Nutrient inflows and levels in Lake Muhazi, Rwanda D. Usanzineza a *, I. Nhapi a, U.G. Wali a, J.J. Kashaigili b, N. Banadda c a WREM Project, Faculty of Applied Science, National University of Rwanda, Box 117 Butare, Rwanda Faculty of Forestry and Nature Conservation, Sokoine University of Agriculture, Box 3003 Morogoro, Tanzania c Uganda Industrial Research Institute, P.O. Box 7086, Kampala, Uganda b ABSTRACT Lake Muhazi, in Rwanda, has experienced a dramatic decrease in fish production since the nineteeneighties, typified by low water transparencies and high turbidity levels. The purpose of this study was to assess the water quality in Lake Muhazi and its tributary rivers, focusing on nutrients. Vertical and horizontal distributions of nutrients in the Lake were assessed and the concentration of nutrients in the tributary rivers was determined for the period July to September 2007. The parameters studied were nitrogen, phosphorous, chlorophyll a, pH, temperature and transparency . Four sampling stations were located within the Lake. The sampling was done fortnightly. The samples were collected at depths of 0.5 m, 2 m, 5 m, and 1 m from the bottom of the lake. Samples were taken using a Van Dorn Bottle water sampler and were preserved and analyzed in the laboratory for TP, TN and chlorophyll-a using Standard Methods. The temperature, transparency (Secchi disc), and pH were also measured. Samples were preserved and stored in cooler boxes. The results revealed that the nutrient levels in the lake are higher than those in the inflow rivers, indicating the possibility of having other sources of pollution than the inflow rivers. The Lake total nitrogen was found to be 0.85±0.22 mg/L, total phosphorus 0.29 ±0.15 mg/L, chlorophyll-a 18.1±10.5 μg/L, and Secchi disc 0.76±0.07 m. The results of this study are higher than those of previous studies. The majority of the measured parameters show that the lake is eutrophic, thus calling for urgent intervention to remedy the situation. Measures recommended include improved management of the catchment to reduce nutrient inflows into the Lake and the development of suitable discharge standards for all lakeshore activities. Keywords: Lake water quality, Lake Muhazi, land use , nutrient levels, nutrient distribution, trophic status 1. Introduction The most common way in which humans affect aquatic ecosystems is through altering nutrient dynamics (Boostma and Hecky, 1993). Most tropical African lakes are facing problems of rapid population growth in the riparian communities, which normally discharge pollutants into the lakes. This has led to the rapid deterioration of water quality in receiving lakes (Wandiga, 2003), and some lakes are experiencing a decrease in fish production. The surface water bodies in Rwanda are facing problems such as declining water levels (ADB, 2003) and infestation by water hyacinth and other aquatic weeds, which is a new phenomenon (MINITERE, 2005). In particular, Lake Muhazi has experienced a dramatic decrease in fish production since the nineteen eighties, and this may have been caused by bad fishing practices (e.g. use of small sized mesh nets, strike and poison fishing) (Mukankomeje et al., 1993). Sources of pollution include tourism development which is increasing around the Lake Muhazi, washing activities in tributary rivers of the lake and around the lakeshore, irrigation practice around the Kanyonyomba River (a tributary of Lake Muhazi), contaminated runoff from a disused mining site at Musha on the banks of Lake Muhazi, and burning of grass for cultivation around the lake. Erosion problems in the lake catchment are exacerbated by poor land cover and unsustainable land use practices. White and Downes (1977) argued that nutrient inflows into lakes promote eutrophication, so control of eutrophication usually means control of nutrient additions. In order to control eutrophication, it is * Corresponding author: Email address. usanzineza@yahoo.fr (D. Usanzineza), Tel. +25008454185 2 essential to conduct quantitative assessment of the important nutrient sources and sinks within the system, so that the effects of reducing a particular input may be assessed. Worldwide, various efforts have been tried to measure the trophic condition of water bodies. Some of these include Fish (1975) in New Zealand who produced a nutrient budget for Lake Rotorua, and the Ecology Division (1972) which took a complex modelling approach to assessing lake status also in New Zealand. While acknowledging the benefits of modelling for assessing nutrients in lakes, assembling data on nutrient loads is both tedious and expensive, and in the absence of resources, rapid assessment may provide a useful basis for developing a management plan for the lake and its catchment. This study is based on a preliminary assessment of nutrient levels in Lake Muhazi as there are no prior studies that could allow a more detailed assessment of the Lake. The purpose of this study was to investigate the nutrient inflows and status of Lake Muhazi by assessing vertical and horizontal distribution of nutrients in the Lake, and nutrient concentrations in the inflow rivers. This study focused only on the water column because nutrient concentration within the water column is important as it is from here that nutrients are taken up by phytoplankton which may then form blooms if excess nutrients are present. There was no equipment for collecting lake sediments. 2. Materials and methods 2.1 Study area Lake Muhazi is situated in the Eastern Province of Rwanda along the northern margin of the Rwamagana District. The location of the Lake is shown in Fig 1. With an altitude of 1,443 m, Lake Muhazi has a total catchment area of 829 km2, a surface area of 34.1 km2, depth of 14 m (max) and 10 m (average) and a volume of 330 x 106 m3. The maximum length of this lake is approximately 37 km with a mean width of 0.6 km (Ministry of Infrastructure, 2007). Lake Muhazi currently has an earth dam at its discharge point into the Nyabugogo River which was constructed in 1999 in an attempt to protect the Lake from drying up. Figure 1. Location of Lake Muhazi in Rwanda. 2.2 Sampling sites, sample collection and analysis The sampling study was conducted from July to September 2007. Samples were collected approximately every two weeks at four sites in Lake Muhazi, at fourteen sites on tributary streams, and at one site on an outflow river of Lake Muhazi (Nyabugogo River) as shown on Fig 2. A GPS was used to identify the 3 sampling sites. At each site in the lake, water samples were collected at depths of 0.5m, 2m, 5m from the water surface, and at 1m from the bottom of the lake. Lake Muhazi N Legend : Samplingsite on Rivers R14 R14 R15 R15 R15 R12 R12 R12 R13 R13 L4 L4 Dam Dam R4 R4 R11 R10 R10 R11 R11 R10 R4 R9 R9 R9 R3 R3 R3 R7 R7 R7 R2 R2 LL33 LakeMuhazi R1 L1 L1 L2 L2 R6 R6 R6 R8 R8 R8 R1 R1 R1 : Ntaruka R2 : Gashogoshogo R3 : Nyakambu R4 : Nyamarebe R5 : Gahurura R6 : Ingarani R7 : Kanyonyomba R8 : Kiruhura R9 : Njume R10: Buganya R11: Nyagatugu R12: Murama R13: Isereka R14: Mwange R15: Nyabuogo : Samplingsite in Lake R5 R5 R5 L1: Kavumu L2: Nyarubuye L3: Kibilizi L4: Rwesero Figure 2. The Lake Muhazi catchment and the location of sampling sites inside the Lake and on tributary rivers for points monitored July-Sept’07. The parameters measured were pH, temperature, Secchi disc depth, chlorophyll-a, total nitrogen (TN) and total phosphorus (TP). A survey of land use activities around the lake was also conducted. Samples were collected using a Van Dorn sampling bottle. The collected samples were stored in 560 ml plastic bottles and placed on ice in cooler boxes. The samples were carefully preserved and analysed according to Standard Methods (APPHA/AWWA/WEF, 2005). The TN and TP samples were unfiltered and digested with alkaline per-sulphate solution, whilst algae was filtered and extracted with ethanol. Secchi disc depths were determined in the field using a 20 cm diameter Secchi disc. The temperature and the pH were also measured in the field using HACH field testing kits. 2.3 Data analysis Descriptive statistics were used to analyze data in terms of mean, median and standard deviation using Microsoft Excel and SPSS software. The median was used to express the difference between data sets as the median is less sensitive to outliers (extreme scores) than the mean and thus a better measure than the mean for skewed distributions (Helsel and Hirsch, 2002). The mean was used for the comparison of data with the water quality guideline as it is considered representative of the whole sample. Its value depends equally on all of the data which may include outliers (Helsel and Hirsch, 2002). For the comparison of the difference between means, a one-way analysis of variance test was used. 3. Results and discussion 3.1 Characteristics of tributary rivers of Lake Muhazi The sampling of water from the tributary rivers was done four times during the period of July – September 2007 and the TN and TP average results are shown on Fig 3. 4 Figure 3. Concentration of TN and TP in water of inflow rivers, July-Sept’07. The concentrations of TP in the tributary rivers of Lake Muhazi ranged from 0.02 mg/L to 0.35 mg/L with the highest value at point R15 (Nyabugogo River). Point R15 is beyond the Lake and in an area where sugar cane cultivation is predominant. Point R1 also had high TP levels and this is attributed to the polluted runoff from Kayonza town. The concentration of TN ranged from 0.08 mg/L to 0.91 mg/L. Point R5 (Gahurura River) had the highest concentration of TN. This is probably due to its location in a large wetland which is exploited for rice cultivation. Similarly, Point R7 (Kanyonyomba River) had 0.43 mg/L concentration of TN and is also under rice cultivation. 3.2 Temperature and pH profiles of the water column The temperature profiles of Lake Muhazi at different sampling sites are shown in Fig 4. 21 22 Temp, oC 23 24 25 26 20 27 22 Temp, oC 23 24 25 26 27 20 0 2 4 1st sampling 2nd sampling 3rd sampling 6 8 10 12 Depth from water surface, m 0 Depth from water surface, m 21 2 4 1st sampling 2nd sampling 3rd sampling 6 8 10 22 Temp, oC 23 24 25 26 27 2 4 1st sampling 2nd sampling 3rd sampling 6 8 10 12 12 (a) Site L1 21 0 Depth from water surface, m 20 (b) Site L2 (c) Site L3 Figure 4. Temperature profiles at the 3 stations monitored in L. Muhazi, July-Sept’07 The data collected shows a difference in temperature of approximately 2°C between the surface and the bottom water, except for the temperature profile measured on the second sampling date where the difference was about 4°C at L1. In general, the water temperature decreases from the surface towards the bottom. The pH of the waters of Lake Muhazi is slightly alkaline (6.2 – 8.5) and did not fluctuate much as is shown in Fig 5. The mean pH of the lake was 7.8. At each sampling site, the pH decreased with the depth as shown in Fig X. The pH at the bottom was lower by 1 to 2 units compared with that of the surface. The pH was higher at the lake surface where high levels of the phytoplankton were observed. This phenomenon occurs mainly in mid-June and at the end of August, and is related to the assimilation of CO2 dissolved by the phytoplankton in water. The pH of the lake was higher as compared to those found in river. The average pH in the rivers varied between 6.1 and 7.1, and the average pH in the lake was 7.8. 5 pH 7 6 8 5 9 1st sampling 2nd sampling 3rd sampling Depth from water surface, m Depth from water surface, m 4 pH 7 8 5 9 6 8 10 12 1st sampling 2nd sampling 3rd sampling 2 4 6 8 10 2 4 8 9 1st sampling 2nd sampling 3rd sampling 6 8 10 12 12 (a) Site L1 pH 7 6 0 0 0 2 6 Depth from water surface, m 5 (c) Site L3 (b) Site L2 Figure 5. pH profiles at the 3 stations monitored in L. Muhazi, July-Sept’08. 3.3 Chlorophyll-a and transparency The chlorophyll-a concentration at various depths were also analyzed. The descriptive statistics of chlorophyll a concentration at various depths are presented in Table 1. Fig 6 shows the vertical and horizontal distribution of Chlorophyll-a in the lake. The average of the measurements taken is 18.1 µg/L (± standard deviation of 10.5 µg/L). It was observed that the chlorophyll-a concentration was high at 0.5 m depth, and the chlorophyll-a concentration increased toward the western part of the lake. The descriptive statistics of chlorophyll a concentration at various depths are presented in Table 4.3. The general average of the measurements taken is 18.1 µg/L (standard deviation equal to 10.5). Table 1. Descriptive statistics of chlorophyll-a in Lake Muhazi. Parameter Number of samples; N 21 Chlorophyll-a (µg/L) Valid N (listwise) Minimum Maximum Range Mean 1.7 37.1 35.4 18.1 21 30 10 22 Mean concentration of Chl a (µg/L) 20 21 20 19 18 17 16 0 15 0.5 m (a) Std. Deviation 10.5 2m Depth 5m L1 (b) L2 L3 L4 Lake sampling sites Figure 6. Average chlorophyll-a concentration in L. Muhazi, July-Sept’07 (a) Mean concentrations at various depths, and (b) Horizontal distribution. 6 Previous studies noted that the maximum concentration of chlorophyll-a, occurred between 1 m and 2 m in depth (Plisnier, 1989; Mukankomeje et al., 1993). This is explained by an optimum in quality (frequency) and quantity (lux) of the light for the phytoplankton. The average Secchi disc transparency results for the three sampling runs done during this study were 0.87 m at L1, 0.83 m at L2 and 0.77 m at L3 (Fig 7). Figure 7. Variation of transparency in Lake Muhazi. Fig 7 shows that the transparency at L1 was nearly always higher by 0.10 m compared to that measured at L3. In 1986 and 1987 at Karambi, not far from L1 on the eastern part of the Lake Muhazi, the average of monthly measurements was 0.81 m. The transparency at Karambi was also nearly always higher by 0.10 m than that measured at Giheta, not far from L3 on the western side of the lake in 1986 and 1987 (Plisnier, 1989). This was the same case during this research. In this lake, the change of the transparency is explained by the change in the phytoplankton levels which are lower when the transparency is high (Plisnier, 1989). It is also important to note that the transparency seems to have slightly increased since the study of Mukankomeje (1993). Mukankomeje (1987) found an average transparency of 0.65 m compared to 0.76±0.07 m from the current study. Perhaps this corresponds to a decrease in the content of sediments coming from erosion since the enactment of the Organic Law of 2005 determining the modalities of protection, conservation and promotion of the environment in Rwanda. The building of the dam in 1999 at the outlet of the lake has also increased the lake level thereby diluting the concentration of suspended material in the lake and also proving a longer hydraulic retention time for the sediments to settle down at the lake bottom. 3.4 Nutrients profiles and their distribution in Lake Muhazi The nitrogen and phosphorus profiles in the Lake are shown in Figs 8 and 9. The nutrient levels in Lake Muhazi ranged from 0.47 to 4.78 mg/L for TN and from 0.08 to 1.02 mg/L for TP. The average TN levels of 0.26 mg/L while TP levels of 0.09 mg/L in inflows rivers seem to be low compared by those found in the lake TN (0.85±0.22 mg/L) and TP (0.29±0.15 mg/L) . The results show that nutrient concentrations are above the recommended concentrations of 0.3 mg/L TN and 0.01 mg/L TP for oligotrophic systems (Mandaville, 2000). Assuming an algal cell composition of C106H180O45N16P (Lee and Jones, 1986), a stoichiometric N:P ratio of 1:7 can be derived. The average N:P ratio in Lake Muhazi is 1:3, suggesting nitrogen is the limiting nutrient. However, it will be misleading to make firm conclusions based on this ratio alone as both nutrients are present in abundance in the lake. Other factors in the pelagic environment like zooplankton grazing, light or micro-nutrients could be the limiting factor for algae or aquatic plant growth in the lake. The abundance of P in the lake could also explain why the bulk of it is not contained in algae as would be normally expected for a productive lake. Lake Muhazi is now eutrophic according to the OECD Fixed Boundary System or Diagnostic Model (OECD, 1982) (TP 100 mg/m3; mean chlorophyll-a 25 mg/m3; mean Secchi disc depth 1.5 m). 7 TN, mg/L 0.4 0.9 TN, mg/L 1.4 1.9 0.4 TN, mg/L 1.4 0.4 1.9 4 6 8 10 12 1st sampling 2nd sampling 3rd sampling 2 4 6 8 10 12 (a) Site L1 0.9 1.4 1.9 0 Depth from water surface, m 1st sampling 2nd sampling 3rd sampling 2 Depth from water surface, m Depth from water surface, m 0.9 0 0 2 4 6 8 1st sampling 2nd sampling 3rd sampling 10 12 (b) Site L2 (b) Site L3 Figure 8: Vertical profile of TN at the 3 Lake sampling sites monitored July-Sept’08 0.1 TP, mg/L 0.3 0.4 0.2 0.5 0.6 0.1 0.2 TP, mg/L 0.3 0.4 0.5 0.6 0.1 0 4 1st sampling 2nd sampling 3rd sampling 6 8 10 2 4 1st sampling 2nd sampling 3rd sampling 6 8 10 TP, mg/L 0.3 0.4 0.5 0.6 2 4 1st sampling 2nd sampling 3rd sampling 6 8 10 12 12 12 0.2 0 Depth from water surface, m 2 Depth from water surface, m Depth from water surface, m 0 (c) Site L3 (b) Site L2 (a) Site L1 Figure 9: Vertical profile of TP at the 3 Lake sampling sites monitored July-Sept’08 3.5 Relation between Secchi disc, Chlorophyll-a and Nutrients Most eutrophication models are based on empirical relationships between abiotic variables. Therefore, a series of regressions were performed considering abiotic variables commonly used in eutrophication studies. The TP and TN were used as independent variables. Both TN and TP did not show a significant and linear relationship with chlorophyll-a as was expected (Figure 10). The coefficient R2 for this regression was 0.0876 for Chlorophyll-a versus TN, and 0.0071 for Chlorophyll-a versus TP. The regressions analysis were also performed between TP, TN and the Secchi disc measurements and results showed that both TN and TP had no significant and linear relationship with the Secchi disc measurement. The coefficient R2 for these regressions were, respectively, 0.1959 for Secchi disc versus TN, and 0.1104 for Secchi disc versus TP. Thus, this set of regressions could not confirm that TP and TN could be good descriptors of the trophic status of Lake Muhazi. y = -0.0515x + 58.749 R2 = 0.0876 40 35 35 30 30 Chl a (µg/L) Chl a (µg/L) 40 25 20 15 25 20 15 10 10 5 5 0 y = 0.0076x + 15.213 R2 = 0.0071 0 0 200 400 600 800 1000 1200 0 TN (µg/L) (a) 100 200 300 400 500 600 700 TP (µg/L) (b) Figure 10: Linear regression between mean concentrations of (a) TN and (b) TP and mean concentrations of chlorophyll-a for L. Muhazi, July to Sept’07 8 90 86 y = -0.039x + 112.03 R2 = 0.1959 84 Secchi (cm) 85 Secchi (cm) y = 0.0125x + 75.282 R2 = 0.1104 88 80 75 82 80 78 76 74 70 72 65 600 70 650 700 750 (a) 800 850 900 950 TN (µg/L) 1000 0 (b) 100 200 300 400 500 600 700 TP (µg/L) Figure 11: Linear regression between mean concentrations of (a) TN and (b) TP and mean Secchi disc measurements for L. Muhazi, July to Sept’07 This study could only show the situation in Lake Muhazi in as far as nutrients in the water column are concerned. It would have been ideal to have major inflows gauged so that the response of the Lake to nutrient loadings could be established. This would also required a year-round monitoring study which was not possible with the resources available. Nutrient levels in the sediments would also be interesting to monitor as they could give the extent to which the lake is already affected. 4. Conclusions and recommendations 4.1 Conclusions From the results of this study, the following conclusions were drawn: 1) The horizontal and vertical nutrient profiles of Lake Muhazi show a lot of variability, making it difficult to rely on the observed nutrient concentration profiles to understand the nutrient dynamics in Lake Muhazi. Overall, the TN concentration averaged 0.85±0.22 mg/L, and TP averaged 0.29±0.15 mg/L, Secchi disc averaged 0.76±0.07 m, and chlorophyll-a averaged 18.1±10.5 µg/L. The lake is now eutrophic. 2) The levels of nutrients in the inflow rivers of Lake Muhazi were considered generally low and always below 1 mg/L for both TN and TP. This could be due to the buffering capacity of the wetlands around each inflow river. 4.2 Recommendations From the results and discussions in this report, the following recommendations were derived: 1) It is crucial and urgent to identify all controllable sources of pollution (e.g., poor soil and land management in agriculture, domestic sewage disposal practices, etc) throughout the watershed of Lake Muhazi and use this information to develop appropriate pollution reduction measures. 2) Immediate action could be taken to reduce major pollution sources such as improving wastewater disposal systems, the use of phosphorus detergents in the lake and its catchment, and rehabilitation of disused mining sites. 3) Long-term water quality monitoring and river flow gauging is required for at least a year to cover seasonal fluctuations. 4) In future studies, it is necessary to measure dissolved oxygen profiles so as to better understand nutrient transformations in the lake. 5) An investigation on the sudden increase in nutrients towards the lake outlet is needed. 6) An estimation of nutrient input from other sources (communities and industries around the lake) or from atmospheric deposition would be also interesting so that a comparison with the input from rivers could be done. Acknowledgements 9 This research was generously supported by funding from the Water Resource and Environmental Management Project in the Faculty of Applied Science and analytical work was conducted at the Faculty of Science Laboratory, National University of Rwanda. References ADB (African Development Bank) (2003). Appraisal report, inland Lakes integrated development and management support project (paigelacpaigelac), Republic of Rwanda. http://www.afdb.org/pls/portal/docs (Accessed December, 2006). APPHA/AWWA/WEF (2005). Standard Methods for the Examination of Water and Wastewater. 21st Edition. American Public Health Association, American Water Works Association and Water Environment Federation, Washington, DC. Bootsma, H.A. and Hecky, R.E. (1993). Conservation of the African Great Lakes: A Limnological Perspective. Conservation Biology. Vol. 7(3): pp 645-656. Chapman, D. (1996). Water quality assessment: A guide to the use of biota, sediments and water in environmental monitoring. Second edition. Chapman & Hall, London. Ecology Division (1972). The prospects and aims of making a rapid assessment of the trophic status of New Zealand lakes. An unpublished report from Freshwater Section to the Officials Committee on Eutrophication. New Zealand DSIR, Auckland. Fish, G.R. (1975). Lakes Rotorua and Rotoiti, North Island, New Zealand, their trophic status and studies for a nutrient budget. Fisheries Research Bulletin 8. Ministry of Agriculture and Fisheries, New Zealand. Helsel, D.R. and Hirsch, R.M. (2002). Hydrologic Analysis and Interpretation: Statistical Methods in Water Resources. Site: http://water.usgs.gov/pubs/twri/twri4a3/ (Accessed October, 2007). International Small-hydro Atlas (2005). Rwanda Water Resources. Internet site: http://www.smallhydro.com/index.cfm?Fuseaction=countries.country & Country_ID=130 (Accessed May, 2007). Kruis, F. (2005). Environmental Chemistry: Selected Analytical Methods. UNESCO-IHE Lecture note LN0188/05/1, Delft, The Netherlands. Lee, G.F., and Jones, R.A. (1986) Detergent Phosphate Bans and Eutrophication, Environ. Sci. Technol. Vol. 20(4), p330-341. Lim, N.C. (2001). Assessment of reservoir water quality and its application to reservoir management in the central plains. MSc Thesis, University of Kansas. Internet site: http://www.kbs.ku.edu/people/assets/staff_cv/lim_thesis.pdf (Accessed May, 2007). MANDAVILLE, S. M. (2000). Lake Data of Relatively Undisturbed lakes within Lava Scotia: Provincial Averages. Canadian Council of Ministers of the Environment, Health and Welfare Canada and Ontario Ministry of the Environment. Internet site: http://www.chebucto.ns.ca/science/swcs/swcs. htmI. Accessed February 2009. Ministry of Infrastructure (2007). Rwamagana, Rwanda Conceptual Master Plan. Government of Rwanda, Kigali, Rwanda. MINITERE (Ministry of Land, Environment, Forestry, Water and Mines) (2005). The United Nations Framework Convention on Climate Change: initial national communication Internet site: http://unfccc.int/resource/docs/natc/rwanc1.pdf (Accessed May, 2007). Mukankomeje, R., Plisnier, P.D., Descy, J.P. and Massaut, L. (1993). Lake Muhazi, Rwanda: limnological features and phytoplankton production. Hydrobiologia 257: pp107-120. OECD (Organisation for Economic Co-operation and Development) (1982). Eutrophication of Waters: Monitoring, Assessment and Control. OECD Co-operative program on monitoring of inland waters (Eutrophication control), Environment Directorate, OECD, Paris, 154pp. Plisnier, P.D. (1989). Etude Hydrobiologique et Développement de la Pêche au Lac Muhazi (Bassin de l’Akagera, Rwanda) (in French) (Hydrobiologic Study and Fishing Development at Lake Muhazi (Akagera Catchment, Rwanda)). Final report (1986-1988). Namur, Belgium. Wandiga, S.O. (2003). Lake Basin Management Problems in Africa: Historical and Future Perspectives. Internet site: http://www.povertyenvironment.net/ (Accessed May, 2007). White, E and Downes, M.T. (1977). Preliminary Assessment of Nutrient Loads on Lake Taupo, New Zealand. N.Z. Journal of Marine and Freshwater Research. Vol. 11(2): p341-356.