Abstract:
The study seeks to understand the complex groundwater systems within Waihi Basin (Coromandel Area, North Island, New Zealand), a river basin covering c.140km2 which is drained by a single river, the Ohinemuri River. These groundwater systems are hosted by rather shallow, fracture ignimbrite aquifers and deeper andesite rocks infilling the basin down to depths of c.2km. All groundwaters are found to mix freely with infiltrating meteoric water throughout the whole basin.
Because of the favourable river basin setting, all existing meteorological and hydrological data collected during 1992 to 1996 were analysed in terms of the mass balance components. Incomplete data sets only allowed the assessment of the three major components for the year 1994, namely total precipitation (0.22km3/yr), total evapotranspiration (0.1lkm3/yr) and total stream discharge (0.13 km3/yr). The ratio of these major components in the other years appears to be similar. Other smaller components in the mass balance were assessed by numerical modeling. For this, the whole basin was subdivided into numerous blocks and three layers with the intersection of the blocks reflecting gross tectonic structures. The observed piezometric levels of the groundwater wells intersecting the two top layers were used as matching parameters retaining infiltration (recharge) as an unknown parameter. The best fit numerical models points to a rather small annual recharge of the groundwater in the basin amounting to c.l% of the volume of the total annual precipitation (0.002km3/yr). The best fit basin model also predicts a somewhat smaller stream discharge (0.1Okm3/yr) than that observed. The difference can be interpreted in terms of change in storage and a discrepancy factor which might reflect the uncertainty in the observed total stream discharge. There is no evidence for any significant in- or outflow beneath the catchment boundaries.
Most waters from groundwater wells in the Waihi Basin exhibit a very low mineralisation as shown by their average value of total dissolved solids (TDS) of c.110mg/kg. Significantly higher values (>3,000mg/kg) occur only within and around the workings of the old Waihi Mine. The waters with low mineralisation can be classified as sodium-bicarbonate waters of low ionic strength. Despite the low strength, their cation ratios are controlled by fluid-rock interactions which fingerprint different types of volcanic host rocks (i.e. rhyolite, ignimbrite, andesite) and traces of distinct alteration minerals (i.e. kaolinite, montmorillonite, illite).
The geochemical constituents in the groundwater vary throughout the year with silica sodium and TDS showing the smallest changes. Their concentration patterns across the whole basin points to the occurrence of slightly higher mineralized, dilute thermal waters in the axial and central part of the basin which discharge in a few wells as artesian flow with slightly elevated (a few degrees only) temperatures. Using a 2-D numerical simulation model, derived in part from the earlier numerical basin model, the associated advective flow of slightly heated deeper groundwater could be modelled using the near-surface discharge temperatures of artesian wells and their location as a matching parameter. It was found that a deep terrestrial heat flow of c.90mW/m2 is required to initiate convection and advective flow. This points to a deep basin temperature of c.118 degrees C at 2km depth. The best fit model points to a discharge centre not only in the basin centre but also to a smaller one lying outside the SE-basin divide. The probable sources for the anomalous terrestrial heat flow are most likely related to paleo-subduction and associated paleo-volcanic processes.
The numerical models simulating the groundwater flow in the Waihi Basin also allow an assessment of the local groundwater anomaly caused by the pumping and dewatering of the Waihi open cast which started in 1990. The decline in waterlevel in the centre of the open pit in the old No.7 shaft was used as matching parameter in modelling the local groundwater flow around the mine. The actual pumping rates were poorly known and were treated as an unknown parameter. The best fit model provides extreme values for the pumping rates in 1997 with a mean of 60 ± 30 kg/s; this agrees with the actual mean annual pumping rates of c.55 kg/s for 1997 and allows the prediction that the deep piezometric level within the andesites around the mine has been depressed to mean sea level and affects an area within a radius of at least 0.5km around the centre of the open cast mine. The average hydraulic conductivities of the rocks within and beneath the cone of depression are also one order of magnitude lower than conductivities in adjacent blocks reflecting presumably the accumulated effects of older and present day blasting activities.
The study has shown that detailed simulation of groundwater flow is required to understand the actual flow regimen of a groundwater basin even in an ideal setting such as the Waihi river basin.