Water Quality
Ambient Water Quality Criteria for Dissolved Oxygen
3.0 Occurrence in the Environment
3.1 Natural Sources / Depletion
3.1.1 Oxygen Cycles And Distribution
The major sources of dissolved oxygen in water are the
atmosphere and photosynthesis by aquatic vegetation (macrophytes
and algae). Atmospheric oxygen dissolves in water (and passes
out of water) according to the set of physical conditions that
were discussed in Section 2. The saturation concentration of
dissolved oxygen is quickly achieved at the air-water interface
and, in shallow, moving water, will be relatively consistent
through the water column. However, since the diffusion of gasses
through water is slow, oxygenation of a larger freshwater system
with a thermocline barrier is dependent upon water circulation
moving aerated water away from the surface (e.g., by wind-induced
current, lake turnover and inflows). Gross (1972) explained
that the movement of oxygen into marine waters is controlled
by some of the same mechanisms, including stratification and
biological activity. Saltwater also is oxygenated primarily
at the surface. Extensive wind mixing and de-stratification
occur in the high latitudes, and the cold, dense waters sink
and move into the temperate zones through bottom currents,
where they upwell along coasts. In areas of restricted circulation
away from these bottom currents, such as in many of British
Columbia's inlets, the deep waters remain anoxic (Gross, 1972).
In both freshwater and marine systems, oxygen supplied at the
surface is steadily diminished with depth according to consumptive
demands.
Generally, dissolved oxygen concentrations are in a constant
state of flux on a daily basis (consumption/production biotic
responses) and seasonal basis (climatic and flow responses).
A normal diurnal oxygen cycle would be sinusoidal with a maximum
concentration late in the day and minimum in early morning. Distributions
of dissolved oxygen at depth, or oxygen profiles, have been studied
intensively in lakes and classified according to lake productivity.
While it is not necessary here to examine all identifiable variations
in oxygen distribution, it is useful to review some typical situations
to clarify the mechanisms which control oxygen content at depth.
In Figure 3, Wetzel (1983) has illustrated the idealized vertical
distribution of oxygen through one annual cycle of two dimictic
lakes.
Figure 3. Idealized Vertical Distribution of Oxygen Concentrations
and Temperature
(_) During the Four Main Seasonal Phases of an Oligotrophic
and an Eutrophic Dimictic Lake
Source: Wetzel, 1983.
At spring turnover, the water column is near 100 percent saturation
(12-13 mg /L O2 at 4 degrees Celsius and sea level) for both
oligotrophic and eutrophic lakes. There is no temperature/density
barrier to internal water circulation mechanisms, and complete
mixing is usually possible. With the onset of summer stratification
the oxygen profiles diverge. In the idealized oligotrophic lake
the oxygen concentration is regulated largely by physical means
as stratification occurs. As summer water temperatures increase
the dissolved oxygen concentration (and solubility) in the circulating
epilimnion decreases. Conversely, a temperature decline in the
metalimnion and hypolimnion causes the oxygen concentration to
increase and the level of saturation will be close to 100 percent
with increasing depth-an orthograde profile (Figure 3). In most
oligotrophic situations; however, there still is some organic
matter settling into the hypolimnion from the productive zone
of the lake and oxidation processes result in undersaturation
as stratification progresses (oxygen renewal by circulation and
photosynthesis not being possible).
In shallow water, the bulk of the loss attributable to oxidation
generally occurs at the sediment/water interface where bacterial
activity and organic matter are concentrated. A considerable
amount of oxygen also is lost in the water column by bacterial,
plant and animal respiration, particularly in deep lakes. Oxygen
depletion also occurs by direct chemical oxidation of dissolved
organic matter (usually masked by biochemical processes). In
eutrophic lakes, hypolimnion depletion may progress to complete
anaerobic conditions shortly after summer stratification-a clinograde oxygen profile (Figure 3). Once this happens, there is a shift
to anaerobic bacterial metabolism and decomposition processes
proceed at a slower rate. The fall turnover of a dimictic lake
is brought on by the cooling of epilimnial water with the resultant
breakdown of the density barrier at the thermocline and oxygenation
of the lower strata. In the winter stratification shown, ice
formation in temperate regions prevents the exchange of atmospheric
oxygen and the concentration profile for an idealized oligotrophic
lake is constant at saturation relative to depth. Again, biotic
influences of respiration and oxidation are normally present
and there is a reduction in oxygen concentration with depth.
In eutrophic lakes, the photic zone is reduced but can remain
active where sufficient light penetration through ice continues
and the normal consumptive demands are lessened by cooler temperatures.
The resultant oxygen profile showing depletion at depth is similar
to, but more gradual than for summer stratification (Wetzel,
1983).
One common variation of the oxygen profiles involves an increase
in metalimnetic oxygen and is termed a positive heterograde curve.
Typically, dissolved oxygen solubility in the epilimnion decreases
with increasing summer temperature and oxidative consumption
in the hypolimnion results in a clinograde curve. This leaves
the highest oxygen concentrations (saturation or higher) in the
metalimnion. When algal communities adapted to this strata (e.g.,
blue-greens) are active, supersaturation of several hundred percent
can result. Rooted aquatic vegetation in the lower littoral zone
also can enrich this zone. The actual depth of this maxima is
dependent on water transparency and is usually in the 3 to 10
m range. Metalimnetic oxygen maxima tend to be more pronounced
in relatively deep lakes that are well stratified; the bottom
profiles are such that a small proportion of the sediments (site
of highest bacterial utilization of oxygen) are in contact with
the metalimnion. A metalimnetic oxygen minimum (or negative
heterograde curve) is more uncommon and can result from a variety of reasons:
1) oxidizable material produced within or outside the catch-basin
is continuously decomposed while it sinks; when it encounters
the dense metalimnetic water it slows down and exacts a proportionately
greater oxygen demand in this zone, 2) respiratory demand from
large concentrations of zooplankton, and 3) in cases where a
gradual bottom slope (site of high oxygen utilization) coincides
with the prevailing metalimnion, horizontal mixing (greater in
the zone) can spread the reduction laterally.
There are other important mechanisms
that control oxygen levels in the littoral zone, which can
be quite different to that of
the pelagic zone. For example, well developed stands of aquatic
macrophytes and associated periphyton will substantially raise
oxygen levels during photosynthesis and consume oxygen during
respiration at night. This diurnal event also gives rise to a
diel cycle due to the net production of oxygen during the growth
seasons. Late in the year, much of the macrophyte standing crop
dies back to the root crown and the concomitant decomposition
can cause a prolonged oxygen demand extending beyond the littoral
areas. Understandably, the proportion of littoral development
ultimately determines the magnitude of these influences. Frodge
et al. (1990) observed that where this proportion was very high,
the dissolved oxygen profile can be radically different over
short distances. The horizontal distribution of dissolved oxygen
in two shallow Washington lakes was associated with the patchy
distribution of aquatic plants. Open water, being well-mixed
in comparison to vegetated areas, had higher sub-surface oxygen
concentrations than beneath either submergent or floating plant
canopies. Plant morphology largely determined the differences
in vertical oxygen profiles and diurnal changes (if any). For
submersed plants such as Myriophyllum sibiricum and Ceratophyllum
demersum, dissolved oxygen concentrations within and immediately
below the canopies were regularly at supersaturation (>30
mg/L) during daytime, but declined rapidly to consistently low
levels when the effects of wind and sunlight were cut-off. In
a typical western Washington lake dominated by Brasenia schreberi,
the floating leaves isolated the entire water column even further
and dissolved oxygen levels of 2 mg/L or less were recorded,
with little recovery during daylight. Brasenia schreberi and
other floating-leaved macrophytes are able to exchange gasses
directly with the atmosphere and make no substantial oxygen contribution
to the water column (Frodge et al. 1990). Profuse phytoplankton
growth in very eutrophic lakes also will produce diel fluctuations,
and if die-off occurs rapidly (e.g., when snow-covered ice blocks
photosynthesis) dissolved oxygen may be completely exhausted
(Wetzel, 1983).
Similar consumptive processes, which surpass
the resupply of oxygen, exist in coastal marine areas and are
often exaggerated by very high water temperatures (lowered oxygen
solubility) and organic effluents from anthropogenic sources.
The odorous hydrogen sulphide present at low tide in coastal
marshes is an indicator of anaerobic metabolism (Gross, 1972).
3.1.2 Oxygen Pathways
There is a plethora of complex pathways involving
oxygen production and utilization in water, the specifics of
which are not a primary concern here. However, brief descriptions
are given of some of the more important redox processes.
Other than atmospheric input, the principal source of dissolved
oxygen in the surface strata is photosynthesis. Phytoplankton
and attached vegetation also represent a major (often predominant)
supply of new organic matter. The simplified photosynthetic reaction
is summarized:
6 CO2 + 12 H2O
--------->C6 H12 O6 + 6 H2O + 6 O2 (Wetzel, 1983)
This reaction is mediated by a light pigment receptor
There is a transfer of carbon to carbohydrate and water is oxidized
to oxygen. This release of oxygen in the photic zone commonly
gives rise to slight supersaturation. In very rare instances,
photosynthetic oxygen supersaturation has caused mortality in
freshwater and marine fishes (NCASI, 1985).
Photosynthesis produces reduced states of
free energy and non-equilibrium concentrations of oxygen, carbon,
nitrogen and sulphur compounds. Respiratory, fermentative and
other reactions which bacteria mediate, act to restore this equilibrium.
This is accomplished by specific bacteria decomposing the thermodynamically
unstable products of photosynthesis and thereby obtaining a source
of energy for their metabolic demands. In large, deep lakes most
bacterial respiration may occur in the water column, whereas
in shallower water bodies with high organic inputs from allochthonous
sources (shoreline vegetation), benthic decomposition may dominate.
When oxidative respiratory processes in the lower strata outpace
the supply of oxygen, anaerobic metabolism proceeds using substitution
(oxidative) compounds. However, the oxidizable byproducts which
evolve (e.g., methane, hydrogen, hydrogen sulphide, carbon monoxide)
are released to the overlying oxic waters and further reduce
oxygen content in the system. In water, the principal reactants
in redox processes are limited to carbon, oxygen, nitrogen, sulphur,
iron and manganese. Some characteristic reactions involving dissolved
oxygen (hydrogen ion acceptor) in oxygen-consuming redox reactions
are listed below (from Wetzel, 1983):
(Iron) FeS2 + 31/2 O2 +
H2O --------> FeSO4 +
H2SO4
2 FeSO4 + 1/2 O2 + H2SO4 ----------> Fe2 (SO4)3 +
H2O
(in acid mine drainage where pH<3)
4 FeCO3 + O2 + 6 H2O ----------> 4
Fe (OH)3 + 4 CO2
(Manganese) 4 MnCO3 + O2 -----------> 2
Mn2O3 + 4 CO2
Mn2+ + 1/2 O2 + H2O -----------> MnO2 +
2 H+
(Sulphur) S + 11/2 O2 + H2O
----------> H2SO4
H2S + 1/2 O2 ----------> S0 + H2O (and see iron above)
(Nitrogen-Nitrification) NH4+ +
2 O2 --------> NO3- +
H2O + 2 H+ (overall reaction)
(Methane) 5 CH4 + 8 O2 ---------> 2
(CH2O) + 3 CO2 + 8 H2O
(Carbon Monoxide) CO + 1/2 O2 ---------> CO2
3.2 Anthropogenic Sources / Depletion
As discussed in the previous section, the dissolved
oxygen content of water is largely controlled by the balance
between input and consumptive metabolism (of oxidizable matter
received). Anthropogenic influences (industrial-including deforestation,
agricultural and municipal wastes) tend to load the latter scale
by the addition of organic effluents. The depletion of dissolved
oxygen in receiving waters is often a ready indicator of wastewater
treatment requirements, and specific empirical procedures have
been devised to measure oxygen demand. Biochemical oxygen demand
(BOD) is a standard microbial incubation procedure that measures
the oxygen required to oxidize organic material and certain inorganic
materials (e.g., sulphides and ferrous iron) over a given time
period (usually five days). Alternately, the measure for the
amount of oxygen required to chemically oxidize reduced minerals
and organic matter of a sample is termed chemical oxygen demand
(COD). Both terms are applied to the level of reducing material
present from a combination of natural and anthropogenic sources
and have particular usefulness in assessing the potential impacts
of effluents. Provincial waste discharge permits normally prescribe
an allowable quantity of BOD according to the volume of effluent
and consideration of the type of receiving water. Few ambient
guidelines exist for BOD/COD values and none are proposed; however,
McNeeley et al. (1979) consider waters with BOD5 levels greater
than 10 mg/L to be polluted and less than 4 mg/L to be reasonably
clean.
Wastes that are primarily nutrient related and/or high in carbon
may enhance or deplete dissolved oxygen levels and commonly do
both depending on the location in the water column. Additional
oxygen is usually produced in the photic zone by increased primary
production following enhancement by inorganic nitrogen and phosphorus,
but a subsequent drop in the nutrient supply can be accompanied
by algal die-off, decomposition and oxygen depletion. Gross (1972)
describes a secondary effect (negative) from a nutrient-enhanced
regime in a coastal environment, which allows blue-green algae
to replace diatoms as the dominant growth. Local zooplankton
are not able to utilize the algae, which then sink and enter
the decomposing cycle rather than the grazing food chain. Ultimately,
the depletion of deep water oxygen exceeds the rate of replenishment
from bottom currents, anoxia develops and benthic organisms die.
In British Columbia, some of the best illustrations of the profound
influences that effluent can have on dissolved oxygen in fresh
and marine waters come from the pulp and paper industry (bleached
kraft and sulphite mill effluents in particular). In some of
these cases, the resultant oxygen deficiency in receiving waters
is the most critical environmental hazard. Inlets which characteristically
have poor water exchange and naturally depressed oxygen levels
are vulnerable, as the assimilative capacity for high BOD discharges
may be minimal. Examples of this can be found in Kay's (1986)
review of coastal water quality, which reported serious dissolved
oxygen deficiencies associated with kraft and sulphite pulping
mills located at Alberni Inlet, Neuroutsos Inlet, Cousins Inlet
and Porpoise Harbour/Wainwright Basin. The Port Alberni case,
where a mill outfall is located at the head of an inlet adjacent
to a freshwater inflow (Somass River), is typical of the processes
which can affect the oxygen budget. The inlet is highly stratified
by a halocline layer, which varies between depths of 2 to 5 m,
separating freshwater and seawater (a natural barrier against
the introduction of atmospheric oxygen). The initial period of
discharge (1949-1956) directly to the harbour resulted in chronic
oxygen depression below the halocline due to the high BOD of
dissolved kraft mill effluents (lignin, carbohydrates, organic
acids, alcohols) and decomposition of solid organic material
on the bottom. Effluent was diverted to the mouth of the Somass
River in 1956 to allow diffusion into the more oxygen-saturated
surface layer; however, low oxygen levels persisted above and
below the halocline. Parker and Sibert (1972) suggested that
a secondary effect of pulping wastes on oxygen levels was caused
by the humic colour, which attenuated light penetration within
the photic zone and inhibited photosynthesis/oxygen production.
Since 1970, effluents at Port Alberni have been aerated and clarified
to reduce BOD and colour, and discharges have been varied to
correspond with flow in the Somass River. Unfortunately, most
of the improvement in dissolved oxygen concentration has been
confined to water above the halocline (Kay, 1986; Parker and
Sibert, 1972).
Aeration of effluent from pulp and paper mills and municipal
biological treatment lagoons is common practice, and in some
cases oxygen is introduced to bring concentrations closer to
that of the receiving water. Such applications of pure oxygen
have demonstrated efficiency and its further use within the forest
products industry is being considered (NCASI, 1985). A potential
hazard with oxygenation is that supersaturation can occur and
mitigation procedures may be necessary. Thermal effluents discharged
to saturated water also can cause supersaturation, since heating
actually reduces oxygen-carrying capacity.
Manipulation of oxygen levels is used in fish
hatcheries to increase the carrying capacity for high fish densities.
Similarly, in lakes which stratify, it is possible to increase
the area habitable by fish through increasing oxygen levels in
the lower strata. Hypolimetic aeration can reduce fish kills
in summer and improve the benthic food supply. Either air or
pure oxygen (to improve efficiency) is injected at depth without
destratifying the water column. A review of 13 lake aerators
by Cook et al. (1986) found that all but one increased the level
of hypolimetic oxygen to at least 7 mg/L. Oxygenation of the
hypolimnion can be accomplished in other ways. Artificial circulation
by injecting compressed air can serve to move deoxygenated water
to the surface and allow equalization with the atmosphere (rather
than provide oxygen by diffusion from the injected air).
The difference from hypolimnetic aeration is that the entire
water column is mixed and water temperature is increased at depth.
Hypolimnetic withdrawals from reservoirs can be used to provide
downstream cooling in summer,with a side benefit being the shortened
detention time of the lower strata water with less likelihood
of anoxic conditions developing.
Reduction of dissolved oxygen by mariculture net-pens is not
as common as in freshwater facilities because of the obvious
differences in waste assimilation capacity. Usually, the majority
of the oxygen demand originates from feed and feces which accumulate
below net-pens, and the effects tend to be localized and transient.
Weston (1986) reported that one and one-half to three times as
much oxygen may go to decomposition processes as to fish respiration.
In what was considered a worst-case situation, some decrease
in dissolved oxygen in and around net-pens was found in a poorly
flushed area in Henderson Inlet, Puget Sound. The effect was
limited to periods of summer stratification when surface and
bottom current velocities were extremely low. Most studies of
water quality around fish farms in British Columbia (particularly
Sechelt Inlet) and Washington State have shown mariculture to
have a negligible effect on dissolved oxygen levels (Weston,
1986).
3.3 Natural Levels In Water And Sediment
3.3.1 Freshwater
The concentration range of dissolved oxygen
in western Canadian surface waters, documented between 1980 and
1985, was from non-detectable to 18.4 mg/L (NAQUADAT, 1985).
In British Columbia surface waters, dissolved oxygen levels are
generally high (greater than 10 mg/L) and close to saturation.
Even in the lower Fraser River, where effluent discharges are
most numerous, median dissolved oxygen concentration from 1970
to 1978 were above 9.5 mg/L (10th percentiles exceeded 7.0 mg/L)
with saturations mainly between 80 and 100 percent (Drinnan and
Clark, 1980). Below the major sewage treatment plant outfall
at Annacis Island, the dissolved oxygen concentration showed
a decrease of less than 1 mg/L in the effluent plume, but levels
returned to normal downstream. In the main river channel, levels
were similar to the median of 10.8 mg/L oxygen for most streams
in British Columbia (Drinnan and Clark, 1980).
At townsites in the upper Fraser River watershed, the assimilative
capacity (flow) in the river and its tributaries is reduced,
and localized decreases in dissolved oxygen are more pronounced.
For example, mean dissolved oxygen levels in the Nechako River,
near effluent discharges at Fort Fraser and Vanderhoof, have
been reported as low as 4.7 mg/L (ENVIROCON, 1981). In the heavily
impounded Columbia River mainstem, dissolved oxygen levels rarely
fall below 8 mg/L. However, some of the tributaries which experience
low flows and elevated summer temperatures may contain less than
6 mg/L oxygen (Stober and Nakatani, 1992).
As discussed in Section 3.1.1, oxygen regimes in lakes are dependent
primarily upon seasonal temperature variation, depth and trophic
status. The Okanagan Valley is a major British Columbia waterway
that contains a number of lakes which illustrate the variability
in oxygen content relative to morphometry and productivity. A
study for the Canada-British Columbia Okanagan Basin Agreement
(1974) determined: epilimnial oxygen concentrations remained
near saturation in all lakes through the summer, the hypolimnia
of Osoyoos, Wood and Skaha Lakes were below saturation for the
same period, and the hypolimnia of Kalamalka and Wood Lake remained
well oxygenated. Based on a trophic index correlated with the
oxygen depletion of hypolimnetic water during summer stratification,
the lakes were classified as: Kalamalka and Okanagan Lakes-oligotrophic,
Osoyoos Lake-mesotrophic and Skaha and Wood Lakes-eutrophic (Canada
BCOBA, 1974). Kalamalka Lake exhibits an `orthograde' oxygen
profile (Figure 3 in Section 3.1.1.) during summer stratification,
while the shallow, productive water column of Wood Lake produces
a `clinograde' profile. The latter progresses to a `positive
heterograde' profile in mid-summer, whereby an improvement in
water clarity maximizes algal production at the thermocline.
Supersaturated oxygen levels (up to 135 percent) maintained by
the hydrostatic pressure at 10 m have been observed (Anon, 1974).
An example of a `negative heterograde' curve, was taken from
Skaha Lake in October, 1979 (Jensen, 1981). Since the study,
dissolved oxygen concentrations at Wood Lake have been improving
(Bryan, 1990).
3.3.2 Marine
The quantity of oxygen dissolved in seawater
is less (by about 20 percent) than in lower density freshwater
at a given temperature and pressure. For example, at a salinity
of 27 g/kg (now referred to as a chlorinity of 15-as per Section
2.1), oxygen solubility ranges from 12.1 mg/L at 0 degrees Celsius
to 7.1 mg/L at 25 degrees Celsius (compared to freshwater solubilities
of 14.6 and 8.3 mg/L respectively, from Table 1). Surface waters
usually are at or near equilibrium with atmospheric oxygen, while
supersaturation is common in the first few tens of metres (photic
zone). This surface-supplied oxygen is diminished by consumptive
demands in the upper several hundred metres and, in the deep
ocean where these demands are limited, oxygen concentrations
are relatively uniform but well below saturation (Gross, 1972).
Coastal waters have much more variable water quality and are
commonly the site of major anthropogenic activity and important
biological production. Most of British Columbia's open coast
tends to be well-mixed through upwelling, tidal currents and
wind forces and consequently is well-oxygenated. In the more
protected inside passages, dissolved oxygen values are extremely
variable and depend largely on the accessibility of deep water
inflow to replace oxygen depleted by biological processes. Thompson
(1981) explained that the dilution of Strait of Georgia water
by Fraser River runoff and the seaward migration of this brackish
surface layer, particularly through the southern passes, results
in an inflow of oceanic water through the Strait of Juan de Fuca.
This flushing action maintains the present oxygen regime of the
strait. Replacement time for surface water (less than 30 m) is
only one month. A 60 m deep sill between Victoria and Port Angeles
presents a partial blockage to the Pacific inflow; however, oxygen-rich
water still is able to penetrate to the deeper basins of the
Strait of Georgia and complete renewal can occur within a year.
The variability of the source water quality also has considerable
influence. Upwelling along the outer coast is most prevalent
in mid-summer, hence dissolved oxygen levels in the Strait of
Juan de Fuca are lowered. Added to the accelerated biological-related
consumption during this period in sub-surface waters, oxygen
levels in the Strait of Georgia reach minimum levels by early
Autum. Conversely, relatively oxygen-rich water moves into the
southern straits during the winter/spring months. Numerous sills
also are in evidence in the northern passage along Queen Charlotte
Strait, Broughton Strait, Johnstone Strait and Discovery Passage.
Some layering of the waters in Queen Charlotte Strait occurs
as the well-oxygenated surface water moves seaward and the stratified
ocean water moves inland. Dissolved oxygen levels at the bottom
of the basin remain above 4.5 mg/L. Active tidal currents coupled
with the movement of water over the irregular bottom topography
keep the remainder of the northern passage well-mixed and oxygen
levels uniform from top to bottom year-round. Thompson's (1981)
September, 1977 profile taken mid-channel from Broughton Strait
to Discovery Passage shows a range of only 5.1 to 6.6 mg O2/L.
The British Columbia coast is further characterized by its numerous
inlets, many of which are the site of intense commercial activity.
Oft-cited works by Pickard (1961; 1963) on the physical and chemical
features of prominent inlets along Vancouver Island and the mainland
are still useful summaries. He found that oxygen saturation values
generally increased slightly with depth in the low salinity surface
layers due to the drop in temperature and, below the brackish
water (halocline), dropped by about 10 percent with the rise
in salinity. Oxygen maxima at the head of an inlet were the result
of river inflows, while maxima at the mouth were most often cases
of phytoplankton causing supersaturation (110 to 130 percent).
In the latter situation, these maxima were at the 3 to 8 m depth
where phytoplankton were most concentrated. In deep water, oxygen
values declined to 35 to 60 percent saturation (3.8-5.7 mg/L);
minimum oxygen levels were sometimes homogeneous from a depth
of 200 to 300 m, to the bottom. Pickard (1961) also found mid-depth
maxima and explained that the frequent inflows of oxygenated,
higher salinity water over the sills near the entrance of inlets
(to replace net surface out-flow) displaced oxygen-deficient
water upwards. This water mass was then held at mid-depth below
the less dense surface layers.
Such occurrences were more typical of low-runoff inlets (e.g.,
Sechelt) where induced estuarine circulation was not sufficient
to replace oxygen. Low-runoff inlets, as typified by those on
Vancouver Island, tended to show greater variability of dissolved
oxygen than large-runoff inlets. Shallow sills were considered
important in limiting deep water circulation, and less internal
runoff (than mainland inlets) contributed to lower oxygen levels.
Below 100 m, dissolved oxygen was usually less than 5.7 mg/L,
in many cases was less than 1.4 mg/L and in a few sites within
the inner basins of Tofino, Nitinat and Saanich inlets, values
of Ong/L were recorded (Pickard, 1963). In Saanich Inlet (24
km long and 230 m deep) which receives low runoff and has a shallow
sill at 75 m, only the water above this level is well-mixed.
A thermocline usually develops in summer at 5 to 10 m and the
deep waters become anoxic. Oxygen is replenished once a year
by the inflow of water over the sill from the Strait of Georgia,
although this supply can be depleted in a few months (Harrison
et al. 1983).
3.3.3 Stream Sediment
Interstitial or sediment oxygen concentrations
are highly variable and can differ markedly from the overlying
water due to a number of independent variables which include:
surface and interstitial water velocity/discharge, hydraulic
gradient, sediment texture and porosity, bottom morphology, daily
water temperature fluctuation and the consumptive oxygen demand
of the substrate. In standing water environments, interactions
between the diverse sediment biota and the concentration of particulate
organic material are the major cause of oxygen depletion (both
through respiration and release of decomposition byproducts in
reduced form). As discussed earlier, oxygen itself will regulate
important redox reactions at the sediment/water interface. Oxygen
penetration into the substrate from the water column is governed
by the rate of turbulence of superficial sediments and by the
oxygen demand per unit volume; typically, diffusion from even
well-aerated water will only supply the top few centimetres (Wetzel,
1983). In some cases, filter feeding organisms can introduce
oxygen to greater depths by physically reworking compacted, oxygen-deficient
mud and extending the habitable range. Similarly, in riverine
environments, the act of redd construction by salmonids through
excavation and backfilling results in localized dispersal of
fine sediments and increased interstitial flows for improved
oxygen delivery to the buried eggs. The following discussion
is confined to stream sediment oxygen levels in spawning habitat.
Koski (1965) found relatively high oxygen
levels (close to saturation) in his examination of 31 coho salmon
redds in three small costal
streams in Oregon. The variability of his measurements was high,
both between redds (mean values of 3.0 to 11.8 mg O2/L)
and within redds (up to 6 mg O2/L at a given time).
Mean minimum interstitial
oxygen concentrations for the three streams were 7.4, 7.5 and
8.7 mg O2/L, which represented reductions of 3.2,
3.5 and 1.4 mg O2/L respectively, from surface water
values (at or near saturation). Hollender (1981) also encountered
high variability both spatially
and temporally in his study of dissolved oxygen in brook trout
redds in two Pennsylvania streams. Stream and interstitial water
temperatures were almost identical and followed the same temporal
pattern. The overall mean dissolved oxygen levels in natural
redds was 8.2 mg/L, within a range of means between 3.7 and 11.6
mg/L. Only about one-quarter of the redds had mean oxygen levels
below 6 mg/L. The interstitial concentrations were lower, but
closely followed those of the surface water which averaged more
than 10 mg/L. In one stream, the substrate/surface water differential
was 2.1 and 2.8 mg O2/L in consecutive years, and
in the second stream was 3.7 mg O2/L for one year.
Surprisingly, the corresponding
geometric mean particle sizes within redds were 2.5, 3.0 and
4.4 mm, respectively (an inverse relationship with dissolved
oxygen). On the strength of the two previously cited studies
on natural redds, the US EPA (1986) suggested that interstitial
oxygen be considered to be at least 3 mg/L lower than that of
the overlying water as a result of the reduced permeability and
consumptive demands in that region. Field data collected by Pyper
and Vernon (1955) suggested that the mean dissolved oxygen in
interstitial water of a typical sockeye spawning stream in British
Columbia would be about 75 percent of saturation. Sowden and
Power (1985) examined rainbow trout redds in an Ontario stream
which had been subjected to moderate sedimentation from agricultural
land. Ground water upwelling appeared to control the dissolved
oxygen content of the redds, although concentrations were variable.
The average mean oxygen level for 19 redds was 5.5 mg/L (range
of mean values was 2.0 to 8.9 mg/L). Unfortunately, no surface
water oxygen concentrations were collected for comparison.
Other researchers have reported levels of dissolved oxygen in
simulated redds, in which eggs are buried manually. Coble (1961)
noted that interstitial oxygen content in two Oregon streams
was closely correlated with sub-surface water velocity and measured
a reduction of about 5 mg/L in artificial redds relative to the
overlying water (interstitial oxygen averaged 6 mg/L). Similarly,
McLean (1988) described the results of siltation on artificial
spawning media in the Little Qualicum River and found that interstitial
oxygen was positively correlated with substrate permeability.
Dissolved oxygen levels in silted channel sections varied between
0.5 and 9.3 mg/L (5.4 mg/L approximate average), and after intensive
cleaning all interstitial oxygen levels were above 9.5 mg/L.
Turnpenny and Williams (1980) recorded interstitial dissolved
oxygen concentrations in artificial rainbow trout redds of between
8 and 11.4 mg/L (80-103% of saturation), which were relatively
high considering the industrial influences at some of their study
sites. At two control sites, away from potential impacts, the
maximum differential between surface water (saturation) and interstitial
water over a 28-day incubation period was 2 to 3 mg/L. Chevalier
and Murphy (1985) reported oxygen levels within simulated redds
along the Tucannon River, Washington. Between February and June,
1980 (period of salmon egg incubation), the differential between
average mean surface oxygen saturation and interstitial water
at five sites was 3.4 mg/L; however, the range of mean values
was considerable (0.19 mg/L to 7.71 mg/L). Low oxygen levels
in the lower reaches reflected the presence of organic matter
and high sediment oxygen demand. During a second period in May,
1981, the relationship between sediment organics and oxygen levels
was less obvious, although the mean surface/sub-surface differential
was similar at 3.75 mg/L (surface saturations were 10.53 to 11.83
mg/L, while interstitial concentrations varied from 3.58 to 11.26
mg/L). The investigations of artificial redds by Turnpenny and
Williams (1980) and Chevalier and Murphy (1985) support the US
EPA's assumption (based on natural redds) that a differential
on the order of 3 mg/L exists between the interstitial environment
and the water column.
Except in cases of upwelling ground water,
the oxygen content of interstitial water is highly dependent
on flow driven by the
overlying hydraulic gradient. Bed composition and particle size
play a major regulatory role in the extent of this downwelling.
Understandably, substrates predominated by fines have more limited
exchange with surface water for oxygen replenishment. Whitman
and Clark (1982) found that restricted porosity resulted in considerable
variation between surface and sediment water oxygen in a Texas
stream. Along a homogeneous, sandy riffle there was a reduction
(from overlying water) of approximately 5.5 mg O2/L
at a sediment depth of 10 cm and interstitial oxygen averaged
about 3.5 mg/L.
The level of fine particles (clay, silt, sand) can become critical
in salmonid spawning gravels.
Some of the most detailed work on the deposition of fine particles
and resultant impairment of fish production has been done in
the Pacific Northwest. The Tucannon River study in southeastern
Washington was primarily concerned with fish habitat loss attributable
to agricultural practices. Chevalier and Carson (1985) investigated
the intricate relationship between sediment fraction size and
flow mechanics of this river and derived an interstitial dissolved
oxygen transfer model for use in predicting salmonid embryo survival.
The hydraulic impairment by fine particles appears to become
critical to egg survival when the sand fraction approaches 20
to 25 percent by volume or when silt / clay / fraction reaches
only one to two percent by volume. A small amount of silt was
found to be particularly effective at blocking interstitial flow
if the fine particles were arranged in layers and not evenly
dispersed (often the case in the Tucannon River). Through modelling
parameters which approximated natural conditions of the river,
the researchers showed that vertical and horizontal diffusion
of oxygen was insignificant below the top 8 cm of substrate materials
relative to convective transport of oxygen by horizontal flow.
However, as pore spaces became plugged, velocity decreased to
the point where diffusion became the only source of oxygen supply.
In British Columbia, a long term study on the effects of logging
on salmonid spawning habitat was conducted at Carnation Creek
on Vancouver Island. Scrivener and Brownlee (1980) determined
that prior to the effects of major siltation (pre-1978) the mean
interstitial oxygen concentrations had followed seasonal trends
relative to surface concentrations: pre-spawning period was 5.4
mg/L, post-spawning period was 6.3 mg/L and pre-emergence period
was 6.6 mg/L, at a mean saturation potential of 12.6 mg/L. Subsequent
declines in dissolved oxygen were attributed to increased consumption
from greater loadings of organics and a decrease in gravel permeability
due to infilling by fine sediments. Through 1980, interstitial
oxygen concentrations ranged between 0.5 and 11.4 mg/L. In August,
1982, interstitial oxygen concentrations ranged from 1.3 to 8.9
mg/L while surface concentrations were 7.5 to 9.5 mg/L. The researchers
noted that gravel size could not account for more than 20 percent
of the variability in dissolved oxygen and a positive correlation
with mean particle size was only apparent in the top 15 cm layer.
Interstitial oxygen varied between 18 and 96 % of surface levels
when the geometric mean particle size was 10 to 12 mm, but was
always greater than 60 % when particle size was more than 18
mm (Figure 4). While other influences (flow and direction, hydraulic
gradient, micro-topography, consumptive demand) obviously were
strong, it was suggested that a better correlation may have been
obtained at times other than during summer minimum flows (Scrivener
and Brownlee, 1989). Most of the study sites probably had adequate
oxygen for normal development prior to 1978 (pre-logging) when
post-spawning interstial oxygen levels were normally distributed
about a mean concentration of 6.3 mg/L. After logging, the comparitive
mean was only 3.2 mg/L (not normally distributed but skewed in
the direction of greater oxygen deficiency), which represents
an oxygen deficient environment for developing chum salmon (Scrivener
and Brownlee, 1980).
Alternatively, workers such as Chevalier and Carson (1985) have
expressed the opinion that the validity of mean particle size
for such comparisons is over emphasized and that fine particles
themselves account for most of the variability in dissolved oxygen
concentrations. Again, an exception to the relationship of substrate
size composition and oxygen content is ground water-fed streams.
Sowden and Power (1985) found that superficial substrate characteristics
may not have a strong influence on the oxygen content of interstitial
water, but rather may be governed by the level of oxygen depletion
in the aquifer source-water. Similarly, interstitial flow rates
can be determined by hydraulic pressure originating in the aquifer.
Figure 4. Relationship Between Interstitial Dissolved Oxygen
(as Percent of Surface Concentration) and Mean Particle Size
in the Top Layer of Streambed
45
sites on Carnation Creek During August, 1982
Source: Reproduced from Scrivener and Brownlee,
1989
