MODFLOW - West Melton Aquifer

The MODFLOW model was used to simulate the groundwater levels and river flows under different climatic conditions and to assess the  effects of increasing abstraction on the Christchurch - West Melton aquifer system river flows and the saltwater interface.

The aim of this study was to provide background information on the groundwater system below Christchurch and its sources with a primary focus on the relationship between groundwater takes and the reliability of base flows in the Avon and Heathcote rivers. This information is then intended be used in part to inform management decisions surrounding Christchurch water takes given forecast increasing risk of low flows as a result of increased ground water uses.   

MODFLOW has the capability of modelling groundwater flows in 3D meaning that its application to the multi-layered aquifer located between Halkett in the west to Pegasus bay in the east was well suited. 

MODFLOW was used to simulate the responses of the Christchurch-West Melton groundwater system to growing abstraction rates under a range of climate conditions. The model was set to run over a total area of 1,150km2 with a resolution ranging from 0.5km to 8km in grid size. The no flow boundaries were set in accordance with landscape features and applicable knowledge of the groundwater system, for example to the south a boundary was set against Banks Peninsula known to be relatively impermeable.  

MODFLOW case study boundaries

Extent and boundary conditions of the Christchurch - West Melton groundwater model

The three dimensional nature of the Christchurch ground water system was developed from the analysis of borehole logs throughout the area. Each of the five aquifers, part of the multi layered geology of the groundwater system, were represented by a single model layer. The vertical boundaries of each model layer were determined from inferred aquitard thickness derived from borehole data.

In this study 3 scenarios were modelled, these scenarios were constructed on past data involving hindcasting. This was done because the rate of demand growth on the groundwater system and future climate are unknown and as a result future modelling could be perceived as highly speculative. Instead, in order to have a sound foundation on which to analyse the effects of varying levels of groundwater abstraction on related base flows, scenarios were modelled on the 35 year period between 1964 and 1999.

The three case studies were then based on the following assumptions:

  • The year of 1998/99 was used as a base case in terms of spatial and temporal patterns.
  • Alternative groundwater abstraction scenarios were defined as percentages of this base case.
  • Modelling of groundwater levels and stream flows occurred using climate based estimates of rainfall recharge for the 35 year period between 1964 &1999.

Case 1 – Across the board change

For this case the 1998/99 water use pattern was used to define higher (or lower) levels of abstraction on a strictly proportional basis. Fourteen abstraction scenarios were evaluated ranging from 70% (i.e. 30% reduction) to 200% (i.e. a 100% increase) of the 1998/99 abstraction.

Case 2 – Layer 1 fixed

This case was designed to represent the outcome of a policy to limit future abstractions from the shallowest aquifer. For all simulations the abstractions from Layer 1 of the model were fixed to match those for the 1998/99 year. Ten alternative abstraction scenarios were evaluated with total abstraction ranging from 110% to 200% of the 1998/99 abstraction with the adjustment factor applied so that increases in abstraction were met by changes in abstraction from the lower model layers only. The option of fixing Layer 1 abstractions would not be irrelevant for scenarios involving a reduction in total abstraction

Case 3 – Unconfined/confined restriction regimes

This case was included to illustrate the comparative effectiveness of restrictions on abstractions from the unconfined and confined. Abstraction rates were based on 130% of the 1998/99 abstraction but alternative restrictions were evaluated for the following scenarios:

  1. No controls
  2. No pumping from within confined aquifer zone when total drain discharge falls to
    700,000 m3/d
  3. No pumping from within confined aquifer zone when total drain discharge falls to
    413,835 m3/d
  4. No pumping from within unconfined aquifer zone when total drain discharge falls to
    700,000 m3/d
  5. No pumping from within unconfined aquifer zone when total drain discharge falls to
    413,835 m3/d



Case 1 – Across the board change.

Each of the fourteen simulation runs was used to provide estimates of Avon River and Heathcote River baseflows (as mean monthly flow) for the 35-year simulation period. For the first of the scenarios run in this case (70%) was determined that the baseflow drops below the 1500 l/s level for 5 of the 35 years (1970, 1971, 1972, 1973 and 1989). This was in contrast to the last scenarios (130%) where the baseflow drops below 1500 l/s for every year except 1980.

Case 2 - Effectiveness of constraints on layer 1

Comparison with the equivalent results for the Case 1 simulations indicates that for the Avon River the strategy of restricting abstractions from layer 1 results in an improvement in minimum low flow of around 100 l/s. For the Heathcote River there is comparatively less advantage: at the 0% change the percentage of years with minimum flows below the specified values is unchanged.

Case 3 – Unconfined/confined restriction regimes

None of these restriction options are particularly realistic options but have been included to provide results consistent with some of the earlier work described in Little and Scott (1999) where the same flow rates were used as restriction triggers. Nevertheless, the results do once again illustrate the potential for different types of response to restrictions in different aquifer zones and suggest that such options should be studied more closely.

The scenarios that were produced by MODFLOW during this study suggested a strong link between groundwater systems and the base flow levels of the Avon and Heathcote rivers. This modelling provided a development on previous understandings of the Christchurch-West Melton system which tended to generate the impression that abstraction on the system resulted only in a net loss of fresh water flowing to the nearby coastline, with little effect on fresh water springs that feed the Avon and Heathcote rivers. It was also noted as a result of this modelling that the Christchurch system was particularly responsive to climatic variation and likely more so than forecast growth in groundwater use.

This case study showed MODFLOWS capabilities in modelling 3D groundwater systems and the additional insights into a system that it can provide. In the case of the Christchurch-West Melton system this modelling informed a future issues and options document to assist the development of alternative management regimes as part of the Section 32 analysis required for the development of a new draft plan.         

Associated Models


MODFLOW is the USGS's modular hydrologic model.


Scott, D (2000): Christchurch – West Melton Groundwater Investigation- Simulation of alternative groundwater abstraction scenarios and their effects on baseflows of the Avon and Heathcote rivers.  Environment Canterbury unpublished report U00/33

Little, R.J. and D.M. Scott (1999): Christchurch - West Melton groundwater: Simulating groundwater augmentation. Canterbury Regional Council Report U98/35.