediment delivered via the stream from
each study reach is required as a quantitative measure of the effec-
tiveness of the BMP's. From a biological standpoint, concentration,
duration, frequency, and composition (organic versus inert inorganic) of
suspended solids are significant as well as that portion of the total
sediment that is deposited on the stream bed following a runoff event
and which constitutes a net increase in stream bed sediment. Measure-
ments should therefore include suspended solids, differentiated into
fixed and volatile fractions, during a range of flow conditions including
intensive measurement during storm events following extended dry periods.
Rationale: suspended solids delivered to streams in surface run-
off, particularly after extended dry periods, may contain substantial
amounts of biodegradable organic material which exerts a significant
dissolved oxygen demand. The reduced dissolved oxygen concentrations,
-------
in combination with additional stresses of high suspended solid concen-
trations and increased stream velocities, may have an adverse biological
effect. Inert suspended matter <3 mm in diameter, delivered via surface
runoff and deposited on the stream bed, produces a homogeneous substrate
which eliminates the niches occupied by many important stream organisms.
Substrates with a range of grain sizes (including fine particles), are most
desirable for a healthy stream community.
3. Particle size distribution and composition of bed materials:
Particle size distribution and composition of stream bed materials
should be determined at approximately the same times and locations as
for benthic invertebrate sampling and salmonid egg survival tests if
used.
Rationale: see item 2.
4. Turbidity:
Turbidity measurements should be made daily at the downstream end
of each study reach with provisions for continuous or short interval
measurements during and immediately following runoff events.
Rationale: turbidity limits light penetration through water.
Light, in combination with other factors such as nutrient availability,
determines the quality and quantity of primary production; e.g., periphy-
ton growth may be the most important part of the base of the food web,
hence, productivity of all trophic levels is affected. High turbidities
for extended periods also give a competitive edge to nonselectively
feeding fish, which are frequently undesirable species such'as carp, and
discriminate against most of the game fish which are sight feeding
predators.
10
-------
5. Nutrients:
Total phosphorus, dissolved orthophosphorus, total organic nitro-
gen, ammonia-nitrogen and nitrate-nitrogen should be measured at the
lower end of each study reach with sufficient frequency to quantify the
impact of the BMP's. The median and range of concentrations of these
nutrients during the growing season are, from a biological standpoint,
most meaningful. It is recommended that samples for nutrient analysis
be collected at least every two weeks during stable flow conditions and
that sampling times be varied sufficiently to include a range of flow
conditions. Short interval sampling (hourly or even more frequently in
some cases) should be conducted during heavy runoff, particularly after
extended dry periods, to characterize the extremes resulting from NPS
inputs.
Rationale: nutrient levels (especially phosphorus) are important
factors in controlling the amount of primary production which occurs in
streams and lakes. While a limited amount of primary production is
desirable, higher concentrations of nutrients may cause excessive and
undesirable quantities of both microscopic and macroscopic plant growth.
A criterion of 100 pg/1 of total phosphorus has been recommended (U.S.
E.P.A., 1976) for flowing streams although there is little scientific
basis for this level. Other recommended total phosphorus criteria
include 50 p.g/1 where streams enter lakes or reserviors and 25 pg/1
within lakes or reserviors.
The nitrogen constituents in water are important for three reasons:
(1) inorganic nitrogen serves as a plant nutrient and in some situations
be the limiting factor for plant growth, (2) ammonia (free, un-ionized)
11
-------
is toxic to fish and may be present in significant amounts during heavy
runoff events and (3) nitrite-nitrates are significant if the stream is
used as a water supply in that concentrations above 10 mg/1 ( as N03-N)
may cause methemoglobinemia in infants.
6. Temperature and dissolved oxygen:
These parameters should be continuously monitored at the downstream
end of each study reach.
Rationale: water temperature determines the metabolic rate of a
stream community and, to a significant degree, determines which species
of aquatic life are present in a stream. Water temperatures could be
influenced by certain kinds of BMP's such as establishment of greenbelts.
Dissolved oxygen (DO) concentrations are critically important to most
stream biota. Most water quality standards specify that at least 5 mg/1
be maintained in trout streams and 4 mg/1 in streams supporting warm-
water fisheries. Dissolved oxygen concentration in a stream at any time
reflects the balance between photosynthesis, respiration and the physical
reaeration which is occurring. Dissolved oxygen concentrations during
heavy runoff periods will be particularly indicative of the quantity of
decomposable organic matter entering a stream from nonpoint sources.
7. Total organic carbon, BOD, COD:
A measure of the degradable organics delivered to a stream, particu-
larly during runoff events, is desired. Although the BOD test is the
only direct measure of the decomposable fraction of the total organics,
the idiosyncrasies of the test are many. For that reason, the measure
of total organic carbon or chemical oxygen demand may be desired as a
substitute for the BOD test with the results used in conjunction
12
-------
with dissolved oxygen data to assess changes in the levels of degradable
organic material delivered to the streams.
Rationale: see item 6 rationale.
8. Other Water Quality Parameters:
Other water quality data such as pH, alkalinity, hardness, conduc-
tivity, trace metals and mineral constituents should be determined
periodically (monthly to quarterly) to characterize the general water
t
quality of the stream. It may be desirable to sample at frequent inter-
vals, at least during one heavy runoff event, to document parameter
changes during a period of rapidly changing stream flows.
Rationale: the general water quality of any stream in which
biological life is studied should be documented because, to a degree,
water chemistry influences the kinds of organisms which are present.
9. Climatological data:
A weather station should be located centrally in each drainage area
to record precipitation intensity and amount, air temperature and light
intensity.
Rationale: climatic factors are the driving force behind NPS
pollution and the functioning of stream ecosystems.
The following factors should be included where biological evaluations of
stream systems are made.
1. Stream Morphometry and Habitat Survey:
The study reach of each stream should be mapped to scale showing
such features as width, depth, composition of bottom substrate (sand,
silt, cobbles, bedrock, etc), riffle and pool areas, sinuosity, stream
13
-------
bed gradient, nature and stability of the stream bank, and riparian
vegetation. The suitability of a stream reach as a habitat to support
an adequate fish population should also be assessed. The habitat
assessment should be made prior to the selection of a stream as a bio-
logical evaluation site. If suitable habitat for fish and other aquatic
organisms is not available; e.g., if the stream has been straightened,
channelized, and riparian vegetation removed, there will be little value
in making a biological evaluation unless the BMP's include restoration
«
of the stream habitat.
Rationale: the scale maps, indicating pertinent physical features,
will be necessary for making estimates of the standing crop of benthic
organisms and fish. The maps will also be of value in comparing the
similarities and dissimilarities between study areas.
The habitat survey will demonstrate whether a stream has potential
to be biologically productive. Unless suitable habitat is present,
biological productivity will be minimal regardless of water quality.
2. Fish:
Fish populations of each study reach should be sampled in a quan-
titative manner at least annually and should include data on kinds,
numbers, lengths and weights, and total biomass of fish present. The
sampling procedure should be nondestructive, i.e. the collected fish
should be kept in live boxes after capture and returned to the stream in
good condition after the appropriate data have been recorded. The
recommended sampling method is electrofishing in stream sections blocked
at each end by nets. Season of the year to sample should be determined
by the life history of the species inhabiting the stream, although, in
14
-------
general, it probably should be done during late summer. It is important
that sampling be conducted at approximately the same time each year so
that the between year data are comparable.
If the stream of interest supports an anadromous salmonid fishery,
additional studies should be conducted on the hatching success of "eyed
eggs" (Everest, 1975). "Eyed" salmonid eggs can be placed in small wire
baskets and 'buried in spawning gravels at depths simulating natural
spawning conditions. Egg development and hatchability is subsequently
monitored. A less quantitative, but still informative procedure, would
be the use of a nylon fry trap described by Phillips and Koski (1969).
This trap consists of a nylon cap placed over a natural redd in a
stream. Emergence of the fry is then monitored during the incubation
and hatching period. While the number of eggs which had been deposited
within each redd are not known, average numbers of fry emerging from a
number of redds could be compared from year to year and between control
•and NPS impacted stream reaches.
Rationale: the fish population of a stream, including the species
present and their abundance, is perhaps the primary concern of the
general public interested in using that stream for recreational
purposes. Density and composition of the fish population is also the
best measure of the health of a stream community, since fish represent
the top of the aquatic food web and therefore reflect the well-being of
the components at lower trophic levels.
3. Benthic invertebrates:
Using Surber samplers, Ponar dredges, or comparable devices which
sample a fixed area of the stream bed, benthic organisms should be
15
-------
sampled at least seasonally in each study reach although more frequent
sampling, particularly when macroinvertebrate biomass "peaks" in the
early spring, would be desirable. Invertebrate drift, using submerged
nets with openings of known size, should also be measured at night at
the same approximate locations as the bottom sampling stations.
Sampling and analysis should be of sufficient intensity and sensitivity
to estimate standing crop, species diversity and similar measures of the
health of the benthic community before and after the implementation of
BMP's in the drainage area affecting the selected stream reaches.
Sampling should be conducted at approximately the same time each year so
that between year comparisons can be made.
Rationale: benthic invertebrates, particularly the insects, are a
major food source for fish in stream ecosystems. Total numbers, number
of taxa, and biomass of benthic invertebrates reflect water quality as
well as the physical characteristics of a stream. A diverse and
abundant population of benthic invertebrates implies a suitable habitat
and water quality for a healthy fish population.
4. Primary productivity:
a. Periphyton - using submerged glass slides or a comparable tech-
nique the rate of production and estimates of total biomass of
attached algae should be determined throughout the study reach
during the growing season. The glass slide technique may be
supplemented by scraping and analyzing the periphyton from natural
substrates in the stream.
b. Phytoplankton - the algal assay bottle test should be used to
evaluate, (1) the potential productivity of the stream water, (2)
16
-------
the growth limiting nutrient, and (3) the presence of any growth
inhibiting substances that may occur in the stream as the result of
NPS runoff. Since these characteristics will probably vary with
stream flow, it is recommended that assays be performed on samples
collected over a range of stream flows including heavy runoff
events.
Rationale: in intermediate size streams and unshaded smaller streams,
periphyton productivity represents a large portion of the food web base
and is influenced by the availability of nutrients and light (turbidity)
both of which are NPS related. One would expect to be able to relate
periphyton productivity not only to changes in nutrient and turbidity
levels, as they might be changed by BMP's, but also the standing crops
of macroinvertebrates and fish.
Phytoplankton, per se, are not particularly important in smaller
stream systems. The purpose of the algal assay bottle test is primarily
as a tool to determine the limiting nutrient, to indicate potential
productivity levels should the water reach a lake or impoundment, and to
determine if toxicants are present.
5. Bacteriology:
Bacteriological indicators of fecal contamination are more closely
related to impacts on man rather than on stream ecosystems; neverthe-
less, relatively little additional effort would be required to gather
data on the impact of-BMP's on a fecal coliform levels in streams.
Collection of fecal coliform data is therefore recommended if livestock
pasturing is a predominant land use in the watersheds of interest.
Samples for fecal coliform samples should be collected over a range of
17
-------
flow conditions and at frequent intervals during significant runoff
events.
General Methodology
When applicable, standard methods of sample collection and analysis
should be used. Such methods are described in detail in Standard
Methods for the Examination of Water and Wastewater (APHA, 1975),
National Handbook of Recommended Methods for Water Data Aquisition
(USGS, 1977) and Biological Field and Laboratory Methods for Measuring
^
the Quality of Surface Water and Effluents (USEPA, 1973).
18
-------
References
American Public Health Association. 1975. Standard Methods for the Examina-
tion of Water and Wastewater. 14th Edition. Washington, D.C.
Everest, Fred H. 1975. Biological Non-Point Pollution Monitoring Guidelines
U.S. Forest Service. Corvallis, Oregon.
Phillips, R.W. and K.V. Koski. 1969. A Fry Trap Method for Estimating Sal-
monid Survival from Egg Deposition to Fry Emergence. J. Fish. Res. Bd.
Canada, 26:133-141.
U.S. Environmental Protection Agency. 1973. Biological Field and Laboratory
Methods for Measuring the Quality of Surface Waters and Effluents.
EPA-670/4-73-001. Cincinnati, Ohio.
U.S. Environmental Protection Agency. 1976. Quality Criteria for Water.
EPA-440/9-76-023. Washington, D.C.
U.S. Geological Survey. 1977. National Handbook of Recommended Methods for
Water Data Acquisition. Washington, D.C.
19
------- |