|
||||||
|
||||||
|
|
|
||||||||||||||||||||||||||||||
|
Water Quality Ambient Water Quality Criteria for Colour in British Columbia: Technical Appendix 4. Aquatic Life 4.1 Effects The presence of coloured substances in aquatic systems has been associated with changes in primary productivity (Ilmavirta and Huttunen 1989; Arvola 1986; Del Giorgio and Peters 1994; Henebry and Cairns 1984; Haynes et al. 1994), depth of the euphotic zone (Eloranta 1978), phytoplankton species composition (Sheath et al. 1986; Vegas-Vilarrubia 1995; Ilmavirta and Huttunen1989; Beauchamp and Kerekes 1989), protozoan colonization rates (Henebry and Cairns 1984), secondary production (Hessen 1985), macro-invertebrate behaviour (Juarez et al. 1987), and macro-invertebrate community structure (Kullberg 1992). Colour can also alter the availability and hence toxicity of heavy metals to fish (Nilsson and Häkanson 1992; Haines et al. 1995; Nilssen 1982; Hutchinson and Sprague 1987). The effects of colour on aquatic biota are discussed in the following sections.
Several studies have noted that brown-water systems with high concentrations of dissolved humics are associated with increased bacterial productivity (Hessen 1985; Tranvik et al. 1991). For example, Hessen (1985) observed that the mean annual biomass of bacteria in holarctic oligotrophic lakes was 7.8-12.1 g C/L in clear lakes, 10.5 g C/L in slightly coloured lakes and 16.2-44.1 g C/L in highly coloured humic lakes (note that colour was not quantified in the humic lakes). Bacterial and zooplankton biomass were also high relative to algal biomass, strongly supporting the possibility that humics are an essential component of the carbon pool in these systems. Similarly, Tranvik et al. (1991) observed a significant correlation between bacterial abundance in 23 oligotrophic lakes in southern Sweden and water colour (Spearman rank correlation, rs = 0.62). Bacterial abundance in the highly coloured lakes (colour = 180 mg/L Pt) was approximately double that observed in clearwater lakes (colour = 5 to 20 mg/L Pt). It should be noted that some anthropogenic coloured substances such as textile dyes may cause reductions in bacterial biomass, although such effects are likely due to chemical toxicity rather than to changes in the optical properties of the water (Tratnyek et al. 1994).
The amount of photosynthetically active radiation in natural waters is of fundamental importance in determining the productivity of vascular plants and phytoplankton. Absorption and scattering by water and dissolved and particulate matter determine the quantity and spectral quality of light at a given depth, which in turn affect the photosynthesis of primary producers. The following discussion relates the effects of colour to primary producers in freshwater and estuarine and marine systems.
Numerous studies have demonstrated a strong positive correlation between primary production and water colour in freshwater lakes and rivers (Del Giorgio and Peters 1994; Ilmavirta and Huttunen 1989; Arvola 1986; Henebry and Cairns 1984). This correlation exists in spite of the smaller euphotic zone that characterizes highly coloured aquatic systems (Del Giorgio and Peters 1994; Ilmavirta and Huttunen 1989; Eloranta 1978). Eloranta (1978), for example, found that the euphotic zone in freshwater lakes in Finland decreased from 10 m to 4-5 m with an increase in colour from 5 to 20 mg/L Pt. Several factors could explain this apparent anomaly. First, most measures of annual productivity are confined to the euphotic zone or epilimnion. Productivity of benthic and periphytic algae would likely be higher in clearwater systems. Second, water colour is often positively correlated with nutrients such as total phosphorus and nitrogen (Del Giorgio and Peters 1994; Ilmavirta and Huttunen 1989; Henebry and Cairns 1984); high concentrations of nutrients have a profound effect on primary productivity, perhaps sufficient to override the reduced light penetration in highly coloured systems. Other possible factors contributing to the positive relationship between water colour and primary production include (i) zooplankton grazing on phytoplankton will be more effective in well illuminated water (Ilmavirta and Huttunen 1989), and (ii) humic acids in highly coloured, low pH systems act as a weak buffer and can therefore reduce the effect of acid deposition on primary production (Kullberg 1992). A more direct means of ascertaining the effects of water colour, both true and apparent, is to examine how changes in water colour affect primary producers while other variables (e.g., nutrient concentrations) are held constant. Carpenter et al. (1996) observed that artificially splitting Long Lake in Wisconsin into two basins led to a three-fold increase in water colour in the east basin, while little change occurred in the west basin. The change in water colour was likely the result of a two-fold increase in dissolved organic matter. The increased water colour in the east basin led to a shift in the vertical distribution of phytoplankton such that the proportion of chlorophyll a in the metalimnion was approximately 0.32 in this basin, compared to 0.62 in the west basin (Christensen et al. 1996). This change was accompanied by a reduction in the depth of oxygenation in the east basin. In addition to the effects on primary productivity, water colour can also affect species composition (Arvola 1986; Ilmavirta and Huttunen 1989; Vegas-Vilarrubia 1995; Sheath et al. 1986). This is expected since algal absorption spectra, and hence photosynthetic efficiency at various wavelengths, differ markedly among algal groups according to the amounts of accessory pigments accompanying chlorophyll a (Atlas and Bannister 1980). Also, zooplankton grazing will be effective at eliminating large, non-motile algae in clearwater systems, but less so in coloured systems (Ilmavirta and Huttunen 1989). Ilmavirta and Huttunen (1989) found that humic-stained lakes (median colour = 150 mg/L Pt) in Finland tended to have larger populations of blue-green algae (which can increase the concentration of the phycocyanin pigment when grown in water with a dominant wavelength in the red portion of the spectrum), motile flagellates (which are efficient at finding optimum light intensities and spectra), and overall higher species richness than did clear lakes (median colour = 5 mg/L Pt). Sheath et al. (1986) also found that a blue-green alga, Phomidium retzii, dominates lowland, brown-water streams in south-central Alaska (colour not quantified), while the chrysophyte, Hydrurus foetidus, dominates clearwater streams, the latter of which also has lower species richness.
Few studies have quantified the effects of colour on estuarine and marine primary producers. Estuaries are often enriched with plant nutrients compared with offshore waters because of inputs from the surrounding landscape. When nutrients are abundant, phytoplankton populations may flourish in the upper waters and cause shading of benthic algae and sea-grasses (Orth and Moore 1983). Reduced light availability has been implicated in the decline of sea-grasses and submerged vegetation in several of Florida's largest estuaries including Charlotte Harbour, Tampa Bay and Indian River lagoon (Orth and Moore 1983; Gallegos and Kenworthy 1996). In Charlotte Harbour, seagrass coverage was reduced by 29% between 1945 and 1982. Light attenuation is particularly high in the northern half of this estuary as a result of basin runoff and inflows from tidal rivers (Arvola 1986). In this area, dissolved matter caused a reduction in the amount of photosynthetically active radiation (PAR), such that much of the bottom was below the depth of 1% surface light. Not only was PAR reduced, but the spectral distribution shifted such that the shorter wavelengths (400-500 nm) were effectively eliminated in the first few centimeters of water in highly coloured waters (greater than or equal to 25 mg/L Pt). This is the part of the spectrum that higher plants such as sea-grasses use most effectively in photosynthesis. Dissolved matter had little effect on light attenuation in much of the southern part of the estuary (colour <5 mg/L Pt), where seagrass coverage is much more extensive. Several studies have indicated that the deep edges of seagrass beds extend to depths at which approximately 20% of surface irradiance penetrates (Gallegos and Kenworthy 1996). When turbidity is low (<2 NTU), a modest colour change from 5 to 7.9 mg/L Pt can decrease the lower limit of the seagrass bed from 2 m to between 1.25 and 1.5 m (Arvola 1986). A laboratory investigation of the effects of coloured Lake Coleman solution surrounding growth tubes with the estuarine diatom Phaeodactylum tricornutum indicated a strong negative relationship between water colour and cell density (Haynes et al. 1994). The source of colour to Lake Coleman (Australia) were pulp and paper mill, domestic and industrial effluents. Isolating the growth tubes from the coloured solution was required to eliminate the effects of nutrients and other water quality variables. The resulting regression equation was:
where CD is algal cell density (x10,000/mL) and Col is colour measured as mg/L Pt. Using this equation, a change in water colour from 10 to 50 mg/L Pt would cause a 15.3% reduction in algal cell density. Coloured pulp and paper mill effluents have also been found to reduce algal productivity in areas with poor flushing along the British Columbia coast. Parker and Sibert (1976) found that a humic-stained effluent with a colour of 1000 mg/L Pt discharged from a British Columbia mill to upper water layers prevented photosynthesis in sub-halocline phytoplankton at the head of Alberni inlet.
Many invertebrate species possess visual receptors with peaks that correspond to the spectral quality of their preferred habitats. For example the opossum shrimp Mysis has an absorption spectrum peak at 515 nm and is found only in deep, clear, photically blue environments (Wetzel 1975). Other species such as the freshwater prawn Macrobrachium rosenbergii have a strong behavioural preference for dark coloured backgrounds, likely because they perceive dark colours as a refuge (Juarez et al. 1987). Thus, changes in the spectral quality of water could have profound effects on the behaviour of some invertebrates. As noted earlier, increased water colour may inhibit the growth of periphytic algae. Invertebrate grazers may therefore be indirectly affected by the presence of water colour, through a decline in their food resources (Kullberg 1992). Kullberg (1992) observed that several taxonomic groups of benthic macro-invertebrates including the shredding trichopterans exhibited a negative correlation to water colour in 20 streams in southern Sweden. This study also noted that species richness declined as water colour increased when the pH was greater than 5.7.
As with invertebrates, many fish species have visual pigments that correspond to the spectral nature of their habitat. For example, freshwater species typical of photically blue environments (e.g., clear, deep lakes) have more pigments in the blue and green portions of the visible spectrum (Wetzel 1975). Thus, changes in water colour could also affect the behaviour of fish. Most studies of the effects of water colour to fish have focussed on the interaction between colour and toxicity of metals. For example, several studies have shown that aluminum, zinc and copper complex with humic substances in coloured water (Nilssen 1982; Winner 1985; Wilson 1972), thus reducing toxicity (Winner 1985; Hutchinson and Sprague 1987). Hutchinson and Sprague (1987) observed that the combined LC50 for a mixture of Al/Zn/Cu to flagfish declined by a factor of 2.1 when apparent colour (i.e., water samples were not filtered) increased from 3 to 10 mg/L Pt. Conversely, mercury availability, bioaccumulation and hence toxicity increase as water colour increases (Nilsson and Häkanson 1992; Haines et al. 1995; Mierle and Ingram 1991). Part of the reason for this relationship is that the mercury brought to an aquatic system from the surrounding catchment area is attached to coloured substances (Nilsson and Häkanson 1992; Mierle and Ingram 1991). This is also likely true for other metals. The difference is that other metals often remain bound to organics once in an aquatic system. Further, in deep water systems, mercury methylation by bacteria under anoxic conditions will likely be enhanced with higher concentrations of humic matter (Nilsson and Häkanson 1992). For deep lakes in Sweden, regression analyses indicate that an increase in water colour from 10 to 50 mg/L Pt was associated with an increase in mercury tissue levels in fish of between 23.4 and 82.8%, depending on fish species, and location, and depth of lake (Nilsson and Häkanson 1992).
Based on a review of the literature, few jurisdictions have derived criteria for the protection of marine, freshwater aquatic life, either for apparent or true colour (Tables 1 and 2). The CCME (1987 and updates) does not have colour guidelines for the protection of freshwater or marine aquatic life. Table 1. Colour criteria for freshwater aquatic life.
Table 2. Colour criteria for marine aquatic life.
With the exception of the colour criterion derived by the Province of Alberta (Alberta Environmental Protection 1994), the criteria for the protection of freshwater and marine aquatic life do not specify a particular true colour value or values, but rather are narrative statements about the effects that are to be avoided (US EPA 1972, 1976; Alaska Department of Environmental Conservation 1979; Ministére de l'environement du Quebec 1992; Australia 1992). No rationale was provided to support the colour criterion by the Province of Alberta.
4.3.1 Apparent Colour The 30-day mean transmission of white light shall be less than or equal to 80% of background levels as measured historically or at appropriate reference sites. This criterion applies to freshwater, estuarine and marine aquatic systems. The 30-day mean calculation should be based on a minimum of five weekly samples taken over a period of 30 days. Estimating percentage transmission of white light requires simultaneous measurements of light intensity at the surface and at a selected depth (generally below 1 m to avoid effects from surface agitation). Percent transmittance is based on total white light and is thus a composite for all visible wavelengths, each of which is variously influenced by water, dissolved matter and particulate matter (Wetzel 1975). Eloranta (1978) measured light penetration in water with a 'Submarine Fotometer' consisting of an underwater selenium photocell equipped with 2 mm thick Schott filters with peak transmissions at 432 nm (blue), 527 nm (green), and 627 nm (red) and connected with a micro-ammeter. Using such an apparatus, percent transmittance (per m) relative to surface incident light could be estimated by averaging across wavelengths. Precision may be increased by including more wavelengths.
The 30-day mean true colour of filtered water samples shall not exceed background levels by more than 5 mg/L Pt or (> 5 true colour units) in clearwater systems (background levels less than or equal to 20 mg/L Pt) or 20% in coloured systems (background levels >20 mg/L Pt). This criterion applies to freshwater, estuarine and marine aquatic systems. The 30-day mean calculation should be based on a minimum of five weekly samples taken over a period of 30 days. In situ single wavelength analysis at 456 nm with the results calibrated against the Hazen measurement scale is the preferred analytical methodology for true colour. Section 1.4 and Bennet and Drikas (1993) describe this methodology in more detail.
4.4.1 Apparent Colour All the criteria available for apparent colour in general are related to changes in colour relative to background volume. The review of the literature on effects indicated that relatively small changes in light attenuation by dissolved organic matter and/or suspended particulates can have a profound impact on the lower limit of the euphotic zone (Eloranta 1978). This effect can lead to reductions in primary productivity (Parker and Sibert 1976; Haynes et al. 1994; Christensen et al. 1996), coverage of submersed macrophytes (Orth and Moore 1983; McPherson and Miller 1987; Gallegos and Kenworthy 1996), and indirect impacts at higher trophic levels in both freshwater and estuarine systems (Kullberg 1992). Therefore, an increase in apparent colour, as a result of anthropogenic inputs, should be of concern. Percent transmission of light is an easily measured parameter that indicates the amount of photosynthetically active radiation available to primary producers at lower depths (Wetzel 1975). Transmission of white light is a function of both components of apparent colour, dissolved and particulate matter, and therefore is a useful monitoring tool for this parameter (Jerome et al. 1994a,b). Transmission of white light will exhibit considerable spatial, year-to-year, and seasonal variation in British Columbia aquatic systems (Jerome et al. 1994a). Thus, it is not useful to specify a single value as the ambient water quality criterion for apparent colour. A more useful approach is to determine whether a particular anthropogenic activity (e.g., road construction, harvesting of forests) is causing a decrease in transmission of white light compared to historical conditions or to background conditions in nearby aquatic systems. We have arbitrarily selected a 20% reduction in the amount of white light transmission for the criterion to ensure minimal impacts on productivity. The sampling design used to determine whether transmission of white light has been reduced by less than 20% as a result of an anthropogenic activity needs to be flexible. Some considerations include availability of historical data and/or reference sites, nature of the anthropogenic activity (e.g., point versus diffuse sources), and various legal and economic issues. Short-term changes in apparent colour (e.g., <24 hours) can arise due to natural events (e.g., storms) and are unlikely to have serious impacts on primary producers. Therefore, when testing whether a particular anthropogenic activity has caused an increase in apparent colour above the water quality criterion, several samples (n greater than or equal to 5) should be taken over a 30-day period and a mean calculated (or median if the data are right skewed).
Changes in the spectral quality of light in water can have a profound impact on primary productivity (Gallegos and Kenworthy 1996), phytoplankton species composition (Atlas and Bannister 1980; Ilmavirta and Huttunen 1989), and foraging behaviour and habitat selection of invertebrates and fish (Kullberg 1992; Juarez et al. 1987; Wetzel 1975). Several studies have shown that increases in true colour of approximately 5 mg_L-1 Pt in clearwater systems can have a profound impact on the depth of the euphotic zone (Eloranta 1978) and photosynthetic rates of algae and macrophytes (Gallegos and Kenworthy 1996). This is the basis for true colour criterion in clearwater systems. Further increases in colour have a less dramatic impact on depth of the euphotic zone (Eloranta 1978). The 20% above background cutoff for coloured systems is somewhat arbitrary. The value could not have been much lower because natural variation would lead to frequent criterion exceedances. Changes in spectral quality are much more influenced by changes in true colour than by changes in concentrations of particulate matter, because the latter have relatively non-selective scattering properties. The appropriate wavelength range for determining spectral quality and hence true colour should be in the blue portion of the spectrum because water absorption is low in this region and because humic and fulvic acids exhibit equal absorbance to the standard Pt-Co reference solution around 410 nm and 445-470 nm (Bennett and Drikas 1993; Hongve and Akesson 1996). Therefore, the Pt-Co colour standard is an adequate measure for true colour in aquatic systems. Note that colourimetric measures of true colour are more precise than are comparator methods, although both can be used to determine true colour. As with apparent colour, the sampling design used to determine whether the true colour criterion has been exceeded as a result of an anthropogenic activity needs to be flexible. Some considerations include availability of historical data and/or reference sites, nature of the anthropogenic activity (e.g., point versus diffuse sources), and various legal and economic issues. Short-term changes in true colour (e.g., <24 hours) can arise due to natural events (e.g., storms) and are unlikely to have serious impacts on primary producers. Therefore, when testing whether a particular anthropogenic activity has caused an increase in true colour above the water quality criterion, five weekly samples should be taken over a 30-day period and a mean calculated (or median if the data are right skewed).
|
|||||||||||||||||||||||||||||||
|
|
