Executive Summary
Aim of the study
The Office of the Auditor General (OAG) requested assistance from NIWA on technical matters underpinning OAG's audit of freshwater quality management by regional councils in New Zealand. This report details results of a study performed by NIWA that responds to two specific questions posed by OAG:
- Do regional councils have effective methods to gather information about and monitor the quality of freshwater?
- Over the last 10 years, what is the state and trends in water quality as indicated by state of the environment monitoring data collected by regional councils and NIWA?
We have addressed these questions for four regional councils specified by OAG:
- Environment Southland
- Taranaki Regional Council
- Horizons Regional Council
- Environment Waikato
To answer the first question we assessed the methods used to monitor the quality of freshwater in each of the following regions: Southland, Taranaki, Horizons, and Waikato. We first obtained information (metadata) from the regional councils that described their State of Environment (SoE) for physical, chemical, microbiological and biological aspects of water quality monitoring programs for rivers, lakes and groundwater. Specifically, we obtained the locations and the details of monitoring sites, the frequency of monitoring, the variables analysed, the QA/QC and data storage procedures. From this information we assessed the network and monitoring programmes from technical perspectives.
To answer the second question we analysed state and trends in water quality data for rivers and streams (hereafter referred to as "rivers") for the ten year period up to and including 2009. We included sites in the National Water Quality Monitoring Network (NRWQN) that is run by NIWA that are within the four regions.
Findings to question 1: Regional council state of environment monitoring of freshwater
We consider that the four regional councils surveyed have well-planned and operated networks for assessing the current state and long term trends in physical and chemical quality of freshwaters. All four councils have monitoring networks with SoE sites for rivers, lakes and groundwater that are distributed over their regions in a reasonably representative manner (i.e. where the number of sites in different catchments or types of water bodies is commensurate with the overall importance and quantity of water bodies of that type). We also consider that all four councils are monitoring a comprehensive suite of relevant physical, chemical, microbiological and biological variables at a suitable frequency, and that they generally have adequate QA/QC and data storage procedures.
We made a specific evaluation of the adequacy of number of sites included in the river SoE monitoring networks in each regions. Our evaluation was based on whether the networks had sufficient statistical power to detect large scale patterns, defined by River Environment Classification (REC) categories, in water quality state and trends. The REC categories grouped the monitoring sites in four categories on the basis of the dominant topography of their catchments; Low-elevation, Hills, Lakes or Mountains. Such groupings provide insights into the causes of spatial patterns of water quality state and trends in relation to environmental and human factors and can be used to describe how well a network of sites represents the overall environmental variation within a region.
We used river water quality data provided by the councils and from the NRWQN for the 10 year period ending 2009 that had been collected at a quarterly or monthly basis. Trend analysis of water quality data, and to a lesser extent the calculation of the central tendency to evaluate state (i.e. mean or median conditions), is only meaningful if calculated using a continuous time series of observations of sufficient length. Not all the river water quality data sets provided by the regional councils were sufficiently complete to provide robust trend analyses for the 10-year period of our trend analysis. We limited our analysis to data sets for which at least 80% of sample occasions had data. Accepting time series with more missed sample occasions would result in more insignificant results. Thus, for sites that were monitored quarterly, we included sites that had data for 32 quarters of 40 possible quarters and for sites that were monitored monthly we included sites that had data for 96 of 120 possible months. These criteria restricted the number of monitoring sites that we used to estimate state and trends. Environment Southland had approximately 56 sites (depending on variable) that met our criteria for trend analysis. Taranaki and Horizons Regional Councils had 13 and 17 SoE sites (depending on variable) that met our criteria respectively. Environment Waikato had a total of 113 SoE sites which met our criteria for trend analysis.
The river water quality monitoring data for Southland and Waikato were able to detect detailed patterns in both state and trends (i.e. statistically significant differences in state and significant overall trends were found for many REC categories and variables). The data for 12 sites in the Taranaki region was sufficient to detect patterns albeit for fewer REC categories and variables than for Southland and Waikato. The Horizons dataset was barely adequate to describe large scale patterns in water quality state and trends in the region. This is because, in the past, Horizons have employed a system of "rolling SoE sites" whereby some sites have been monitored on a rolling basis, i.e. once every three years 12 months of monthly sampling has been undertaken. This practice is no longer carried out by Horizons Regional Council and the number of SoE sites in the monitoring network has been increased to 63. The large differences between regions in the total number of currently active SoE sites reflect, to a degree, the size of the regions.
We do not consider that any of the regions have too few sites to describe regional patterns in water quality. More sites would provide more detailed information at a larger number of specific locations. A case could be made for further SoE sites added to the current Taranaki regional network of 12 sites (which is comparatively low compared to the 113 sites in Waikato, for example). However, we consider it unlikely that the overall regional picture of water quality state and trends would be greatly different if the number of sites in any of the regions were increased. An important point is that if SoE monitoring is to be of maximum benefit (i.e. be suitable for robust trend analysis) it must be frequent and consistent and this requires an ongoing commitment by the regional councils. However, we acknowledge that trend analysis, while desirable, is not necessarily the most important aspect of SoE monitoring. A trade off between cost, coverage of a region's water bodies and continuity of the time series has to be made.
We did find that Environment Southland and Horizons need to lower detection limits for some water chemistry variables in order to detect trends in currently high quality water bodies. Ideally councils should measure flows on sampling occasions, so as to assist with interpreting water quality data and to enable both trend analysis and estimation of contaminant loads. Where flows are not measured at water quality sites it is common practice to estimate flows from suitable (close) gauged rivers. We consider that the uncertainties associated with flow estimates should be evaluated so that the robustness of trend analyses based on these flows can be assessed.
The four regional councils surveyed monitor biological characteristics in rivers including periphyton (algae that grows on the bed of rivers) and macro-invertebrates (invertebrate animals that live on the bed of rivers). Biological organisms integrate and express the effect of water quality and habitat over time and provide an index of the ecological health of waters. Because living organisms express (in part) the effect of the historic flux of contaminants, they need not be sampled as frequently as water quality. The biological sampling programs of all councils started in the mid 1990s and there has been a consistent effort to monitor at least annually. Environment Southland, Horizons Regional Council and Environment Waikato sample invertebrates annually, whereas Taranaki Regional Council sample invertebrates twice a year. Annual biological sampling is generally performed during summer in order to assess conditions during the period of low flow, which is generally the period of greatest ecological stress. Annual sampling is subject to occasional bias by atypical conditions such as unseasonal flooding but is generally considered to be suitable (and is the protocol of the NRWQN, for example).
All four regional councils carry out SoE monitoring on lakes. These programs are focussed on the management of individual lakes (e.g., cyanobacteria blooms during summer, long term eutrophication). These programs are targeted to specific "iconic" lakes within each region, which is practical and appropriate. All four regional councils have extensive groundwater monitoring SoE programmes that represent the major aquifers in the regions. Monitoring is predominantly carried out on an annual basis and in accordance with the national protocols.
Findings to question 2: State and trends in water quality
We analysed state and trends in river water quality data because SoE monitoring of freshwaters is most comprehensively and consistently carried out on this type of water body (in terms of the time period for which monitoring has been carried out, sample frequency, variables analysed and intensity of sampling). This study did not analyse state and trends in lake or groundwater quality data. Lake data is collected in a less consistent manner across the regions due to differences in the distribution of lakes (e.g., Taranaki and Horizons have few iconic lakes) and because of differences in the intensity of lake management issues. There are also differences in how groundwater is monitored across the regional councils reflecting differing regional focuses of the groundwater programmes. Most groundwater monitoring programmes indicate stable water chemistry other than for nitrate, which is usually monitored in separate (non-SoE) programmes. There have been recent national studies of the state and trends of lakes1 across New Zealand for the ten year period up to and including 2009 and groundwater2 up to and including 2008. In addition, we did not analyse state and trends in biological variables (e.g., macro-invertebrates and periphyton data). Again, this was because of differences in the time period for which biological monitoring has been carried out, sampling frequencies and range of variables analysed by the four regions. These differences reflect differing regional focuses for biological monitoring programmes.
We used data supplied by the councils and from the NRWQN to assess river water quality state in each region. Ten physical and chemical variables were assessed; black disc water clarity, conductivity, dissolved reactive phosphorus, total phosphorus, ammonium, oxidised nitrogen, total nitrogen, E. coli, faecal coliform, and in a small number of cases total suspended solids. Nutrient species (oxidised nitrogen, total nitrogen, ammoniacal nitrogen, dissolved reactive phosphorus and total phosphorus) were included because they stimulate the growth of plants, including algae, which can be either suspended in the water column of lakes and rivers or attached to substrates (periphyton). Nutrient contamination results from point and non-point source discharges and is strongly associated with intensive land use. High nutrients can promote excessive ('nuisance') growth of plants that, in turn, can smother habitat, produce adverse fluctuations in dissolved oxygen and pH, and impede flows and block water intakes. Excess plants in water bodies also have and detrimental effects on aesthetics and human uses causing changes to water colour, odour and the general physical nature of the environment. Nitrogen and ammoniacal nitrogen are also toxicants that can adversely affect both ecosystems and human health. Effects of nitrogen on human health mean that it is a key variable that is monitored in ground water. We note that Regional councils routinely measure dissolved oxygen and pH which are water quality variables that are strongly influenced by the growth of plants in water bodies. These variables fluctuate over the course of a day due to the metabolic cycles of plants. This means that one-off samples of dissolved oxygen and pH are not particularly useful as SoE variables because they must be interpreted with reference to the time of day that the sample was taken.
Conductivity is routinely measured by the four councils but conductivity itself does not have adverse effects, at least in New Zealand's 'dilute' waters. However, conductivity is a general indicator of ionic constituents including nutrients. Trends in conductivity can indicate changes in water quality due to human activities. Visual water clarity and suspended solids are monitored because they are associated with the attenuation of light due to contaminants that are suspended in the water column and because settleable solids have the potential for smothering the beds of rivers and downstream water bodies. Visual clarity is generally measured as the sighting range of a black disc (e.g. MfE 1994). Low visual clarity has ecosystem effects, including changes in animal behaviour. Water clarity also has implications for contact recreation due to its effect on human visibility through water. Councils include bacterial variables in monitoring for rivers and lakes. Faecal coliforms and E. coli indicate the presence of human or animal faeces and the associated risk of infectious disease from waterborne pathogens for both humans via contact recreation and drinking water and livestock via drinking water.
Overview of state analysis
We used the median value of each of the variables at the sites (or the 95th percentile for E. coli) as a measure of the state and compared these to guideline "trigger values" for water quality. The trigger values are not national standards but rather, have been devised to assess the levels of physical and chemical stressors which might have ecological or biological effects. Rather than implying that there will be ecological and biological effects caused by increased levels of physical and chemical stressors, exceedances of trigger levels (referrer to here as "failing" guidelines) indicates cause for further investigation of water quality issues.
Our analysis showed that water quality state between sites within regions was highly variable. The state of individual sites also showed strong variation between variables within sites (i.e. sites can meet guidelines for some variables and not for others). As outlined in Table A, across all regions water quality had strong relationships with REC Topography categories with the highest water quality (e.g., highest clarity, lowest conductivity, lowest nutrients and lowest indicator bacteria) generally occurring in the Mountain, Lake and Hill Topography categories. Low-elevation sites usually failed water quality guidelines for many variables and Hill Topography categories often failed (Table A).
Overview of trend analysis
Trend direction and strength for the ten physical and chemical variables over the ten year period from 2000 to 2009 were quantified using the non-parametric Seasonal Kendall Sen Slope Estimator (SKSE). The SKSE is a commonly used method for estimating trends in data that are subject to appreciable seasonality such as water quality data. Values of the SKSE were normalised by dividing by the median and normalising to 100 to give the relative SKSE (RSKSE; %), allowing for direct comparison between sites measured as per cent change per year. A positive RSKSE value indicates an increasing trend, while a negative RSKSE value indicates a decreasing trend. The RSKSE values were also associated with a test of significance. If P is "small" (i.e. P < 0.05), it can be concluded that the observed trend is most unlikely to have arisen by chance.
Table A:
Water quality state by region and variable for sites grouped by REC Topography category. NA = no sites in the category. Pass = the median of the site median values was acceptable with respect to water quality guidelines3. Fail = the median of the site median values was not acceptable with respect to the water quality guideline. The categories L, H, Lk and M refer to catchments dominated by Low-elevation, Hill, Lake or Mountain topography.
| Region | Variable | REC Topography category | |||
|---|---|---|---|---|---|
| L | H | Lk | M | ||
| Southland | Clarity | Fail | Pass | Pass | Pass |
| Dissolved reactive phosphorus | Fail | Pass | Pass | Pass | |
| Escherichia coli | Fail | Fail | Pass | Pass | |
| Faecal coliforms | Fail | Pass | Pass | Pass | |
| Ammoniacal nitrogen | Fail | Pass | Pass | Pass | |
| Total nitrogen | Fail | Pass | Pass | Pass | |
| Total phosphorus | Fail | Pass | Pass | Pass | |
| Oxidised nitrogen | Fail | Pass | Pass | Pass | |
| Taranaki | Clarity | Fail | Pass | NA | NA |
| Dissolved reactive phosphorus | Fail | Fail | NA | NA | |
| Escherichia coli | Fail | Fail | NA | NA | |
| Faecal coliforms | Pass | Pass | NA | NA | |
| Ammoniacal nitrogen | Pass | Pass | NA | NA | |
| Total nitrogen | Fail | Pass | NA | NA | |
| Total phosphorus | Pass | Pass | NA | NA | |
| Oxidised nitrogen | Fail | Pass | NA | NA | |
| Horizons | Clarity | Fail | Fail | NA | Fail |
| Dissolved reactive phosphorus | Fail | Fail | NA | Pass | |
| Escherichia coli | Fail | Fail | NA | NA | |
| Faecal coliforms | NA | NA | NA | NA | |
| Ammoniacal nitrogen | Pass | Pass | NA | Pass | |
| Total nitrogen | Fail | Pass | NA | Pass | |
| Total phosphorus | Fail | Pass | NA | Pass | |
| Oxidised nitrogen | Fail | Pass | NA | Pass | |
| Waikato | Clarity | Fail | Fail | Fail | Pass |
| Dissolved reactive phosphorus | Fail | Fail | Fail | Fail | |
| Escherichia coli | Fail | Fail | Pass | NA | |
| Faecal coliforms | Fail | Pass | Pass | NA | |
| Ammoniacal nitrogen | Pass | Pass | Pass | Pass | |
| Total nitrogen | Fail | Pass | Pass | Pass | |
| Total phosphorus | Fail | Fail | Fail | Pass | |
| Oxidised nitrogen | Fail | Pass | Pass | Pass | |
Table B:
Ten-year overall trends by region and variable for sites grouped by REC Topography category. NA = less than 3 sites in the category (therefore an overall trend could not be assessed), NS = No significant trend for the category, Deg = a degrading trend for the category, Imp = an improving trend for the category. The categories L, H, Lk and M refer to catchments dominated by Low-elevation, Hill, Lake or Mountain topography.
| Region | Variable | REC Topography category | |||
|---|---|---|---|---|---|
| L | H | Lk | M | ||
| Southland | Clarity | NS | NS | NS | NA |
| COND | NS | Deg | NS | NA | |
| Dissolved reactive phosphorus | NA | Imp | NS | NA | |
| Escherichia coli | NS | Imp | NA | NA | |
| Faecal coliforms | NS | Imp | NA | NA | |
| Ammoniacal nitrogen | Deg | NS | NS | NA | |
| Oxidised nitrogen | Deg | Deg | NS | NA | |
| Total nitrogen | Deg | NS | NS | NA | |
| Total phosphorus | NS | NS | NS | NA | |
| Taranaki | Clarity | Deg | NS | NA | NA |
| COND | NS | NS | NA | NA | |
| Dissolved reactive phosphorus | NS | NS | NA | NA | |
| Escherichia coli | NS | NS | NA | NA | |
| Faecal coliforms | NS | NS | NA | NA | |
| Ammoniacal nitrogen | Deg | NS | NA | NA | |
| Oxidised nitrogen | NS | NS | NA | NA | |
| Total nitrogen | NS | NS | NA | NA | |
| Total phosphorus | NS | NS | NA | NA | |
| Horizons | Clarity | NS | NS | NA | NA |
| COND | NS | NS | NA | NA | |
| Dissolved reactive phosphorus | NS | NS | NA | NA | |
| Escherichia coli | NS | NS | NA | NA | |
| Faecal coliforms | NA | NA | NA | NA | |
| Ammoniacal nitrogen | NS | NS | NA | NA | |
| Oxidised nitrogen | NS | NS | NA | NA | |
| Total nitrogen | NS | NA | NA | NA | |
| Total phosphorus | NS | NA | NA | NA | |
| Waikato | Clarity | Deg | Deg | Deg | NA |
| COND | Deg | NS | Deg | NA | |
| Dissolved reactive phosphorus | Imp | Imp | NS | NA | |
| Escherichia coli | Deg | NS | NS | NA | |
| Faecal coliforms | Deg | NS | NS | NA | |
| Ammoniacal nitrogen | Imp | NS | NS | NA | |
| Oxidised nitrogen | Deg | Deg | Deg | NA | |
| Total nitrogen | Deg | Deg | Deg | NA | |
| Total phosphorus | Imp | Imp | NS | NA | |
Trend direction and strength at individual sites showed strong variation across variables. An overview of the trends in each region's river water quality is provided by grouping sites by REC Topography categories (Table B). We deemed that there was an overall trend for a REC Topography category if the number of sites that exhibited that trend were greater than could be expected by chance. This was formally evaluated with the Binomial test under the null hypothesis that degrading and improving trends were equally likely. In this manner we found overall degrading trends in clarity in Taranaki and Waikato, degrading trends in conductivity in Waikato, improving trends in Dissolved reactive phosphorus in Southland, Taranaki and Waikato, a degrading trend in E.coli in Waikato, improving trends in Ammoniacal nitrogen in Horizons and Waikato and a degrading trend in ammoniacal nitrogen in Taranaki, degrading trends in oxidised nitrogen in Southland, and Waikato, degrading trend in total nitrogen in Southland and Waikato and improving trend in total phosphorus in Horizons and Waikato. When these trends were broken down by REC categories there was a predominance of degrading trends in Low-elevation and Hill Topography and the Pasture Land-cover categories. These results suggest that water quality degraded over the ten year period in Low-elevation areas and in catchments dominated by pastoral land cover (often in low elevation areas). There were however, generally improving trends in dissolved reactive phosphorus and total phosphorus in all of the regions.
The degrading trends are consistent with increasing intensification of agricultural land use in New Zealand's low elevation and hill country pastoral landscapes. However, there were improving trends in dissolved reactive phosphorus and total phosphorus. The improving trend in phosphorus shown in this study is consistent with a recent national study4 and may be attributable to two factors. First, there has been increase in phosphorus fertiliser costs over the last decade (an 86% rise in 2008 alone). Second, there has also recently been very active management of soil phosphorus (Olsen-P) levels by the pastoral industry. However, the degrading trend in nitrogen may be attributable to increased farm production. For example, there has been a 20% rise in dairy-farm production. This increase in production is associated with leaching of nitrogen from pasture soils for which there are not currently adequate mitigation methods.
Two points of caution need to be borne in mind in using the state and trends analysis in this report to draw conclusions concerning regional councils' management of freshwater. First, we compared the existing state to non-statutory guideline values. To fully assess whether regional councils are meeting (their own) standards, the standards defined in statutory plans would need to be compared with the state information derived in this study. Second, trends provide information about change in water quality over time but also need to be considered within the broader statutory framework that regional councils have set. That analysis is outside the scope of this report.
Reviewed by:
Rob Davies-Colley
Approved for release by:
Clive Howard-Williams
1: Verburg, P., K. Hamill, M. Unwin and J. Abell. 2010. Lake Water Quality in New Zealand 2010 : Status and Trends. NIWA Client Report: HAM2010-107. Hamilton. 52p.
2: Daughney, C.; Randall, M. 2009. National Groundwater Quality Indicators Update: State and Trends 1995-2008, GNS Science Consultancy Report 2009/145. 60p. Prepared for Ministry for the Environment, Wellington, New Zealand.
3: The guidelines are provided in Table 2 of this report.
4: Ballantine, D., D. Booker, M. Unwin and T. Snelder. 2010. Analysis of National River Water Quality Data for the Period 1998–2007. NIWA. NIWA Client Report: CHC2010-038 72p.