The aim of the research introduced in this paper is to develop a nutrient export coefficient model relating nutrient loading from land use categories in the watershed to algal community structure and biomass in Laurel Creek.
The export coefficient modelling approach aims to predict the concentrations of total nitrogen and total phosphorus at any site in the stream as a function of the export of nutrients from each nutrient source in the catchment above that site (Johnes et al, 1996). The model is constructed using data collected on the spatial distribution of land use and fertilisers applied to each land use type; the numbers and distribution of livestock and human populations in the watershed; and the input of nutrients to the catchment through nitrogen fixation and atmospheric deposition. Export coefficients are derived from the literature to determine the loss of nutrients from each identifiable source to the stream.
Export coefficient modelling has been successfully modified and used in Britain (Johnes et al, 1996). It has also recently been adapted to allow determination of nutrient loading to lakes in England and Wales, in the development of a new lake classification and monitoring scheme for the National Rivers Authority (Johnes et al, 1994). The export coefficient modelling approach has a number of very clear advantages over more detailed, process-based modelling approaches, in particular the simplicity of the model format and the operation of the model using a spreadsheet system. The model provides an effective means of evaluating the impact of land use and land management on water quality (Johnes et al, 1996). The model predicts and is calibrated against total nitrogen and total phosphorus concentration data.
Numerous studies have demonstrated the sensitivity of attached algal community structure and biomass to nitrogen and phosphorus loading (e.g. Whitton et al., 1991; Kelly and Whitton, 1995). There are two important arguments in favour of biological monitoring of freshwater systems. Firstly, because organisms have an integrating response to their environment, fluctuations in water quality which may be missed by intermittent chemical analysis are recorded. Secondly, if we wish to maintain healthy, diverse biological communities, it is more appropriate to monitor the aquatic community rather than physio-chemical variables only (Cox, 1991).
Water quality samples were taken from each Laurel Creek site (see Fig. 1) every two weeks from May to September in 1995 and 1996 and analyzed for: total nitrogen; ammonia; nitrate; total phosphorus; soluble reactive phosphorus; total dissolved phosphorus; biological oxygen demand; alkalinity; total suspended solids; and chlorophyll. Depth, conductivity, temperature, pH and total dissolved solids were measured in the field. In addition, samples were taken about once a month from May 1995 to March 1997 for total nitrogen and total phosphorus analysis.
Algal samples were collected once a month from each site from May to September in 1995, and from May 1996 to March 1997. Sampling procedures were based on those of Pipp and Rott (1993) and Rott and Duthie (1997). At each site, separate samples of diatoms were taken by scraping rock surfaces, siphoning sediments and collecting macrophytes. Diatom slides were prepared from each sample by digesting a subsample in hydrogen peroxide and mounting in Hyrax. Diatoms were identified by microscopic examination, and 200 diatoms were counted from each sample. A fixed area of rock surface was scraped at each site, and ash free dry weight and chlorophyll measurements were made to calculate attached algal biomass. Non-diatom algal community structure was mapped using a quadrat, and subsamples were taken and identified.
Semi-quantitative data analysis was based on epilithic (those growing on rock surfaces) diatom samples. Temporal and spatial changes in diatom community structure have been identified, in relation to the measured physical and chemical parameters. To identify the most discriminant environmental variables, multivariate evaluation has been made using detrended correspondence analysis (DCA) and canonical correspondence analysis (CCA) using the program package CANOCO version 3.1 (Ter Braak, 1988, 1990). DCA is an ordination technique that was used to look at patterns and variations in the species data without using environmental variables. CCA is another ordination technique that identified the environmental variables, of those measured, most important in explaining the variation in the species data.
The results presented are for 1995 epilithic diatom counts for May, July and September. Using DCA, gradient lengths indicated that numerical techniques based on a unimodal species response model were most appropriate for analysis. CCA was therefore used to explore the relationships among the distribution of the epilithic diatom taxa and the measured environmental variables.
A series of CCAŐs was run to assess colinear environmental variables (i.e. variables with variance inflation factors, VIFŐs, greater than 20). Variables with VIFŐs greater than 20 were almost perfectly correlated with the other variables and had no unique contribution to the ordination (Ter Braak, 1988). The variance with the highest VIF was omitted from the environmental data set, and the CCA was rerun until all VIFŐs were less than 20. Canonical coefficients and approximate t-tests were then used to identify environmental variables which were important in explaining the directions of variance in the distribution of diatom taxa. The significance levels of the CCA axes were assessed using Monte Carlo permutation tests (999 random permutations).
In May, July and August, conductivity increased to very high levels at sites 5 and 6 (see Fig. 2). High conductivity indicates a high concentration of ions, and the high levels at sites in the urban area are an indication of urban run-off.
Biological oxygen demand (BOD) is an indication of the microbial activity in the water, for example bacterial processes, which uses up oxygen. In May and June, BOD was higher in the headwater sites than during other months (see Fig. 3), which may be indicative of agricultural inputs such as livestock waste or manure used as fertiliser. In July, August and September, BOD was very low in headwater sites and increased downstream of campus and Silver Lake. This indicates that inputs from the impoundments and urban run-off are causing an increase in BOD at these sites. One important source is waterfowl faeces.
Total nitrogen is very variable in concentration (see Fig. 4). Concentrations generally tended to be higher in headwater sites, indicating the run-off from agricultural land. Ammonia generally increased at sites downstream of the impoundments (see Fig. 5). The toxicity of ammonia is dependent on the temperature and pH conditions of the water, and according to Canadian water quality guidelines (Canadian Council of Resource and Environment Ministers, 1987) ammonia reached concentrations that were toxic to aquatic life at sites 3 and 4 in July and August, given the pH and temperature conditions. Although the concentration of ammonia was higher at all sites in September, levels did not exceed the guidelines because water temperature was lower. Likely sources of ammonia are microbial activity and waterfowl waste from the impoundments.
Headwater sites generally had lower concentrations of total phosphorus than sites in the urban area, with peaks in total phosphorus observed downstream of the impoundments (see Fig. 6). A large proportion of the phosphorus measured reflects that bound to particulate material, and high concentrations are related to the characteristic turbidity of the creek in the urban area. Total dissolved phosphorus (see Fig. 7) tends to be higher in headwater sites, and the proportion of dissolved phosphorus decreases in the urban area (see Fig. 8), indicating the importance of phosphorus bound to particulate material downstream of the impoundments.
The concentration of total suspended solids indicates the amount of particulate material borne in the water. It is a significant water quality parameter because it reduces light penetration and so interferes with aquatic plant growth (Flanagan, 1990). The concentration of total suspended solids generally peaked at sites 3 and 4, and was high at sites 5 and 6 too (see Fig. 9). Inputs of particulate material from the impoundments are important in Laurel Creek, and include phytoplankton cells. Planktonic chlorophyll a analysis shows peaks in chlorophyll a downstream of the impoundments in May and, especially, July and August (see Fig. 10), indicating phytoplankton production is important in terms of particulate loading to the stream at certain times of the year.
As determined in 1995 counts, attached algal community structure for Laurel Creek sites 1 to 6 was dominated by species typically indicating eutrophic conditions (Winter, unpublished data). A relatively high proportion of species indicating hypereutrophic conditions was found at sites in the urban area.
After deleting colinear variables (see methods), CCA identified 10 environmental variables that explained directions of variance in the species data along one or more of the first two axes (see Fig. 11) total suspended solids (TSS); biological oxygen demand (BOD); total nitrogen (TN); alkalinity (alk); nitrate (NO3); pH; total dissolved solids (TDS); temperature (temp); total phosphorus (TP); and soluble reactive phosphorus (SRP). These environmental variables are represented by arrows on the ordination plot, and those with the longest arrows are most highly correlated with the axes. The CCA axes constrained to these environmental variable were significant (P 0.01). Species / environmental correlations were high, which indicates that the 10 environmental variables accounted for the major gradients in the composition of the epilithic diatom assemblages in this data set.
In conclusion therefore, Laurel Creek is impacted by agricultural inputs in the headwaters, and urban run-off downstream. The impoundments in the system have a further deleterious effect on water quality, and these influences are reflected in the attached algal community structure.
Cox, E.J. (1991) "What is the basis for using diatoms as monitors of river water quality?" in Whitton, B.A et al (eds.) (1991) The use of algae for monitoring rivers. University of Innsbruck, Austria pp. 33-41.
Flanagan, P.J. (1990) 2nd Edn. Parameters of water quality. Interpretation and standards. Environmental Research Unit, Ireland.
Grand River Conservation Authority (GRCA) (1993) Laurel Creek Watershed Study.
Grand River Conservation Authority (GRCA) (1994) Laurel Creek. Stewardship News.
Johnes, P., B. Moss & G.L. Phillips (1994) Lakes - classification and monitoring. A strategy for the classification of lakes. National Rivers Authority R&D Project Record 286/6/A.
Johnes, P., B. Moss & G. Phillips (1996) "The determination of total nitrogen and total phosphorus concentrations in freshwaters from land use, stock headage and population data: testing of a model for use in conservation and water quality management." Freshwater Biology. 36: 451-473.
Kelly, M.G. and B.A. Whitton (1995) "The Trophic Diatom Index: a new index for monitoring eutrophication in rivers." Journal of Applied Phycology. 7: 433-444.
Pipp, E. & E. Rott (1993) Ecological evaluation of Austrian rivers by means of algal phytobenthos. Blaue Reihe 2, Ministry of Environment, Youth and Family, Vienna, Austria.
Rott, E. & H.C. Duthie (1997) "Diatoms for monitoring organic pollution and eutrophication in the Grand River, Ontario." in press.
Ter Braak, C.J.F. (1988) CANOCO - a Fortran program for canonical community ordination by (partial) (detrended) (canonical) correspondence analysis (version 2.1). Tech. Rep. No. LWA-88-02. Institute of Applied Computer Science, Statistical Department Wageningen, 6700 AC Wageningen, The Netherlands.
Ter Braak, C.J.F. (1990) CANOCO - a Fortran program for canonical community ordination. Microcomputer Power, Ithaca, N.Y.
Whitton, B.A., E. Rott & G. Friedrich (eds.) (1991) The use of algae for monitoring rivers. University of Innsbruck, Austria.