Hydrogeologist Andrea Broughton’s Comments re Takaka FLAG “Summary of Interim Decisions”

Groundwater Solutions International
115 Tasman St, Mt Cook
Wellington 6021

31 January 2017
Lisa McGlinchey
Tasman District Council

Dear Lisa

Thank you for the opportunity to provide comment on the Takaka FLAG “Summary of Interim Decisions for Water Quantity and Quality Management in the Takaka Freshwater Management Unit, November 2016 (updated 16 December 2016)”.

I am providing a partial submission as I have not had the chance to review the Science Panel report (not yet available to the public) and the section of Dr Roger Young’s report that deals with his ideas regarding “A Geological Contribution to the Nitrogen Load to Te Waikoropupu?” as I understand that report was not available to the public until a few days ago. My comments outlined below concern statements made in the FLAG document relating to groundwater, especially in Section 8.7.

I hold a Master’s Degree in Applied Science in Hydrogeology and Groundwater Management, from the University of New South Wales (1992), and a Bachelor of Science (Honours) in Geology (1989) from Victoria University of Wellington. I am a member of the International Association of Hydrogeologists (IAH), a member of the New Zealand Hydrological Society; a member of the Geoscience Society of New Zealand; a credited member of Landscaping New Zealand. Since 1992 I have been employed as a Project Hydrogeologist by the NSW Government; a Hydrogeologist and Senior Hydrogeologist by Australian and New Zealand environmental consultancies; and since 2006 I have operated as a Sole Trader under the name Groundwater Solutions International contracting to many environmental, farming and community groups in Australia and New Zealand, NSW Environmental Defence Office and the Greater Wellington Regional Council.

My professional expertise is in hydrogeology, geophysics (in the groundwater context) and geology. I have researched and examined hydrogeology to a Master’s level. I have researched and published technical reports, and produced hydrogeological maps for the NSW State Government. I have co-authored scientific papers and presented at conferences, and was a guest speaker for radio programmes. I have conducted groundwater educational workshops for farming and community groups. I was invited to speak as a groundwater expert on a Panel of Expert Scientists as part of the Coal Seam Gas Science and Law Forum held at NSW Parliament in March 2014. I am familiar with the hydrogeology of New Zealand and Australia in general, and have specific knowledge of NW Nelson, Wellington, Kapiti Coast, Wairarapa, Taranaki and Hawkes Bay regions in New Zealand; and the Surat and Gunnedah Basins in NSW and Queensland, Australia, as part of the Great Artesian Basin.

Please find my comments as follows:


p48 Section 8.7.1.

It is misguided to state that by creating two management zones, the Arthur Marble Aquifer (AMA) Recharge Area and the AMA Confined Area, to “reflect the different considerations for water allocation” will in fact “allow for sustainable water use and protection of the AMA and Te Waikoropupu Springs.”

Based on advice from Tasman District Council staff, and their consultants, FLAG are proposing to split the AMA into two distinct units in the Takaka Water Management Plan. The two AMA zones being:

* the Unconfined Arthur Marble Recharge Zone, and
* the Confined Arthur Marble Aquifer Zone.

However, the Unconfined AMA and the Confined AMA are intrinsically linked to form the groundwater capture zone for the Te Waikoropupu springs. Te Waikoropupu springs are recharged/supplied/fed from both the confined and unconfined zones of the Arthur Marble Aquifer, upgradient of the Te Waikoropupu Springs complex, and not merely from the unconfined Arthur Marble Aquifer alone. The confined Arthur Marble Aquifer, in which Te Waikoropupu springs is located, is recharged by the unconfined Arthur Marble Aquifer.

Restricting the “take” of a parcel of water at one end of the aquifer (the Unconfined AMA), only to have it taken downgradient (the Confined AMA), does not “allow for sustainable water use and protection of the Arthur Marble Aquifer and Te Waikoropupu springs.” The Arthur Marble Aquifer needs to be managed as a single zone in order to effectively protect Te Waikoropupu springs. It is frankly irrational to suggest that you are protecting Te Waikoropupu springs by, for instance, stopping a bore owner pumping a “parcel” of surface/groundwater in the AMA Unconfined one (when the Unconfined Zone allocation is, say, 500 l/s) but then allowing another bore owner in the Confined AMA to pump out that “parcel” of water. Therefore, the Springs do not receive that “parcel“ of water.

Added to this, TDC investigations indicate that groundwater takes between 2 to 8 years to flow from the point of recharge in the Unconfined AMA to the discharge point at Te Waikoropupu springs. So if TDC opens up the Confined AMA to greater allocations, then a “parcel” of water that was not pumped out 2 to 8 years ago so that the Springs can receive it, is now going to be pumped out of the Confined AMA zone upgradient of the Springs. Whatever the allocation limit is in the unconfined zone must apply to the confined zone, upgradient of the Springs (or apply to any part of the AMA that is found within the catchment area of Te Waikoropupu springs).


p49 Section 8.7.2 1st sentence, 3rd paragraph AND p50 2nd para AND p51 Table 10.

Where is the evidence proving there is groundwater from AMA “flowing out to sea in Golden Bay” and via sub-marine springs?

The idea that “groundwater in the Arthur Marble Aquifer must be flowing out to sea” has been arrived at by TDC as a result of a shortfall of 6-8 m3/s in their water balance modelling for the Arthur Marble Aquifer (Mueller 1991, 1992; Stewart and Thomas, 2008; Thomas and Harvey, 2013). There is no geochemical, geophysical, bathymetric or geological evidence that offshore sub-marine springs exist. Vague anecdotal evidence is not sufficient to allow these sub-marine springs to play such an important component in the conceptual groundwater model which the Takaka Water Management Plan is based upon.

Edgar (1998) completed a detailed thesis as part of her Masters in Environmental Science. She produced a considerable document with a detailed literature research. She discussed in great detail the physiology, climate, geology, geomorphology, hydrology and hydrogeology of the Takaka Valley groundwater system but of interest here is her hydrometric analysis to determine the water balance for the Arthur Marble Aquifer.

Based on her studies she refuted any need to include offshore submarine springs to account for the extra 6‐8 m3/s that hasn’t been explained by researchers to date (Mueller 1991, 1992; Stewart and Thomas, 2008). Realistically she looked at ranges of values instead of providing averages and mean values of systems. The karstic environment produces groundwater systems that exhibit extreme hydraulic behaviour. Fractures and conduits in karstic marble aquifer fill quickly but empty just as fast. Depending on the amount of sediment found in these systems determines how much storage there actually is. Mueller (1991) discussed the likelihood of storage in the Arthur Marble Aquifer saying sediment is not seen appearing at Te Waikoropupu Springs but it certainly enters the recharge sinks in the Upper Takaka River. Therefore it is possible then the Arthur Marble Aquifer could store a considerable amount of groundwater.

Edgar (1998) concluded the following on her water balance for Te Waikoropupu Arthur Marble Aquifer:

1. “Two alternate methods are outlined, namely the annual water balance method and the flow balance method.”
2. “Input and output data for an annual water balance for AMA are presented. The precipitation input is derived, along with the sum of inflows, and the total outflow.”
3. “The resulting changes in storage for AMA are calculated as ‐0.44 m3/s.”
4. “Components for the flow water balance are listed, and the generation of mean flows using TIDEDA is outlined.”
5. “The resulting change in storage for AMA is calculated at 220 l/s.”
6. “The results of both water balances suggest that AMA is relatively balanced.”

TDC has developed an oxygen-18 mass balance mathematical model based on a conceptual groundwater flow model (Stewart & Thomas, 2008). This groundwater model is based on the notion the water balance model has a shortfall of 6‐8 m3/s and that AMA is flowing out to sea via sub-marine springs.

This groundwater model forms the basis of the proposed new groundwater allocations plans for the Confined AMA set out in the FLAG document but is based on a conceptual model and mathematical modelling which are unsubstantiated due to the following reasons:

a. TDC has incorporated groundwater discharge to “sub-marine springs” into their groundwater management model treating it as fact. However, there is no geophysical and geochemical data to support whether these off-shore springs exist, let alone contain fresh water.

b. TDC has asserted a diorite dyke intrudes into the shallow levels of the Arthur Marble Aquifer in the mid-valley area. Mr Thomas has stated that this diorite intrusion splits the Arthur Marble Aquifer flow into “shallow” and “deep” groundwater flow systems, to explain the geochemistry and isotopic signatures of the groundwater found in the Arthur Marble Aquifer and Te Waikoropupu Springs. TDC state the diorite dyke directs groundwater to Te Waikoropupu springs and the rest out to sea via “sub-marine springs”. However, this diorite intrusion has not been ‘ground-truthed’ by drilling to characterise the geometry and depth of the diorite intrusion. No data analysis has been presented by TDC from monitoring bores along the coast to characterise groundwater flow from the AMA out to sea.

c. TDC’s mathematical model has been presented as unique, which is incorrect and misleading. I have showed in Appendix 1 how manipulating the non-measured contributions can produce different model outcomes.

FLAG’s comment that stated “Consumptive water takes from the Confined AMA Zone have not been included in the AMA Recharge zone accounting……..as it does not take into consideration of the amount of aquifer water going out to sea” is therefore incorrect.


p49 Section 8.7.2 4th para

Why does FLAG then have “key concerns in the AMA Recharge Zone’ with ‘risks to, and protection of, the water quality in both the confined and unconfined parts of the Arthur Marble Aquifer” if the FLAG don’t consider the Confined AMA zone is linked to the AMA Recharge Zone? The comment they make is true but is contrary to the big picture they are providing to the public which is not true.


p50 Section in 1st para (third point ‘Waingaro River’) and 3rd para

The Waingaro River contribution to the AMA Recharge Zone has gone from 100%, under the 500 l/s interim policy, to only 8%. I understand this has been determined from hydrometrics, that is, looking at river flow “gains” and “losses” to/from groundwater along the main rivers in each sub-catchment. Those portions of the Waingaro River that lose flow to the AMA groundwater system are destined for Te Waikoropupu springs. I also understand the conceptual groundwater model and oxygen-18 mass balance modelling given in Stewart and Thomas (2008), is what FLAG are referring to in their comment “The new approach to determining which takes are managed under which zone, reflects the better understanding of how the various catchments contribute to flows at Te Waikoropupu springs.”

Edgar (1998) and Mueller (1991, 1992) were used by Andrew Fenemor as the basis to come up with the 500 l/s in the first place. TDC knew about the “gains” and “losses” in flow to the Unconfined AMA Recharge Zone, including the Waingaro River, back in the 1990s. So I still think most of the numbers have changed due to the Stewart and Thomas (2008) model. I have to assume that, given the groundwater model is fundamentally unsubstantiated, these comments are flawed. If not where is the evidence for this 8%?


p51 Table 10

Regarding the claim that “The catchment account for the informal (no legal status) and proposed regimes are summarized below” – It is incorrect to refer to the 500 l/s allocation limit for Te Waikoropupu springs Recharge Zone as “informal (no legal status)”. The 500 l/s allocation limit was part of the Takaka Valley Water Management Policies, drafted by Andrew Fenemor (then Manager of Resources of the Nelson Marlborough Regional Council). They are not “informal” policies, but rather “interim” ones, pending a formal plan change in the Tasman Resource Management Plan. Mr Fenemor has indicated that these policies were formally adopted by NMRC in 1991 (on 20 August 1991, to be exact), and they continued as policies of Taman District Council after the regional council was dis-established. Here is the specific language of the Takaka Valley Water Management Policies that references the allocation limit to protect Te Waikoropupu springs:

“2. The Regional Council has set an interim limit of 0.5 m3/sec for total abstractions from the recharge zone for the Pupu Springs system.”

The 1991 policies indicated that this 500 l/s allocation limit was necessary for the “protection of Pupu Springs recharge area(s) from contamination and significant reductions in recharge” and stated that Te Waikoropupu springs “were regarded as sufficiently important to warrant interim policies.” I have read the document setting these interim policies. They are not referred to as “informal” in any part of that document.


p62 Section 8.9 Science Panel Recommendations for Te Waikoropupu springs.

Given one of the key concerns in the AMA Recharge Zone is:

“Risks to, and protection of, the function of the aquifer and the stygofauna community (animals that live in groundwater systems) which are thought to play a role in creation of the outstanding water clarity in Te Waikoropupu spring.”

I am disappointed that no attempt has been made to undertake a rigorous study to provide a baseline data set, nor any recommendations made to do so in the future as stated by the Science Panel. I would like to have the opportunity to review the whole document, given its importance, when it is made publically available.

Kind regards,
Andrea Broughton
Principal Hydrogeologist
Groundwater Solutions International

Appendix 1: TDC Oxygen-18 Mass Balance Mathematical model and Conceptual Groundwater Flow model

There is an issue surrounding the “unique solution” status Stewart and Thomas (2008) have presented by using their oxygen-18 mass balance model in Table 4. Below is an example of why the model cannot be defined as a unique solution:

The authors stated the following:

“The paper examines the evidence on recharge sources to the Arthur Marble Aquifer and establishes a new recharge model using the most reliable information. This recharge model is used with estimates of the average oxygen-18 isotope values of the three groups of recharge sources to derive an oxygen-18 isotope mass balance for the inputs and outputs of the AMA by means of a spreadsheet (Table 4).

The results show that the Main Spring is sourced mainly from the karst uplands (74%), with smaller contributions from the Upper Takaka River (18%) and valley rainfall (8%). In contrast, Fish Creek Springs are fed mainly by Upper Takaka River (50%), with valley rainfall (25%) and karst uplands (25%).The unexpected consequence of the mass balance is that much of the Upper Takaka River contribution to the aquifer (58%) must bypass the springs and be discharged via offshore springs and seeps.”

However, the following example shows how manipulating the non‐measured contributions can produce different model outcomes:

Keeping the TDC measured inflows and outflows with their oxygen-18 isotope concentrations the same, for Main Spring and Fish Creek Springs, but altering the proportion of each of the three contributing sources to those flows (that being, Karst Uplands, Upper Takaka River and Valley Rainfall) can lead to a different spring source scenario. Table 4 below is from Stewart and Thomas (2008):

After solving the various equations to provide the matrix solutions the model allows you to alter the Numbers, as written in the explanation box in the above spreadsheet. The following conclusions could then be drawn:

This recharge model example shows that the Main Spring is sourced mainly from the Karst Uplands (65%), with a smaller contribution from the Upper Takaka River (23%) and Valley Rainfall (12%). Fish Creek Springs are fed mainly by Karst Uplands (69%) similar to the Main Springs, with a smaller contribution from the Upper Takaka River (30%) and very little from Valley Rainfall (<1%).

The flow out to the supposed “marine springs” was kept in this example. However, even if they existed then the mass balance would have much of this flow still coming from the Upper Takaka River (79%) with contributions from Valley Rainfall (14%) and Karst Uplands (7%).

So following Stewart and Thomas (2008) reasoning then much of the Upper Takaka River contribution to the AMA (61%) must bypass the springs and be discharged via offshore springs and seeps. However, although there seems to be 6450 litres per second flowing out via marine springs, there is no evidence of this happening via discrete discharge vents based on temperature and electrical conductivity measurements. There is of course the possibility it is being kept in storage. Depending on the amount of sediment found in these karstic systems determines how much storage there actually is. Mueller (1991) discussed the likelihood of storage in the Arthur Marble Aquifer saying sediment is not seen appearing at Te Waikoropupu springs but it certainly enters the recharge sinks in the Upper Takaka River. Therefore it is possible then the Arthur Marble Aquifer could store a considerable amount of groundwater.

For either one of the springs, there are three degrees of freedom (the three contributing sources of water) and two constraints (the volume of water and the concentration of isotope). Therefore there is one degree of freedom that allows the data to be manipulated. Obviously, there are secondary constraints that bound the system like the model can’t have negative numbers.

In the example shown here the numbers can be freely altered. This indicates the GNS/TDC modelling does not actually tell us anything and any opening up of Arthur Marble Aquifer confined zone to greater groundwater allocation based on GNS/TDC conceptual model could lead to negative effects on the Te Waikoropupu Springs ecosystem.

The authors have based both their oxygen-18 and conceptual model on averages and means, along with many assumed conditions (as is typical of most models). I don’t understand how Stewart and Thomas (2008) can deal with averages and means as inputs into their model and then come up with a unique solution as shown in Table 4 of their report.

In nature the karstic marble groundwater system operates in extremes. The groundwater (and pressure) level increases quickly with sustained recharge, but also loses it rapidly during drought or over abstraction. I am not convinced this model actually emulates this. The new conceptual recharge model should be dealing in ranges and scenarios.

I have not seen a sensitivity analyses included as part of this model showing how sensitive the specific input parameters are to both the model outputs, as they were not included in the Stewart and Thomas (2008) report.

The authors think a major conduit or conduits must be involved in channelling flow to Te Waikoropupu Springs (as well as the presence of the diorite sill), because of the very substantial flow (7.4 m3/s on average from the deep system at the Main Spring) and because such a large proportion of the known recharge of the deep system emerges at the springs (80% at Main Spring). However, the authors think such major conduits probably act as collecting vessels (or short‐circuits through to the springs) from smaller channels/fissures and are not present throughout the Arthur Marble Aquifer system.

An anonymous reviewer of the Stewart and Thomas (2008) paper had several questions regarding this and these are given below along with the authors’ comments:

1. How is this seen in the modelling? The authors accept the modelling is limited and that they are not clear about the relative roles of fissured‐porous matrix and karstic channels in the deep system.

2. If the dispersion system diffusive exchange is invoked between the (mobile water) fissures and the (relatively immobile water) porous matrix, or is it diffusion from channels (secondary fissures), and how is this adapted in the authors modelling? The authors stated that they are trying to understand this and will address this future work.

3. Why is it possible that older water is resident in less accessible parts of the porous matrix? Surely the porous matrix is low‐permeability and any exchange is then diffusively controlled rather than head‐potential controlled. The authors agreed this was a problem for the age modelling because the porous matrix could act to suck the age out of the fissure (tritium, CFCs).