Identifying Leakage to Groundwater from Lake Rerewhakaaitu Using Isotopic and Water Quality Data

Journal of Hydrology (NZ) 47 (2): 85-106 2008 © New Zealand Hydrological Society (2008)

Identifying leakage to groundwater from Lake Rerewhakaaitu using isotopic and water quality data

Robert R. Reeves1, Uwe Morgenstern2, Christopher J. Daughney2, Michael K. Stewart3 and Dougall Gordon4 1 GNS Science, Wairakei Research Centre, Private Bag 2000, Taupo 3352, New Zealand. Corresponding author: R. [email protected] 2 GNS Science, Avalon, P O Box 30368, Lower Hutt 5040, New Zealand 3 Aquifer Dynamics Ltd, 20B Willoughby St, Lower Hutt 4 Environment Bay of Plenty, PO Box 364, Whakatane 3158, New Zealand

Abstract Introduction Stable isotopes (18O, 2H), water dating, and The quality and quantity of recharge to a major anion and cation water chemistry groundwater system can affect the sustainable of springs, groundwater bores, a stream use of groundwater resources, particularly and a lake have been used to identify areas where groundwater is used for domestic or of discharge from Lake Rerewhakaaitu. public water supplies, because public health Groundwater aquifers and spring discharges can be affected by contamination. Thus, to the north and west of Lake Rerewhakaaitu identification and safe management of the are found to contain up to 79% lake water. source of groundwater recharge is essential Hierarchical cluster analysis is used to for protection of groundwater resources. group sites into three clusters based on Natural sources of groundwater recharge to the major anion/cation water chemistry. the groundwater aquifers include streams, Defining meaningful hierarchical clusters was rivers, lakes and rainfall. difficult due to the similarity of the chemical Groundwater is used extensively in New compositions between sites. The three Zealand for agricultural, horticultural, clusters defined in this study do not clearly domestic and public water supplies. White identify the sites containing lake water. (2001) estimates up to 26% of the New However, the hierarchical cluster analyses can Zealand population depend totally on be used to infer sites that are likely to contain groundwater for drinking, and a further 24% proportions of lake water in the groundwater, are partially dependent on groundwater as a and therefore provides a method to identify source of water. sites that may contain lake water. Identifying the source of the water to a The groundwater and spring discharges to groundwater system commonly requires the the north and west of Lake Rerewhakaaitu use of chemical or isotopic tracers (Quast et flow into a different surface water catchment, al., 2006), usually, in conjunction with other thus providing a hydrological link between techniques such as water chemistry and/or two different surface water catchments. This physical hydrology. Chemical and isotopic has implications for catchment management tracers are particularly useful in areas where in the greater Rerewhakaaitu area. hydrological information important to

85 interpreting the data, such as bore geology, the nutrient cycle in the Rotorua Lakes area. is missing. This will assist EBOP to form new policies Water chemistry, combined with the and develop remediation programmes where naturally occurring stable isotopes oxygen-18 necessary. and deuterium, have been used successfully Nutrient levels, nutrient inputs, nutrient to distinguish between rainfall, local transport mechanisms, nutrient transport groundwater and lake water as the source of times, lake processes, and effects of land water in geothermal systems (e.g., Stewart, use are key scientific aspects that need to 1978; Darling et al., 1996; Ojiambo et al., be better understood to solve the issues. To 2001) and groundwater systems (e.g., Stewart date, much scientific work in the Rotorua and Morgenstern, 2001; Ojiambo et al., Lakes area has focused on Lakes Rotorua and 2001; Chowdhury, 2004; Kristmannsdóttir Rotoiti (e.g., Lakes Water Quality Society and Ármannsson, 2004). Inc, 2003; Hamilton, 2003; Morgenstern Determining the mean residence time of and Gordon, 2006), in line with the priorities groundwater can significantly enhance the set by EBOP. The work reported in this paper understanding of a groundwater system. The may assist EBOP develop policies in the Lake mean residence time can be taken as the average Rerewhakaaitu area, where understanding of travel time for land-use effects to be observed the groundwater system is more limited. and groundwater recharge times, and can be The objective of this study was to identify used to determine the percent age fractions whether lake water is a source of recharge to of recharge water to a groundwater aquifer. the major aquifers around Lake Rerewhakaaitu Tritium and chlorofluorocarbons (CFCs) are and, if it is occurring, what the flow paths two tracers that have been used extensively are. The study was carried out near Lake in New Zealand to understand groundwater Rerewhakaaitu using a combination of water systems (e.g., Taylor, 1994; Stewart et al., chemistry, stable isotopes and water dating 1999; Morgenstern and Gordon, 2006; techniques. Morgenstern et al., 2004; Morgenstern and Stewart, 2004; Morgenstern, 2005). Lake Rerewhakaaitu is one of twelve lakes Setting in the Rotorua area (the ‘Rotorua lakes’) Lake Rerewhakaaitu is situated approximately monitored regularly by Environment Bay 26 km southeast of Rotorua, New Zealand of Plenty (EBOP). Regular cyanobacterial and covers an area of approximately 630 ha. blooms in Lakes Rotorua and Rotoiti since The surface water catchment is approximately 1997 have heightened public concerns 5300 ha in area, with elevations ranging from about the declining water quality of these 435 m above sea level (asl) at the lake to two lakes, and possibly in the other nine approximately 1100 m asl on Mt Tarawera Rotorua Lakes (Lakes Water Quality Society, in the north of the catchment (Fig. 1). The 2001). Nutrient inputs (nitrogen and elevation of the top of the surface water phosphorus), their effects on the lakes, and catchment to the east, south and west of how the lakes can be managed to maintain Lake Rerewhakaaitu is between 420 m and or improve their current water quality are 530 m asl. The surface water catchment is issues requiring study. In response to the mainly to the south of the lake, with narrow public concern, EBOP has been coordinating catchment areas to the north, east and west. research programmes involving iwi, regional Mean rainfall is about 1600 mm per year. government, local government and scientific Volcanic processes in the last 22 ka have organizations to improve understanding of moulded the landscape, resulting in the

86 formation of many of the lakes found in the McIntosh et al. (2001) estimate 77% of the region. Mt Tarawera is made up of several catchment is pasture, 15% exotic forest, 6% Okataina rhyolitic domes ranging in age indigenous forest/scrub and 2% ‘other’. Of from 0.7 ka to 21 ka. Three recent (<11 ka) the 77% pasture, 70% is used for dairy cows deposits around Lake Rerewhakaaitu are and 7% for sheep/cattle. Dairying and use of attributed to eruptions from Mt Tarawera fertilizers has intensified between 1990 and (Nairn, 2002): Waiohau pyroclastics (11 ka); 2000, with relatively high application rates Kaharoa pyroclastics (0.7 ka); and mud and (>300 kg/ha/year) of nitrogen-based fertilizers ejecta from the 1886 eruption. The thickness (McIntosh et al., 2001). of the Waiohau pyroclastic deposit is in excess Lake Rerewhakaaitu is shallow, having a of 30 m, while the thickness of the Kaharoa mean depth of 6.3 m and a maximum depth pyroclastic deposit is between 10 and 30 m of 15 m. The lake receives water from rainfall, thick in the area around Lake Rerewhakaaitu streams and shallow groundwater systems (White et al., 2003). The thickness of ejecta (White et al., 2003). The lake is described from the 1886 eruption is too small to be as a mesotrophic lake in average condition mapped except near vent zones and the shores (Scholes and Bloxham, 2005). Mean total of Lake Rotomahana (Nairn, 2002). nitrogen and mean total phosphorus (as P) Lake Rerewhakaaitu is formed in an old are 0.38 g m–3 and 0.008 g m–3 respectively deep narrow valley cut into the Rangitaiki for the period 1991–2001. No net trend Ignimbrite (White et al., 2003). The valley is observed in the total nitrogen and total is thought to have drained a large catchment phosphorus data over this period. However, to the south of the current lake, northwards an increase in nitrogen and phosphorus to the Tarawera River. Infilling and damming occurred between 1995 and 1997. McIntosh of the valley probably occurred during the et al. (2001) suggest that this increase could 11 ka eruption of Mt Tarawera (Waiohau be due to an increased lake level mobilizing pyroclastics) and probably caused a small nutrients in the sediment of the lake margin. lake to form. The lake reached its present form with further damming from the 0.7 ka Mt Tarawera eruption (Kaharoa pyroclastics). The Rangitaiki Ignimbrite is considered to be the base unit in this study and is described by Nairn (2002) as a moderately welded, dark grey, crystal-rich tuff. The unit includes coarse tuffs, pumice breccias and air fall deposits. It is the oldest unit (0.34 Ma) mapped in the Rerewhakaaitu area. The unit outcrops to the east and in the steeply-incised valleys to the west of Lake Rerewhakaaitu. Land use in the catchment is dominantly dairy farming. Figure 1 – Map of the Lake Rerewhakaaitu area.

87 Mangakino Stream and Awaroa Stream east side of the lake (White et al., 2003). A are the only two streams that flow into Lake small groundwater flow from the Rangitaiki Rerewhakaaitu. Both streams flow into a Ignimbrite aquifer could be entering Lake wetland at the southern end of the lake, Rerewhakaaitu from the northeast. An and both streams have a shallow groundwater assessment of groundwater flow directions aquifer as their source. Mangakino for the other aquifers is difficult to determine Stream has a mean flow of 24 l s–1 (from due to a lack of data. Lake Rerewhakaaitu measurements between 1995 and 2001) is estimated to be losing water at a rate of measured approximately 1 km upstream from 556 l s–1 to groundwater aquifers (White the lake. Total nitrogen concentrations have et al., 2003). Most of this is probably going increased from a median of 1.21 g m–3 in the into the Rangitaiki Ignimbrite aquifer; 1970’s to 2.36 g m–3 in 2000 (McIntosh et however, other smaller aquifers need to be al., 2001). Total phosphorus concentrations considered. have remained at about 0.045 g m–3 over A number of small seeps and springs are this period. Awaroa Stream is commonly dry. observed near the base of gullies and depressions Water samples collected from Awaroa Stream near the northern and northwestern parts of in the 1970s and the 1990s show both Lake Rerewhakaaitu. One of these springs (Te total nitrogen and total phosphorus have Kaue Spring) contains significant amounts approximately tripled (median total nitrogen of lake water (Taylor et al., 1977) based on to 8.43 g m–3 and median total phosphorus stable isotope measurements. White et al. to 0.681 g m–3) over this period (McIntosh (2003) speculates that Lake Rerewhakaaitu et al., 2001). may also be losing water through a subsurface The only surface water outlet to Lake flow along the pyroclastic buried valley cut Rerewhakaaitu is an upper tributary of into the Rangitaiki Ignimbrite flowing under Mangaharakeke Stream that flows from the the lake to the north. southeast corner of the lake to the Rangitaiki River. This tributary only flows when the lake level is high. Methods All major geological units in the Lake GNS collected water samples from existing Rerewhakaaitu catchment, except the groundwater bores, springs, streams and the Okataina Rhyolites, are identified as having lake from the Lake Rerewhakaaitu study area groundwater aquifers. The Rangitaiki in January-February 2006. Grab samples were Ignimbrite is the main groundwater aquifer obtained for spring, stream and lake samples. that is used to supply water for domestic and Groundwater bores were purged until the farming purposes. Lake Rerewhakaaitu is a field parameters temperature, conductivity, possible source of recharge to the Rangitaiki pH, and turbidity were stable prior to Ignimbrite aquifer because groundwater sampling. Bore samples were collected as levels in the Rangitaiki Ignimbrite aquifer close to the bore-head as possible. The field are generally lower than the level in Lake parameters were recorded immediately prior Rerewhakaaitu (White et al., 2003). to sampling. Water samples were collected in Groundwater flow directions are variable bottles appropriate for the required analyses in the Rangitaiki Ignimbrite aquifer near using appropriate methods (Rosen et al., Lake Rerewhakaaitu. Lake Rerewhakaaitu 1999; van der Raaij, 2004 pers. comm.). Water is situated on a groundwater divide, with samples were filtered (0.45 micrometer pore groundwater flowing to the west on the size), acidified and chilled in the field where west side of the lake and to the east on the required.

88 Water samples were analysed for alkalinity anions. In general, this left between one and (HCO3) and pH measured on a Metrohm three samples that could be used to calculate autotitrator, magnesium (Mg), Calcium (Ca), a median at a site. Censored data is given the iron (Fe), manganese (Mn), potassium (K), value of zero in the ion balance calculation. silica as SiO2 (SiO2) and sodium (Na) using Medians of the samples at each site were a Thermo Jarrell Ash inductively-coupled calculated with Microsoft Excel to analyse the plasma optical emission spectrometer, chloride spatial distributions of the water chemistry (Cl), sulphate (SO4), nitrate as nitrogen and to use in the hierarchical cluster analyses. (NO3-N), bromide (Br), fluoride (F) and Several data processing steps were required dissolved reactive phosphate as phosphorus to obtain medians. Firstly, censored data (DRP) using a Dionex ion exchange was converted into values by using half the chromatograph, ammonia as nitrogen (NH3- detection limit. Where multiple detection 18 N) using an auto analyzer, oxygen 18 ( O) limits existed (e.g., Br, NO3-N), the highest and deuterium (2H) using a mass spectrometer, detection limit was used. Results were tritium using low-level liquid scintillation changed to half the highest detection limit for spectrometry, and chlorofluorocarbons samples with results less than half the highest (CFCs) using a gas chromatograph. detection limit. This was done to ensure data Data collected by GNS were augmented all had the same lower limit. with data collected by EBOP in the Lake Rerewhakaaitu area over the last three years. Hierarchical cluster analysis Not all water samples collected by EBOP Differences in water chemistry within an were analysed for the major cations/anions aquifer and from different aquifers can be above as some surveys were designed to used to identify the factors above using investigate nutrient levels only. The sampling hierarchical cluster analysis (Daughney and methodology or methods of the EBOP Reeves, 2005; Guler et al., 2002). Hierarchical analyses were not known. Data supplied by cluster analysis is a multi-variate statistical EBOP has been modified to be consistent technique used to ‘group’ samples based on with the GNS chemical parameters to enable any number of variables (Back, 1961, 1966; the data to be used for statistical processing. Morgan and Winner, 1962; Seaber, 1962). The modifications were: The groups are called hydrochemical facies, • Alkalinity as HCO3 has been calculated and represent groundwaters with similar from Alkalinity (as CaCO3). chemical composition. • Soluble iron is used as an indicator for Water quality data was processed for the iron. hierarchical cluster analyses as described • Soluble manganese is used an indicator for by Daughney and Reeves (2003) using the manganese. Statgraphics software package. Medians • Dissolved reactive phosphorus (DRP) is for each analyte were transformed (where used as an indicator for phosphorus. required) and tested for normality and Ion balances were calculated for all water homoscedasticity using the Kolmogorov- samples to ensure data quality. Only samples Smirnov test (p=0.1). Once a suitable with ion balances (Freeze and Cherry, 1979) transformation had been selected, all within +/– 10% were retained for hierarchical medians were scaled according to their cluster analysis. This reduced the number of z-scores to ensure that each analyte imparted samples that could be used from the dataset. a similar degree of influence to the clustering It also restricted samples that could be used to algorithm (Helsel and Hirsch, 1992; Guler those that contained major cations and major et al., 2002).

89 A hierarchical cluster analysis algorithm The isotopic composition of the water using the nearest neighbour linkage rule, sample from Mangakino Stream (site with the square of the Euclidean distance as BOP120101) plots close to the meteoric a measure of similarity, was used to identify water line, suggesting the stream source is residual samples. Residual samples were either rainwater or a shallow groundwater identified and removed from the dataset. The aquifer with an isotopic ratio similar to cluster analysis algorithm using Ward’s linkage rainfall. The isotopic concentration at this with the square of the Euclidean distance as site is considered to represent the second a measure of similarity was then used on the end member in a simple rainfall – lake water data without the residuals. The hierarchical mixing model. cluster analysis produces a membership list Sites Lake Rerewhakaaitu, BOP210218, (site and cluster), a dendrogram (graphical BOP180476, BOP180417, 76, BOP180489, method of visualising cluster memberships), Medium Spring and Te Kaue2 Springs all and centroid values (averages) for each show enriched isotopic concentrations relative cluster. Centroid values are reduced back to to the meteoric water line. All these sites occur ‘real’ values by using the opposite function on the western side of Lake Rerewhakaaitu. used to transform the data and compared to The isotopic ratio of Site BOP210218 is the the water dating, stable isotope, geology and closest to Lake Rerewhakaaitu (Fig. 2). This hydrological data. site is probably the most strongly influenced by the lake water as a source of recharge. Site Results BOP210218 is a spring situated near the Stable Isotopes northwest corner of Lake Rerewhakaaitu. Oxygen-18 and deuterium results range from All sites with enriched isotopic –1.56 ‰ to –7.2 ‰ and –9.1 ‰ to –41.9 ‰ concentrations relative to the meteoric water respectively (Table 1, Table 2). Groundwater line plot close to a mixing line (Fig. 2) between sourced from infiltrating rainfall is expected water from Mangakino Stream and water to fall on the meteoric water line (Fig. 2) from Lake Rerewhakaaitu. This suggests a appropriate for the district (given by the lake component of water in the groundwater equation δ2D = 8 × δ18O + 13, Stewart and at these sites. A simple mixing model using Morgenstern, 2001). isotope mass balances (Cook and Herczeg, The isotope data clearly indicate that 1999) can be used to estimate the percentage water from Lake Rerewhakaaitu is different from most of the groundwater collected from around the lake (Fig. 2). The δ18O and δ2H ratio is consistent with isotopic enrichment caused by evaporation from the lake surface. This gives the water a unique isotopic signature and allows the water to be used as a groundwater recharge ‘tracer’. Stewart and Taylor (1981) demonstrate good vertical mixing of δ18O in Lake Rerewhakaaitu, making the lake water a homogeneous source of δ18O to the groundwater recharge. The isotopic concentration of the lake represents Figure 2 – Oxygen-18 verses deuterium with the the first end member of a simple rainfall – mean residence times (MRT) of samples in lake water mixing model. brackets.

90 of the lake water that contributes to the collected and analysed by GNS. Data was groundwater at each site (Table 3). Average sampled between March 2003 and February contributions of lake water are estimated to 2006. Not all sites were sampled in each range from 79% at site BOP210218 to 13% sampling round, and the suite of parameters at site Te Kaue2 Springs. analysed differed between sampling rounds. All other sites have similar isotopic In general, median water quality of the concentrations and plot close to the meteoric groundwater and streams in this study is good water line (Fig. 2). These sites probably have compared to the New Zealand Drinking Water rainfall recharge as their main source of Standards (NZDWS) (Ministry of Health, groundwater recharge, with little or no input 2005) (Table 1). Median manganese (both from the lake. total and dissolved) is the only parameter above the Drinking Water Standards Water age dating maximum allowable value (0.4 g m–3). This Eleven sites were sampled for tritium and/or occurs at sites BOP180451, BOP180489 CFCs. Mean residence times were calculated and BOP210218. All three sites occur to the for each sample (Table 1). Residence times northwest of Lake Rerewhakaaitu and also range from 1 year at site BOP120101 contain high median concentrations of both (Mangakino Stream) to 145 years at site total and dissolved iron, and trace amounts of BOP180486. The majority of the mean ammonia. This demonstrates that anaerobic residence times are greater than 30 years. conditions are present in the groundwater Sites BOP210218 and BOP180489 have aquifer in this area. ambiguous mean residence times of 4/44 Median dissolved iron concentrations range years and 7/40 years respectively. Ambiguous from below the detection limit (<0.02 g m–3) residence times can occur due to tritium to 13.2 g m–3. In most of these cases, a high concentrations falling within a specific median dissolved iron concentration is assoc range, coinciding with duplicate atmospheric iated with a high median dissolved manganese tritium concentrations occurring either concentration. side of the ‘tritium bomb peak’ (Stewart Median NH3-N concentrations range and Morgenstern, 2001). CFC data cannot from below the detection limit (<0.01 g m–3) be used reliably to identify the ‘correct’ to 62.5 g m–3. Only nine of the 52 sites with mean residence time, as both sites have NH3-N data have median NH3-N strongly anaerobic waters. Anaerobic water concentrations above the detection limit. can degrade CFC concentrations, making Site 76 has the highest median NH3-N a residence time interpretation based on concentration of 62.5 g m–3. The median CFCs at these sites impossible. The youngest NO3-N concentration at this site is mean residence times at sites BOP210218 0.052 g m–3. Low dissolved oxygen, low and BOP180489 have been used in this oxidation-reduction potential, high median paper based on hydrological and chemical ammonia, high median iron, high median conditions (e.g., hierarchical cluster analyses) manganese and low median nitrate are con at these sites. ditions indicating a reducing groundwater environment at this site (Table 1). Major anion and cation chemistry Median nitrate-nitrogen concentrations Major anion and cation concentrations used range from below detection (0.02 g m–3) in this work have been collated using data to 10.17 g m–3. The six sites with median from three sampling rounds sampled and nitrate concentrations below the detection analysed by EBOP, and one sampling round limit can be divided into three areas: 1) sites

91 –3 F 0.095 0.025 0.032 0.16 0.03 0.18 0.11 0.16 0.09 0.14 0.175 0.08 0.075 0.075 0.035 0.13 0.085 g m –3 TP g m 0.052 0.0535 0.046 0.022 0.022 0.076 0.257 0.0205 0.015 0.0095 0.194 0.026 –3 DRP g m 0.044 0.079 0.022 0.073 0.039 0.054 0.007 0.006 0.081 0.264 0.026 0.0155 0.011 0.022 0.197 0.0245 –3 4.9 8.4 3.7 8.3 4.5 6.3 9.1 6.5 2.7 8.9 7.9 2.9 6.7 Cl 11.5 10.3 13.6 12.3 g m –3 Ca 6.6 5.4 5.92 6.65 5.3 2.43 3.4 4.11 6.39 3.49 2.395 9.27 7.8 3.25 5.975 g m 10.045 10.75 –3 Br <0.1 <0.1 <0.1 <0.1 0.36 <0.1 <0.1 g m -N –3 3 0.13 <0.01 62.5 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 g m NH

pH 6.25 7.58 7.11 6.52 5.95 6.3 6.28 6.8 6.075 6.2 6.65 5.875 5.7 6.1 6.07 6.8 6.2 3 O and Mean Residence times are from a single sample from times are Residence O and Mean –3 18 δ 40 28 66 24 23.9 33 42.8 23.6 29.3 28.25 30.3 26.1 31.55 32 38 28.35 g m HCO 333 H, 2 D – Dissolved Fraction D – Dissolved Fraction Total T – δ –3 6.6 1.1 9.5 2.1 8.8 2.1 8.05 7.5 6.8 8.5 7.25 5.4 6.9 8.8 9.9 DO g m 10 C 7.4 9.2 9.7 7 7.8 6.9 o 13.9 14.4 15 14.15 13.2 14.85 14 17.6 13.3 15 13.35 Temp –2 540 480 477 538 514 492 mV ORP

79 96 90 66.5 79 EC 584 129 121 140 145.5 100 186 236 136.5 180 123.5 169 0.9 3 5.5 1.7 4.5 71 10.1 NTU Turbidity

1 1 4 4 3 1 4 2 2 3 2 2 2 2 3 2 2 Points EC – Electricial Conductivity (uS/cm @ 25°C) EC – Electricial potential ORP – Oxidation reduction Oxygen DO – Dissolved above NZ MAV

– Chemical and isotopic results for waters from the Lake Rerewhakaaitu catchment and environs the Lake Rerewhakaaitu for waters from – Chemical and isotopic results Site above NZ guide line value Notes 70 76 BOP120101 BOP180417 BOP180445 BOP180448 BOP180449 BOP180451 BOP180454 BOP180456 BOP180457 BOP180458 BOP180459 BOP180460 BOP180462 BOP180463 BOP180464 Table 1 Table

92 1 5

81 Years MRT O 18 ‰ –6.44 –3.7 –6.57 –3.15 –6.13 –5.94 –6.66 δ H 2 ‰ δ –37.5 –22.7 –34.1 –29 –35.2 –28.4 –23 –3 B g m 0.0165 0.009 0.023 0.0165 0.032 0.0195 0.027 0.022 0.047 0.03 0.006 4 –3 0.18 7.6 0.92 2.4 6.5 0.9 9.4 5.6 3.9 2.9 2.1 1.55 SO 19.6 13.2 16.55 56.05 10.4 g m –3 Na 7.6 9.4 8.81 9.7 8.97 6.125 6.97 g m 13.1 10.1 11.8 12.45 10.3 13.65 22.25 13 12.9 11.1 2 –3 5.2 SiO 73 63.4 71.75 73 84.3 73 49.2 73.8 74.4 84.7 70.9 93 80.6 74 83 80.6 g m –3 K 5.6 7.13 5.4 6.9 4.63 4.9 3.61 8.34 4.77 2.73 7.69 9.705 6.7 3.56 6.505 g m 16.6 10.25 -N –3 3 g m 0.052 2.78 3.4 2.33 0.467 6.225 3.02 0.1975 7.17 3.23 8.185 8 0.373 6 1.8 NO <0.03 <0.03 T –3 Mn g m 0.0132 0.0806 0.0019 0.0019 0.0005 0.000525 0.000425 0.0032 0.01245 0.0006 0.000675 1.47 D –3 Mn g m 0.01 0.12 0.0101 0.0545 0.006 0.01275 0.0032 0.0117 1.43 <0.005 <0.005 <0.005 <0.005 <0.005 <0.005 <0.005 <0.005 –3 Mg 3.9 2.28 4.12 2 1.3 2 2 2.35 1.79 2.11 3.885 3.56 3.68 4 2.56 1.98 3.5 g m –3 T Fe g m 0.185 0.565 0.14 3.25 0.05 <0.02 <0.02 <0.02 <0.02 <0.02 <0.02 <0.02 –3 D 0.11 0.39 0.12 2.43 0.02 Fe g m 13.2 <0.02 <0.02 <0.02 <0.02 <0.02 <0.02 <0.02 <0.02 <0.02 <0.02 <0.02 Site 76 BOP120101 BOP180417 BOP180445 BOP180448 BOP180449 BOP180451 BOP180454 BOP180456 BOP180457 BOP180458 BOP180459 BOP180460 BOP180462 BOP180463 BOP180464 70

93 –3 F g m 0.16 0.14 0.085 0.11 0.14 0.1045 0.0845 0.1 0.14 0.12 0.17 0.11 0.16 0.08 0.12 0.21 0.16 –3 TP g m 0.036 0.0475 0.097 0.059 0.074 0.029 0.106 0.09 0.076 0.0465 0.183 0.044 0.064 0.0115 0.0625 0.113 0.062 –3 DRP g m 0.025 0.022 0.1015 0.058 0.067 0.016 0.1025 0.092 0.074 0.044 0.177 0.039 0.058 0.0085 0.062 0.118 0.062 –3 Cl 4.1 6.5 9.1 4.85 3 4.8 5.95 6 3.8 9.3 6 5.35 4.2 4.6 g m 10.7 13.2 10.5 –3 Ca 7.825 2.48 3.6 2.845 2.295 2.69 7.43 5 3.45 6.71 3.77 7.375 3.3 3.1 3.07 g m 15.95 10.03 –3 Br <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 g m -N –3 3 g m 0.085 0.0075 0.0075 NH <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01

pH 5.775 6.25 6.525 6.14 6.06 6.33 6.65 6.4 6.3 6.25 6.8 6.05 6.45 6.075 6.2 6.5 6.2 3 –3 g m 46.2 31.55 41.05 28.2 25.35 23.4 29.3 30 48 28 38.1 24.2 31.2 22.8 26 39.1 25 HCO –3 3.9 6.5 6 3.1 9.4 8.9 6.8 3.6 6.2 6.9 7.1 9.8 8.65 9 3.4 10.1 DO 10.9 g m C o 13.9 22.35 14.55 13.8 10.65 11.45 13.5 12.3 13.4 13.3 16 13.2 13.6 14.25 13.6 13.6 12.8 Temp 506 522 580 387 564 498 452 457 mV ORP

EC 78.5 63 79 84 95 90 87 89 280 170 167 105 128 139 122 143 154 0.9 2.1 8.2 0 1.1 1.1 0.9 0.7 NTU Turbidity

2 2 2 3 3 3 2 2 3 3 2 3 2 2 3 3 2 Points

Site BOP180465 BOP180466 BOP180467 BOP180469 BOP180470 BOP180471 BOP180473 BOP180474 BOP180476 BOP180477 BOP180478 BOP180480 BOP180481 BOP180483 BOP180484 BOP180486 BOP180488

94 34 76 Years MRT 145

O 18 ‰ δ –6.54 –6.51 –7.2 –3.89 –6.08 –6.85 –6.99 –6.79 H 2 ‰ δ –32.1 –36.1 –41.9 –21.4 –33 –33.3 –33.6 –40.4 –3 B g m 0.036v 0.028 0.021 0.0295 0.0235 0.015 0.009 0.009 0.0135 0.0135 0.011 0.022 0.008 0.049 0.018 0.016 0.008 4 –3 9.3 1.6 1.75 1.2 9.7 4.5 3.3 1.8 2.2 2.7 2.1 SO 35.6 41.85 32.55 16.25 12.2 g m 16.05 –3 Na 8.84 7.455 8.15 9.2 6.9 9.89 8.13 g m 20.1 11.65 11.3 15 15.85 10.2 11.3 12.4 10.2 10.6 2 –3 81.7 60.4 77.5 75.3 78 75.85 81.6 82.7 58.7 74.6 79.55 85.1 80.1 81.8 78.85 SiO g m 108.85 –3 K 9.845 5.945 4.745 4.19 4.7 4.41 2.305 3.98 5.6 5.7 3.5 6.92 5.19 9.195 4.3 4.3 4.56 g m

3 –3 NO g m 0.1 1.1 1.81 2.015 2.735 0.5875 1.5 2.42 3.9 0.613 4.8 2.95 5.635 2.6 2.21 10.17 <0.03 T –3 Mn g m 0.0014 0.2655 0.00075 0.0017 0.0912 0.0204 0.000375 0.1034 0.000525 0.0006 0.000875 0.0011 0.3095 <0.005 <0.005 <0.005 <0.005 D –3 Mn 0.266 0.0153 0.101 0.258 g m <0.005 <0.005 <0.005 0.04705 <0.005 <0.005 <0.005 <0.005 <0.005 <0.005 <0.005 <0.005 <0.005 –3 Mg gm 6.28 2.78 6.155 1.05 1.825 1.32 1.02 1.46 3.57 2.5 2.39 2.68 1.54 2.47 1.47 1.96 1.2 –3 T 3.3 0.1 1.36 0.035 0.025 0.41 Fe g m <0.02 <0.02 <0.02 <0.02 <0.02 <0.02 <0.02 <0.02 <0.02 <0.02 <0.02 –3 D Fe g m 1.895 0.03 0.55 0.22 <0.02 <0.02 <0.02 <0.02 <0.02 <0.02 <0.02 <0.02 <0.02 <0.02 <0.02 <0.02 <0.02 Site BOP180465 BOP180466 BOP180467 BOP180469 BOP180470 BOP180471 BOP180473 BOP180474 BOP180476 BOP180477 BOP180478 BOP180480 BOP180481 BOP180483 BOP180484 BOP180486 BOP180488

95 –3 F g m 0.18 0.1 0.09 0.08 0.11 0.0575 0.12 0.2 0.066 0.125 0.12 0.11 0.12 0.16 0.051 0.067 0.06

–3 TP g m 0.095 0.02 0.008 0.051 0.118 0.004 0.034 0.21 0.046 0.042 0.045 0.041 0.045 –3 DRP 0.096 0.016 0.01 0.009 0.069 0.035 0.021 0.101 0.056 0.037 0.045 0.033 0.05 0.071 g m <0.004 <0.004 <0.004 –3 9 Cl 5.3 6.6 7.8 7.7 5.2 6.8 5.1 5.25 4.6 4.3 5 3.2 5.2 5.6 5.8 12.75 g m –3 Ca 3.6 3.34 6.05 3.53 2.87 6.045 2.22 2.16 3.6 4.255 5.78 4.25 3.72 2.88 2.2 7.6 g m 10.5 –3 Br <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 g m -N –3 3 0.08 0.08 0.2 0.09 0.12 g m <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 NH

pH 6.62 6.1 6.05 5.9 6.2 6.35 6.3 6.1 6.14 6.68 7.1 6.5 7.4 7.5 7.49 6.43 6.23 3 –3 g m HCO 64.9 27.3 22.3 20.7 30 20.15 18.3 20.1 22 52.1 54.2 28.9 28.1 16.5 23 42 50 –3 1.5 8.4 6.8 7.7 6.3 7.95 8.1 2.8 9.1 3.45 7 6.2 7.7 9 7.1 4.3 DO g m 10.3 7.3 6.8 C 13.7 13.8 15.3 14.7 16.8 14.3 13.7 15.3 10.3 14.9 15.1 15.6 18.6 17.3 11.9 o Temp mV 343 485 378 490 440 469 ORP 95 98 82 77 76 89 90 57

EC 126 118 110 105 106 101 172 183 106.5 0 0 2.3 5.3 0.1 0..3 NTU Turbidity

3 2 1 2 2 2 2 1 2 3 2 1 1 1 1 1 1 Points

Site BOP180489 BOP180490 BOP180491 BOP180493 BOP180494 BOP180495 BOP180496 BOP180502 BOP180512 BOP210218 BOP210219 BOP210249 BOP210250 BOP210251 Lake spring Medium Kaue2 Te Springs

96 59 43 50 Years MRT 7 or 40 4 or 44 O 18 ‰ δ –3.98 –6.55 –2.37 –1.56 –4.59 –5.409 H 2 ‰ δ –9.1 –26.7 –34.9 –16.4 –34.5 –36.5

–3 B 0.014 g m 0.0195 0.044 0.044 0.016 0.0125 0.017 0.014 0.015 0.011 0.011 0.009 0.016 4 –3 SO 2 3.9 5.5 3.2 6.95 5.1 4.3 4.7 1.585 3.2 9.7 2.6 g m 14.4 10.7 13.2 14.9 25 –3 Na 9.54 9.68 8.6 8.955 8.21 8.51 8.6 8.26 9.02 9.33 7.79 8.68 6.5 g m 10.4 16.7 11.1 14.5 2 –3 1 SiO 70.8 68.5 64.3 67 42 61 70 gm –3 K 1.99 4.99 4.35 5.83 5.43 6.615 3.7 4.14 4.5 3.235 5.17 4.57 4.28 4.39 1.1 5.5 5.3 g m -N –3 3 2.32 2.54 2.73 0.355 4.44 2.24 1.45 2.7 0.247 1.68 0.106 2.36 2.9 2 g m <0.03 <0.03 <0.03 NO

T –3 Mn g m 0.0017 0.0127 0.0172 0.0095 0.0009 <0.005 0.174 0.0549 0.0168 0.0349 0.0088 1.75 0.4675 D –3 Mn 0.0107 0.0047 0.007 0.168 0.028 0.0372 0.0061 g m 0.449 1.865 <0.005 <0.005 <0.005 <0.005 <0.005 <0.005 <0.005 <0.005 –3 Mg 1.48 1.43 2.37 1.33 1.03 2.04 0.85 0.86 1.8 2.49 2.73 1.68 1.78 1.6 2.4 3.6 4.9 gm

–3 T Fe 6.505 0.04 0.41 0.08 2.29 0.25 0.08 0.38 0.09 gm <0.02 <0.02 <0.02 12.7 –3 D Fe 5.7 0.06 4.45 2.715 0.08 0.25 0.05 0.05 g m <0.02 <0.02 <0.02 <0.02 <0.02 <0.02 <0.02 <0.02 <0.02 Site BOP180489 BOP180490 BOP180491 BOP180493 BOP180494 BOP180495 BOP180496 BOP180502 BOP180512 BOP210218 BOP210219 BOP210249 BOP210250 BOP210251 Lake Medium spring Medium Kaue2 Te Springs

97 Table 2 – Summary of water sample sites NGMP facies is as defined by Daughney and Reeves (2005)

Site Type Points Cluster Facies Aquifer 70 Bore 1 1 1A-2 Rangitiki Ignimbrite 76 Bore 1 outlier 2B Rangitiki Ignimbrite BOP120101 Stream 4 1 1A-2 BOP180417 Bore 4 2 2A Rangitiki Ignimbrite BOP180445 Bore 3 1 1A-2 Rangitiki Ignimbrite BOP180448 Bore 1 2 1B-2 Rangitiki Ignimbrite BOP180449 Bore 4 2 1B-2 Rangitiki Ignimbrite BOP180451 Bore 2 3 2A Rangitiki Ignimbrite BOP180454 Bore 2 1 1A-2 Rangitiki Ignimbrite BOP180456 Bore 3 2 1A-2 Rangitiki Ignimbrite BOP180457 Bore 2 2 1B-2 Rangitiki Ignimbrite /Oneku ? BOP180458 Bore 2 1 1A-2 Onuku Pyroclastic BOP180459 Bore 2 1 1A-2 Onuku Pyroclastic ? BOP180460 Bore 2 1 1A-2 Alluvial ? BOP180462 Bore 3 1 1A-2 Perched BOP180463 Bore 2 2 1B-2 Rangitiki Ignimbrite/Onuku ? BOP180464 Bore 2 2 1A-2 Rangitiki Ignimbrite BOP180465 Bore 2 1 1A-2 Onuku Pyroclastic ? BOP180466 Bore 2 outlier 2A Rangitiki Ignimbrite /Oneku ? BOP180467 Bore 2 1 1A-2 Rangitiki Ignimbrite BOP180469 Bore 3 2 1B-2 Rangitiki Ignimbrite BOP180470 Bore 3 2 1A-2 Rangitiki Ignimbrite BOP180471 Bore 3 2 1B-2 Onuku Pyroclastic BOP180473 Bore 2 2 1B-2 Rangitiki Ignimbrite ? BOP180474 Bore 2 2 1B-2 Rangitiki Ignimbrite ? BOP180476 Bore 3 1 1A-2 Rangitiki Ignimbrite BOP180477 Bore 3 2 1A-2 Onuku Pyroclastic BOP180478 Bore 2 N/A 1B-2 Rangitiki Ignimbrite ? BOP180480 Bore 3 1 1A-2 Rangitiki Ignimbrite BOP180481 Bore 2 2 1B-2 Rangitiki Ignimbrite ? BOP180483 Bore 2 1 1A-2 Rangitiki Ignimbrite BOP180484 Bore 3 2 1B-2 Rangitiki Ignimbrite BOP180486 Bore 3 2 2A Rangitiki Ignimbrite BOP180488 Bore 2 2 1B-2 Onuku Pyroclastic ? BOP180489 Bore 3 3 2A Kaharoa/Waiohau/Rerewhakaaitu BOP180490 Bore 2 N/A 1B-2 Rangitiki Ignimbrite BOP180491 Spring 1 1 1A-2 Onuku Pyroclastic ? BOP180493 Spring 2 N/A 1B-2 Kaingaroa Ignimbrite ? BOP180494 Spring 2 N/A 1B-2 Kaingaroa Ignimbrite BOP180495 Spring 2 1 1A-2 Kaingaroa Ignimbrite ? BOP180496 Spring 2 N/A 1B-2 Kaingaroa Ignimbrite ? BOP180502 Bore 1 N/A 1B-2 Rangitiki Ignimbrite BOP180512 Bore 2 2 1B-2 Rangitiki Ignimbrite BOP210218 Spring 3 3 2A Rangitiki Ignimbrite BOP210219 Spring 2 N/A 1B-2 Rangitiki Ignimbrite/Onuku ? BOP210249 Spring 1 N/A 1B-2 Rangitiki Ignimbrite/Onuku ? BOP210250 Spring 1 N/A 1B-2 Rangitiki Ignimbrite/Onuku ? BOP210251 Spring 1 N/A 1B-2 Rangitiki Ignimbrite Lake Lake 1 2 1B-1 Medium spring Spring 1 1 1A-2 Rangitiki Ignimbrite Te Kaue2 Springs Spring 1 1 1A-2 Rangitiki Ignimbrite

98 Table 3 – Estimated percentage of lake water contribution to the groundwater.

Site δ2H (‰) δ18O (‰) % Lake water BOP210218 –16.4 –2.37 79 BOP180476 –21.4 –3.89 55 BOP180417 –22.7 –3.15 60 76 –23 –3.7 54 BOP180489 –26.7 –3.98 45 Medium Spring –34.5 –4.59 25 Te Kaue2 Springs –36.5 –5.409 13

Lake, BOP210218 and BOP180489 in the All sites were classified into the six northwest, 2) site BOP180417 located close hydrochemical facies (Table 2) developed to the western shore of Lake Rerewhakaaitu, by Daughney and Reeves (2005) to observe and 3) sites BOP180486 and BOP180449 how data collected in this study compared to located close together to the southeast of, and data collected in the National Groundwater close to, Lake Rerewhakaaitu. Forty-two sites Monitoring Programme (NGMP). Sites have median nitrate-nitrogen concentrations in this study fall into five of the six greater than 0.14 g m–3, suggesting ground hydrochemical facies defined by Daughney water in the area has been affected by and Reeves (2005). Namely, seventeen sites anthropogenic sources. This assumes fall into category 1A-2, one site (Lake) into background (pre-development) nitrate- category 1B-1, twenty-two sites into category nitrogen concentrations calculated for the 1B-2, six sites into category 2A and one site Rotorua aquifers (Morgenstern et al., 2004) (site 76) into category 2B. Most sites fall into are the same for the Rerewhakaaitu aquifers. the threshold 1 (clusters prefixed by a ‘1’) Median dissolved reactive phosphorus category. This category represents ‘Surface- (DRP) concentrations range from below dominated, oxidised, unconfined aquifers the detection limit (<0.002 g m–3) to 0.264 with a low to moderate TDS (total dissolved g m–3. High (>0.1 g m–3) median dissolved reactive phosphorus levels were measured at five sites (BOP180457, BOP180478, BOP180463, BOP180473 and BOP180467) south of Lake Rerewhakaaitu, and two sites (BOP180512 and BOP180486) to the east of Lake Rerewhakaaitu. Median dissolved reactive phosphorus concentrations generally increase with depth (Fig. 3). Fish (1978) suggested that dissolved phosphate entering Lake Rerewhakaaitu is rapidly adsorbed by colloidal allophanic clays in the Rotomahana mud deposits, reducing dissolved phosphorus concentrations in the lake and limiting algae Figure 3 – Median dissolved reactive phosphorus growth. (DRP) versus depth.

99 solids)’. From within this category, seventeen analytes is probably due to large number of of the sites (category 1A-2) have been samples that contain censored data for these classified as most probably having human analytes. impact on the aquifer. The 1B-1 classification Sites 76 and BOP180466 were identified (surface waters from a carbonate or clastic as outliers using the nearest neighbour aquifer) for the Lake water sample is marginal, linkage rule in the hierarchical cluster based on the five censored parameters in the analysis algorithm. The water chemistry water analyses. The NGMP hydrochemical of Site 76 is clearly different from all the facies classifications assigned to the Lake other sites, having a HCO3 concentration Rerewhakaaitu data are mostly consistent an order of magnitude higher than any with what would be expected, e.g., thirty-nine other site, and very high NH3-N, Fe and of the forty-seven sites have been identified K concentrations compared to all the other as coming from a volcanic or volcaniclastic sites. Site BOP180466 abstracts water from a aquifer (categories 1A-2 and 1B-2). reducing environment, based on the median dissolved oxygen, median Fe (total and Hierarchical cluster analysis dissolved), median Mn (total and dissolved) Median concentrations of 14 analytes (SiO2, and median NH3-N concentrations. This K, Mg, Mn, Na, NH3-N, NO3-N, SO4, Ca, site has significantly higher median SO4, Cl, DRP, F, Fe and HCO3) were calculated SiO2 and Ca concentrations compared to the for 41 sites. Only sites where the full list other sites in the study extracting water from of analytes was available were used for the reducing environments. This may be caused analysis (Table 1, Table 2). Twelve sites were by effects of local land use, and therefore not used because the SiO2 analyses were makes this site an outlier. Both of these sites missing. Other analytes such as boron and were taken out for the remaining hierarchical bromide were not used due to the limited cluster analysis clustering process. number of sites that contained this data. All but six analytes were normalised using a log- normal distribution. F and Na required no transformation as the data was already normally distributed. HCO3, DRP, NO3-N and SiO2 were power transformed by the exponents –2, 0.5, 0.5 and 3 respectively. Fe, Mn and NH3-N have poor log-normal distributions so other transformations were investigated, but could not improve the results of the Kolmogorov-Smirnov test for these three analytes. The inability to obtain a normal distribution using Figure 4 – Dendogram from the hierarchical cluster analysis some transformation for these clustering algothithm showing the 3 clusters.

100 Three clusters were determined by the Fe). Sites BOP180483 and BOP180462 have authors to be the optimum number of negligible concentrations of NH3-N and Fe, meaningful clusters that could be interpreted with either a median NH3-N or median Fe from the hierarchical cluster analyses for the concentration equal to the detection limit. 39 sites (Fig. 4). A large geographic spread of Water from this cluster is oxidized. Cluster sites classified as cluster 1 or 2, both inside 1 has the highest centroid median of Cl, Mg, and outside of the surface water catchment, NO3-N and SO4 concentrations (Table 4) is observed (Fig. 5). This suggests that the compared to the centroid medians from the surface water catchment has little to do other clusters. This suggests that these waters with the groundwater catchments and that may be affected by land-use activities. Nitrate groundwater is moving from/to the Lake is the dominant nutrient leaching into the Rerewhakaaitu surface water catchment, and water around these sites given that this cluster similar chemical processes are occurring in has the highest median concentration of groundwater aquifers inside and outside of NO3-N, but the lowest median DRP con the Lake Rerewhakaaitu catchment. No link centration compared to the other clusters. between bore elevation or bore depth to the Cluster 2 includes 18 bores and the clusters was observed. Lake Rerewhakaaitu site. Sites with this Cluster 1 includes four springs, one classification (excluding the lake site) occur to stream and twelve bores, and geographically the east and the south of Lake Rerewhakaaitu. covers the whole study area. All sites in This cluster has the highest centroid median this cluster, except sites BOP180483 and DRP and SiO2 concentrations and lowest BOP180462, have median NH3-N and Fe centroid medians of Ca, Cl, HCO3, Mg concentrations below the detection limit and Na compared to the centroid medians –3 –3 (0.01 g m for NH3-N and 0.02 g m for from the other clusters (Table 4). All sites except for BOP180471 have no NH3-N (site BOP180471 has a median NH3-N concentration equal to the detection limit). Five sites in this cluster (BOP180448, BOP180470, BOP180417, BOP180471 and BOP180486) contain Fe and Mn, with the Lake site containing Fe. All these six sites are close (within 1.5 km) to the Lake. Sites BOP180448, BOP180474, BOP180481, BOP180484, BOP180488 and BOP180471 are more like each other than like other sites (Figure 4), given that they plot together in one sub-cluster of cluster 2. These sites are located to the south of Lake Rerewhakaaitu. Figure 5 – Map showing the geographical distribution of sites, with their assigned cluster and percentage lake water component. No other obvious geographic

101 Table 4 – Centroid median concentrations of hierarchical cluster analysis clusters.

Cluster Ca Cl DRP F Fe HCO3 K Mg Mn Na NH3-N NO3-N SiO2 SO4 1 7.77 8.94 0.025 0.080 <0.02 28.8 6.86 3.28 0.0047 12.91 0.005 4.16 72.9 14.3 2 3.29 4.93 0.069 0.128 0.02 28.0 4.06 1.76 0.0052 9.52 0.005 1.35 77.4 2.9 3 3.98 6.33 0.044 0.155 3.35 51.0 2.85 1.95 1.0619 11.31 0.108 0.10 56.7 1.4

All concentrations are in g m–3 distributions for sites in sub-clusters of cluster sites further away. The mean residence time of 2 are observed. the groundwater gets older as the proportion Cluster 3 includes two bores and one spring, of the lake water gets smaller at sites (except all located to the northwest part of the study Site BOP180476) on the west and north of area. These three sites are more like each other the lake. than the other sites sampled. Cluster 3 has the Bores BOP180417 (153 m deep) and highest median centroid concentrations of F, BOP180476 (91 m deep) have large Fe, HCO3, Mn and NH3-N and the lowest proportions of lake water (60% and 55% centroid median concentrations of NO3-N respectively) (Table 2), considering their and SiO2 than the other clusters (Table 4). depths. Bore BOP180476 is cased to 39 m, This cluster represents anaerobic waters from therefore has a 52 m water inlet into the bore. a reducing environment. Further work is required to determine the Cluster 2 includes the lake site. depth range at which recharge is occurring Interestingly, the lake site is chemically more from the lake. No casing data exists for bore like groundwater that occurs to the southwest BOP180417. Both bores are regularly used and to the northeast than to groundwater to to supply water to dairy farms. The increased the northwest of Lake Rerewhakaaitu. Sites groundwater recharge required to account in cluster 3 plot closer to the lake site than for drawdown and recovery may affect the sites Te Kaue2 Springs, Medium Springs and naturally occurring ratio of lake water to the site 70 (all from cluster 1) (Fig. 4). Different native aquifer water. sources of water for the groundwater A significant reduction in the proportion (including springs) in the northwest part of of lake water can be seen between sites the study area may explain this. Also, sites BOP180476, Medium Springs and Te BOP180454, BOP180483 and BOP180480 Kaue2 Spring (Table 3). All these sites are in are grouped (Cluster 1) together and represent the Rangitaiki Ignimbrite aquifer. This the cluster 1 sites to the southeast of Lake suggests that: Rerewhakaaitu. 1) Lake water is been diluted by native Discussion groundwater or rainwater between sites Oxygen 18 and deuterium data show that BOP180476 and the springs. Lake Rerewhakaaitu contributes water to the 2) The springs represent typical mixing springs and groundwater to the north and west proportions in the Rangitaiki Ignimbrite of Lake Rerewhakaaitu. The contribution of aquifer in this area, whereas site lake water to the groundwater at sites on the BOP180476 has increased proportions of northwestern side of the lake increases to the lake water, due to pumping causing water northwest. Sites close to the lake typically from the lake to preferentially recharge the have larger proportions of lake water than aquifer.

102 3) Mixing of lake water in the aquifer is some clustering of sites with suspected lake variable with depth. This is difficult to water contributions does occur, namely, the determine given the large water intake at cluster 1 and cluster 3 sites mentioned above. the bore. Significant differences in aquifer conditions The large number of springs and seeps between these clusters are evident. The cluster that are observed in the area around site 3 sites (BOP210218 and BOP180489) BOP210218 to site 76 (northwest side of the are in strongly reducing environments and lake) probably contain high proportions of lake have very young mean residence times water. The high proportion of lake water at (4 and 7 years respectively), whereas the these sites (79% and 54% at sites BOP210218 cluster 1 sites (Medium Spring, Te Kaue2 and site 76 respectively) suggests significant Spring and BOP180476) are in an oxidizing lake leakage in this area. Also, the three sites environment and have older mean residence included in the hierarchical cluster analyses times (43, 50 and 34 years respectively). The older mean residence times at these sites in this area (BOP180451, BOP210218 and suggest either lake water is mixing with older BOP180489) form their own cluster (Cluster groundwater, or, that the groundwater flow is 3). This cluster is partially characterized by slower from the lake to these sample points. high concentrations of Fe, Mn and NH -N. 3 These possibilities enable the groundwater The high Fe, Mn and NH -N concentrations 3 to be altered via contaminants in the older are also measured at site 76, not included groundwater or from increased water-rock in the hierarchical cluster analyses. Water interaction. flow from the lake to the springs is probably Water in the Rangitaiki Ignimbrite aquifer rapid, given that the mean residence time on the eastern side of the lake is interpreted to of water from site BOP210218 is young be recharged by rainfall and/or other rainfall- (mean residence time of 4 or 44 years). The recharged groundwater aquifers, based on the series of blast vents along the western edge available data. Oxygen-18 and deuterium of Lake Rerewhakaaitu (Nairn, 2002) may isotope data for sites plot in a group close to have increased permeability in the volcanic the meteoric water line, suggesting no lake material, enabling lake water to leak in this water signature. No distinctive pattern is area. This also suggests that water leaking from observed in the hierarchical cluster analyses Lake Rerewhakaaitu is a significant con of major anions and cations. The oldest mean tributor to the groundwater recharge of the residence time (145 years) is measured at Kaharoa/Waiohau/Rerewhakaaitu aquifer. site BOP180486, approximately 700 m east The groundwater of this aquifer flows to from the eastern edge of Lake Rerewhakaaitu. the west into Lake Rotomahana (White The site is a bore that draws water from et al., 2003). approximately 26–69 m below ground. Water Hierarchical cluster analyses of the water having a long mean residence time would be quality data has not produced clusters that expected with slow-moving groundwater on clearly identify groundwater that has been the top of a groundwater divide, as mapped recharged from the lake. This is because there by White et al. (2003). is not a large variation in water chemistry in Elevated iron concentrations in the the study area, making meaningful clustering groundwater in the Lake Rerewhakaaitu difficult. Overall, there is little difference area may be due to the leaching of iron from between the median centroid concentrations material erupted from the 1886 Mt Tarawera for clusters 1 and 2, which contain all but eruption. Nairn (2002) shows that basaltic three sites used in this analyses. However, material erupted from this event has the

103 is the dominant form of aquifer recharge in the Lake Rerewhakaaitu catchment; however, aquifers to the west and north of Lake Rerewhakaaitu also have a lake water contribution. This may have environmental catchment management implications for both the Lake Rerewhakaaitu and Lake Rotomahana catchments. An area of anaerobic groundwater is identified in the northwest part of the study area, containing elevated levels of iron, manganese and ammonium. Recent volcanic Figure 6 – Location of sites where the median iron (dissolved or deposits may be the source total) concentration is above the detection limit. of iron in the groundwater. Nutrient levels in most aquifers highest concentrations of hematite (Fe2O3) are above the expected background levels, of any volcanic material tested in this area, suggesting some effect of land use on the approximately 5–50 times higher in hematite water quality in the study area. than the representative ignimbrites and Defining meaningful hierarchical cluster pyroclastic formations in the area. Iron could analysis clusters based on the major anion/ be leaching out of this material (and eroded cation water chemistry was difficult due to sediment) as water passes through this layer the similarity of water chemistry between of sediment. It would be expected that this the sites. Similarities in major anion and sediment would be abundant near the vents cation water chemistry are probably due to (e.g., Mt Tarawera, along the west side of Lake groundwater not having time for significant Rerewhakaaitu) and in low-lying areas where water-rock interactions (based on the low sediment has been moved due to erosion mean residence times) in aquifers of similar (e.g., Lake Rerewhakaaitu, stream valleys). material. The three clusters defined in this Sites with detectable median iron (total or study do not show clear ‘clustering’ between dissolved) concentrations tend to follow sites that contain lake water and those that this trend (Fig. 6). More work is required to don’t. However, the hierarchical cluster confirm that this material is the main source analyses can be used to infer sites that may have of iron in these groundwater aquifers. lake water components in the groundwater, Summary and therefore provide a method to identify sites that may require further study. Oxygen-18, deuterium, tritium, chloro fluorocarbon and major water chemistry data have been successfully used to identify areas Acknowledgements of lake leakage from Lake Rerewhakaaitu. The authors would like to thank the bore The data suggest that rainwater infiltration owners who allowed us to sample their

104 bores/springs. This project was funded by Hamilton, B. 2003: A review of short-term EBOP and the Foundation of Research and management options for Lakes Rotorua Technology, non-specific output funding and Rotoiti. A report for the New Zealand contract. Ministry for the Environment, December 2003. http://www.envbop.govt.nz/media/pdf/ lakes-rotorua-rotoiti-jan04.pdf Helsel, D.R.; Hirsch, R.M. 1992: Statistical References Methods in Water Resources. Studies in Back, W. 1961: Techniques for mapping of Environmental Science v. 49, Elsevier, hydrochemical facies. United States Geological Amsterdam. 529 p. Survey Professional Paper 424-D: 380-382. Kristmannsdóttir, H.; Ármannsson, H. 2004: Back, W. 1966: Hydrochemical facies and Groundwater in the Lake Myvatn area, groundwater flow patterns in northern part of northern Iceland: Chemistry, origin and Atlantic Coastal Plain. United States Geological interaction. Aquatic Ecology 38(2): 115-128. Survey Professional Paper 498-A. Lakes Water Quality Society 2001: Proceedings Chowdhury, A.H. 2004: Hydraulic interaction and Report, Rotorua Lakes 2001, A symposium between groundwater, Brazos River and oxbow on research needs in the Rotorua Lakes. 22-23 lakes – Evidence from chemical and isotopic March, 2001, Rotorua, New Zealand. compositions, Brazos River basin, Texas. Lower Lakes Water Quality Society 2003: Proceedings, Brazos River Oxbow Project, Texas Water Rotorua Lakes 2003, Practical Management Development Board Report 41. for good Lake Water Quality. 9-10 October, Cook, P.G.; Herczeg, A.L. 1999: Environmental 2003, Rotorua. Tracers in Subsurface Hydrology, Kluwer McIntosh, J.J.; Ngapo, N.; Stace, C.E.; Ellery, Academic Publishers. P 529. G.R.; Gibbons-Davies, J. 2001: Lake Darling, W.G.; Gizaw, B.; Arusei, M.K. 1996: Rerewhakaaitu Project, EBOP Environment Lake-groundwater relationships and fluid-rock Report 2001/15. p. 84. interaction in the East African Rift Valley: Ministry of Health 2005: Drinking-water Isotopic evidence. Journal of African Earth Standards for New Zealand 2005. Wellington: Sciences 22(4): 423-431. Ministry of Health. Daughney, C.J.; Reeves, R. 2003: Definition Morgan, C.O.; Winner, M.D. Jr. 1962: of hydrochemical facies for New Zealand’s Hydrochemical facies in the 400 foot and 600 groundwaters using data from the National foot sands of the Baton Rouge area, Louisiana. Groundwater Monitoring Programme. United States Geological Survey Professional Institute of Geological & Nuclear Sciences Paper 450-B: B120-121. Science Report 2003/18. 68 p. Morgenstern, U.; Reeves, R.R.; Daughney, C.; Daughney, C.J.; Reeves, R.R. 2005: Definition Cameron, S.; Gordon, D. 2004: Groundwater of hydrochemical facies in the New age and Chemistry, and Future Nutrient Zealand National Groundwater Monitoring Load for Selected Rotorua Lakes Catchments. Programme. Journal of Hydrology (NZ) 44 (2): Institute of Geological & Nuclear Sciences 105-130. Science Report 2004/31. 73 p. Fish, G.R. 1978: Lake Rerewhakaaitu – an Morgenstern, U.; Stewart, M.K. 2004: Stream apparently phosphate-free lake. New Zealand water dating and future nutrient load to Journal of Marine and Freshwater Research 12 Lake Taupo, NZ. International Workshop (3): 257-263. on the Application of Isotope Techniques in Freeze, R.A.; Cherry, J.A. 1979: Groundwater. Hydrological and Environmental Studies, Prentice Hall, New Jersey. 604 p. UNESCO, Paris, 6-8 September, pp. 173-174. Guler, C.; Thyne, G.D.; McCray, J.E.; Turner, Morgenstern, U. 2005: Wairarapa Valley A.K. 2002: Evaluation of graphical and Groundwater – Residence time, flow pattern multivariate statistical methods for classification and hydrochemistry trends. Institute of of water chemistry data. Hydrogeology Journal Geological & Nuclear Sciences Science Report 10: 455-474. 2005/33, 36 pp.

105 Morgenstern, U.; Gordon, D. 2006: Prediction Stewart, M.K.; Fox, V.; Van der Raaij, R. 1999: of Future Nitrogen Loading to Lake Rotorua, Conceptual model of the groundwater GNS Science Report 2006/10, 28 p. resource between the Ashley and Rakaia Nairn, I.A. 2002: Geology of the Okataina rivers, Canterbury, using the CFC dating Volcanic Centre, scale 1:50 000. Institute of technique. New Zealand Hydrological Society Geological & Nuclear Sciences geological 1999 Symposium Proceedings, Napier, 23-26 map 25. 1 sheet + 156 p. Lower Hutt, New November, p 76. Zealand: Institute of Geological & Nuclear Stewart, M.K.; Morgenstern, U. 2001: Age and Sciences Limited. source of groundwater from isotope tracers. In Ojiambo, B.S.; Poreda, R.J.; Lyons, W.B. 2001: Groundwaters of New Zealand, M.R. Rosen Ground water/surface water interactions in and P.A. White (eds.). New Zealand Hydrological Lake Naivasha, Kenya, using delta O-18, delta Society Inc., Wellington, pp. 161-183. D, and H-3/He-3 age-dating. Ground Water Taylor, C.B.; Freestone, H.J.; Nairn, I.A. 39(4): 526-533. 1977: Preliminary measurements of tritium, Quast, K.W.; Lansey, K.; Arnold, R.; Bassett, deuterium and oxygen-18 in lakes and R.; Rincon, M. 2006: Boron Isotopes as an groundwater of volcanic Rotorua region, New Artificial Tracer.Groundwater 44(3): 453-466. Zealand. Institute of Nuclear Sciences, DSIR Rosen, M.R.; Cameron, S.G.; Reeves, R.R.; (INS – R -227). Taylor, C.B. 1999: New Zealand guidelines Taylor, C.B. 1994: Hydrology of the Poverty for the collection of groundwater samples for Bay Flats aquifers: New Zealand; recharge chemical and isotopic analysis. Institute of mechanisms, evolution of the isotopic Geological & Nuclear Sciences Limited Science composition of dissolved inorganic carbon and Report 99/9. 80 p. groundwater ages. Journal of Hydrology 158: Scholes, P.; Bloxham, M. 2005: Rotorua Lakes 151-185. Water Quality 2005. EBOP publication van der Raaij , R. 2004: personal communication. 2005/24. p. 71. White, P.A. 2001: Groundwater resources in New Seaber, G.E. 1962: Cation hydrochemical facies Zealand. In: Groundwaters of New Zealand. of groundwater in the Englishtown Formation, M.R. Rosen and P.A. White (eds.), New New Jersey. United States Geological Survey Zealand Hydrological Society, Wellington, 47- Professional Paper 450-B: B124-126. 75. Stewart, M.K. 1978: Stable isotopes in waters from White, P.A.; Nairn, I.A.; Tait, T.; Reeves, R.R. the Wairakei geothermal area, New Zealand. 2003: Groundwater in the Lake Rerewhakaaitu In Stable Isotopes in the Earth Sciences, DSIR catchment. GNS client report 2003/62, p. 62. bulletin 220, 230 p. Stewart, M.K.; Taylor, C.B. 1981: Environmental isotopes in New Zealand hydrology. New Zealand Journal of Science 44: 295-311.

Manuscript received 28 November 2007; accepted for publication 31 July 2008

106