Groundwater age in the Wairarapa
This dissertation focuses on the catchment-scale evaluation of groundwater age as a function of space and time in the 270 km² Middle Wairarapa catchment. The simulation of the mean age and point distribution of ages, contributing to a regional age estimate, is a novel demonstration of the recently developed groundwater software, Cornaton (2012). The Wairarapa is in the southern North Island of New Zealand and is a dynamic water catchment exhibiting complex interactions between its rivers and shallow aquifers. Groundwater has been widely utilized since the 1980s for agriculture, horticulture and drinking water; increasing land use development (i.e. irrigation and nutrient application) requires effective regional management of both the quantity and quality of water resources. Groundwater age provides insights into groundwater flow and transport processes and thus enables better management of groundwater resources. Subsurface water age information enables the interpretation of recharge influence, zones of sensitivity for sustainable abstraction, as well as contamination risk from land-use intensification to drinking water supplies. It is accepted that groundwater is composed of a mixture of water with different ages, however, until very recently mean age has been the primary indicator for groundwater age assessment. Mean age alone can misrepresent the potential for contamination from young water; for example, a groundwater sample with an old mean age may still contain a significant fraction of young water; therefore, a fuller understanding of the age distribution in both time and space is important for groundwater management. The ability to simulate the full distribution of groundwater age within transient numerical groundwater models has only been very recently enabled, through implementation of the time-marching Laplace transform Galerkin technique (TMLTGT), and is demonstrated in this dissertation. A transient finite-element groundwater flow model originally developed by Greater Wellington Regional Council was converted to simulate transport of the age tracer tritium and groundwater age using the Ground Water (GW) software. Observed tritium concentrations were utilized in the calibration using the Monte Carlo and Gauss-Marquardt-Levenberg methods. Following the calibration of the transport model the GW software was then used to derive pumping well capture zones and directly simulate age throughout the Middle Wairarapa Valley catchment. The advective dispersive equation and the TMLTGT were used for transient mean-age and transient simulations of the full distribution of groundwater age. The results are presented as maps and graphs of both mean age and age distributions throughout the Middle Valley, covering a 15 year simulation period. The mean-age simulations indicated the groundwater age in the valley was strongly influenced by seasonal changes and extreme climatic events. Significant variations existed, from high rainfall recharge percolating young water throughout the domain, to dry extended droughts limiting recharge and increasing the age throughout large sections of the Middle Valley. Age distributions were shown to be strongly influenced by abstraction pressures, depth and geology. Abstractions were shown to skew the age distribution, creating both older and younger mean-ages depending on the location of the observation point, and several simulations indicated the potential misrepresentation of young (potentially contaminated) water quantified as old by mean-age assessment. These results show the dynamic nature of the Middle Valley groundwater system and its inherent vulnerabilities. The Wairarapa transient age distributions are one of the first such examples in New Zealand, and they demonstrate the potential of the information interpreted from age estimates to more effectively manage groundwater resources.