You are here: Publications > State of the Environment > Groundwater Quality in New Zealand > Online version >

3 Groundwater quality trends 1995–2006

3.1 Categorisation of monitoring sites based on trends in groundwater quality

Daughney and Reeves (2006) developed a classification scheme for the NGMP in order to simplify the interpretation of groundwater quality trends at the national scale (see Table 9). As for the classification scheme based on medians (see section 2.1), this classification scheme for trends was developed for the NGMP and has not been rigorously tested for its applicability to non-NGMP monitoring sites. Nonetheless, it provides a convenient means of summarising the patterns of change among several different groundwater quality parameters. Trend-based categorical assignments are based only on groundwater quality and do not specifically account for factors such as well depth or aquifer confinement (see section 4).

Table 9: General characteristics of trend-based groundwater categories defined by Daughney and Reeves (2006)

View general characteristics of trend-based groundwater categories defined by Daughney and Reeves (2006) (large table)

Across New Zealand, for all monitoring sites for which a trend-based categorical assignment could be made (n = 918), the majority (68%) fall into trend category WR, indicating that changes in groundwater quality are occurring slowly, if at all (see Table 10). The remaining sites are partitioned roughly equally into trend categories DIL, DEG, IMP and SiO₂. Figures 7 and 8 display the trend-based categorical assignments in map form (site-specific trend category assignments are given in Spreadsheet 1). The regions with the proportions of monitoring sites assigned to trend category DIL are Northland (18%), Gisborne (14%) and the Bay of Plenty (12%), possibly due to changes in hydrogeological regime related to the pumping of coastal aquifers. The regions with the highest proportions of monitoring sites assigned to trend category DEG are Southland, Tasman and Waikato. Southland is also the region with the highest proportion of monitoring sites assigned to trend category IMP. This indicates that groundwater quality is improving over time in some parts of Southland and becoming worse in others. Note that Southland also shows proportionally high current levels of impact (status-based category 1A; see Table 6). This indicates that one of the regions in which groundwater is currently most affected is also a region in which groundwater quality is changing (improving or degrading) most rapidly over time. Some regions in which the current status of groundwater quality is quite good, such as the West Coast, also show an above-average proportion of monitoring sites in trend category DEG. Although this assessment is based on a small number of monitoring sites, it may indicate that one of New Zealand’s least affected regions may not remain that way for very long.

Table 10: Percentage of monitoring sites assigned to each trend-based groundwater quality category, and total number (n) and percentage (%n) of sites in each category overall

Council	DIL	DEG	IMP	SiO₂	WR
ARC	4.2%	0.0%	4.2%	0.0%	91.7%
EBOP*	12.3	7.0	15.8	1.8	63.2
ECAN	1.4	8.3	1.8	0.0	88.4
ES	0.0	28.9	40.0	8.9	22.2
EW	8.2	17.5	19.6	1.0	53.6
GDC	14.3	13.0	3.9	16.9	51.9
GWRC	2.8	7.0	23.9	1.4	64.8
HBRC	4.0	8.0	4.0	0.0	84.0
MDC	4.3	4.3	13.0	8.7	69.6
MWRC	3.2	9.7	12.9	3.2	71.0
NRC*	17.8	15.6	15.6	17.8	33.3
ORC	7.8	14.4	5.6	6.7	65.6
TDC*	6.3	25.0	6.3	0.0	62.5
TRC	0.0	14.3	0.0	14.3	71.4
WCRC	0.0	12.5	0.0	0.0	87.5
n	53	106	94	38	627
%n	5.8	11.5	10.2	4.1	68.3

* Regions in which categorical assignments may be biased by groundwater under marine influence.

Figure 7: Assignment of North Island monitoring sites to trend-based groundwater quality categories

See figure at its full size (including text description).

Figure 8: Assignment of South Island monitoring sites to trend-based groundwater quality categories

See figure at its full size (including text description).

3.2 Quantification of groundwater quality trends at the national scale

For most of the 33 parameter categories defined in section 1.2, a significant increasing or decreasing trend is detectable at about 25% of the monitoring sites at which the trend assessments could be performed (see Table 11). Trend assessments for trace elements are not given in Table 11 because they are analysed at relatively few sites and many of the reported concentrations are near or below the detection limit. For the remaining parameters, when all monitoring sites are considered together as a single group the proportions of sites showing increasing and decreasing trends are the same to within about 10%, as observed in a previous study of the NGMP data (Daughney and Reeves, 2006).

Previous studies in New Zealand and overseas have shown that for parameters such as Cl, K and NO₃-N, the proportion of sites showing increasing trends can exceed the proportion of sites showing decreasing trends by more than 10%, presumably due to intensification of agricultural activity (Frappaporti et al, 1994; Daughney and Reeves, 2006). However, such patterns are not detectable in this investigation, even if the data set is limited to shallow wells (< 10 m deep). We acknowledge that for some specific sites the ability to detect trends for some parameters is limited by the low number of measurements made between 1995 and 2006. However, the general conclusions reached above are not altered if the analysis is limited to sites where the relevant parameter has been measured more than three times, more than five times, or more than 10 times.

Table 11: Total number of monitoring sites at which trend assessments could be performed (n), and percentages without significant trends (%N) or with significant increasing (%INCR) or significant decreasing (%DECR) trends

Parameter		n	%INCR	%DECR	%N
Major	Ca	903	17.9%	8.5%	73.5%
	Cl	915	18.8	10.3	70.9
	HCO₃	941	20.2	10.6	69.2
	K	895	11.6	7.6	80.8
	Mg	889	13.0	13.4	73.6
	Na	911	12.4	13.7	73.9
	NO₃-N	878	13.1	11.0	75.9
	SiO₂	564	4.4	4.4	91.1
	SO₄	852	20.5	11.0	68.4
	TDS*	252	11.9	8.3	79.8
Minor or biological	B	280	7.5	4.6	87.9
	Br	175	8.6	14.3	77.1
	F	258	5.0	5.4	89.5
	Fe	643	6.8	23.5	69.7
	Li	75	4.0	4.0	92.0
	Mn	539	7.6	19.3	73.1
	NH₄-N	652	9.8	9.2	81.0
	PO₄-P	467	16.3	9.0	74.7
	MaxMicro	297	7.1	3.0	89.9
Other	Cond.	964	16.5	14.2	69.3
	pH	952	13.8	10.4	75.8
	Temp.	864	6.5	6.0	87.5

* All trend assessments use measured TDS instead of calculated TDS.

In general, the rate of change in groundwater quality is slow, with national median values for absolute trend magnitude being less than 0.5 and 0.01 mg/L per year for most major and minor elements, respectively (Figure 9, Table 12). Compared to relevant median concentrations from Table 7, this equates to national median relative trend magnitudes of less than ±2% and ±5% per year for most major and minor elements considered in this study, respectively. Similar absolute and relative rates of change for major and minor elements have been reported for the NGMP (Daughney and Reeves, 2006). For comparison, a “purely arbitrary” cut-off of ±1% has recently been used to identify relative trends in river water quality that are both statistically significant and meaningful from an environmental management perspective (Scarsbrook, 2006). Further work is required to determine what threshold should be used to identify relative trends in groundwater quality that are important from a management perspective.

Absolute rates of change are uncorrelated or only weakly correlated to median concentrations. For example, there is no relationship between the median concentration of NO₃-N at a particular site and the rate at which NO₃-N is changing over time. Trend assessments for NO₃-N, Fe and Mn, microbiological parameters and salinity are discussed in more detail below due to their recognised importance as indicators of groundwater quality in New Zealand (see section 2.2).

Table 12: National absolute and relative rates of change in groundwater quality parameters (units per year)

Parameter		Absolute trend (units per year)				Relative median trend (%)*
Parameter		Units	Minimum	Median	Maximum	Relative median trend (%)*
Major	Ca	mg/L	-25.47	0.27	10.54	1.7%
	Cl	mg/L	-26.93	0.16	49.06	1.1
	HCO₃	mg/L	-33.57	0.78	18.26	1.2
	K	mg/L	-3.48	0.02	4.41	1.5
	Mg	mg/L	-5.47	-0.02	2.96	-0.4
	Na	mg/L	-69.00	-0.08	33.86	-0.5
	NO₃-N	mg/L	-1.96	0.03	2.91	0.2
	SiO₂	mg/L	-2.72	0.01	3.73	-0.04
	SO₄	mg/L	-3.36	0.14	11.98	2.2
	TDS	mg/L	-24.94	2.62	37.60	1.8
Minor or biological	B	mg/L	-0.025	0.001	0.164	5.3
	Br	mg/L	-0.899	-0.003	0.263	-3.2
	F	mg/L	-0.118	-0.002	0.014	-1.9
	Fe	mg/L	-2.170	-0.006	3.401	-18.3
	Li	mg/L	-0.001	0.000	0.004	-0.2
	Mn	mg/L	-1.047	-0.002	1.864	-15.2
	NH₄-N	mg/L	-1.366	0.000	0.522	6.1
	PO₄-P	mg/L	-0.474	0.001	0.033	6.2
	MaxMicro	cfu/100 ml	-13.56	0.68	391.96	–
Other	Cond.	uS/cm	-149.0	1.48	159.3	0.7
	pH	pH units	-0.40	0.01	0.35	0.3
	Temp.	°C	-8.46	0.02	1.11	0.2

* Relative median rates of change (% per year) are calculated by dividing the median absolute trend magnitude by the relevant median concentration from Table 7.

Figure 9: National percentiles for rates of change in groundwater quality parameters, including (a) major elements and (b) minor elements and microbiological indicators

See figure at its full size (including text description).

Nitrate

Significant increasing trends in NO₃-N concentration are detectable at around 13% of the monitoring sites considered in this investigation (see Table 11). Significant decreasing trends in NO₃-N concentration are detectable at a roughly equivalent percentage of monitoring sites. In agreement with the trend-based categorisation, sites with detectable trends in NO₃-N are found in all regions of the country (see Appendix 1), especially in Waikato and Southland, but also in regions such as the West Coast, where water quality is generally good, and Gisborne, where many groundwater systems are oxygen-poor and would not be expected to contain significant concentrations of NO₃-N.

Iron and manganese

Significant trends in Fe and/or Mn are detectable at about 30% of the monitoring sites at which trend calculations could be performed (see Table 11). The majority of these sites show decreasing trends in Fe and/or Mn over time, which is unusual compared to other parameters considered in this study (most of which show roughly equal proportions of increasing and decreasing trends). The disproportionately high number of sites with decreasing trends in Fe and/or Mn might be caused by improvements in sampling and analytical methods between 1995 and 2006. The statistical methods used for trend detection require the replacement of censored values with some fraction of the detection limit. If the majority of results at a particular site are below the detection limit, then the statistical tests could conclude that a decreasing trend is present if the detection limits have decreased over time (eg, due to the improvement of analytical methods).

Improvements in sampling methodology could also lead to decreasing trends in Fe and/or Mn. Within the 11-year timeframe of this study, several regional councils started to field-filter samples for Fe and Mn analysis, whereas in the past a raw sample might have been collected for this purpose. Dissolved Fe and Mn concentrations in filtered samples are often many times less than total concentrations in raw samples, and so the change to collecting the former could lead to identification of decreasing temporal trends.

It is also possible that human influence could cause the disproportionately high number of decreasing trends in Fe and/or Mn. This could arise, for example, by over-abstraction of groundwater near a stream, which could potentially lead to an increased proportion of oxidised surface water (with low Fe and Mn concentration) being drawn into an aquifer. It will be important to repeat these analyses again in a few years to identify the cause and significance of trends in Fe and Mn, ideally with more information on the source of the water being monitored.

Microbiological parameters

Trend assessments were performed using the microbiological parameter with the highest reported value for each sampling event at each site (MaxMicro). Under this approach, the microbiological indicators display an extremely wide range of trend magnitudes. The extreme values in calculated trends presumably arise because of a small number of analytical results that are substantially above background. Sporadically elevated microbial counts could be related to flooding events at sites with poor well-head protection or to contamination during sampling, making site-specific trend assessments for microbiological indicators difficult to interpret.

However, national-level statistics can be interpreted, with caution. In this study, significant increasing and decreasing trends are detectable at about 7% and 3% of sites, respectively, and the median absolute rate of change is near zero. Similar results are obtained for most of the non-biological water quality parameters considered in this report. Overall it appears that microbial contamination of New Zealand’s aquifers is not worsening very dramatically over time, but this tentative conclusion should be re-tested by repeating a similar study in a few years.

Salinity

Electrical conductivity is a useful measure of salinity and TDS. Significant increasing and decreasing trends in conductivity are detectable at roughly equal proportions of monitoring sites considered in this study (16.5% and 14.2%, respectively). Sites with increasing trends in conductivity are found in most areas of New Zealand, especially Waikato, Southland and Wellington. At some sites, especially those assigned to category 1 (oxidised), the trends in electrical conductivity are mirrored (in direction and relative magnitude) by trends in NO₃-N. More research on these relationships may indicate that they can be used in combination as indicators of human influence.

[ Previous|Next