• About this site|
• Site index|
• Contact us|
• Jobs

An Assessment of Regional Council Water Quality Data (1996–2002) in Support of National State of Environment Reporting (ENZ07)

1 Introduction

2 Methods

3 Summary Tables

4 Changes in Water Quality State (1997–2002) and Land Use Differences

5 Summary

6 References

Appendix 1: Breakdown of Monitoring Sites by REC Classes

2 Methods

2.1 The dataset

Larned et al (2003) compiled a water quality dataset comprising physico-chemical measurements from 15 regional councils and unitary authorities and NIWA’s National River Water Quality Network (NRWQN: 77 sites). The full dataset included multiple sampling events from 996 sites across the country, but a number of these sites did not meet predefined criteria for state and trend analyses (see Larned et al 2003). Following data grooming, a final dataset consisting of 622 sites for state and 386 sites for trend analysis was used to assess spatial and temporal patterns in water quality. The River Environment Classification (REC) was used as a spatial framework to partition variability.

In this report, we summarise data from a 545 sites, excluding the 77 sites from the NRWQN that were included in the assessment of Larned et al (2003). Spatial and temporal patterns in water quality indicators at these NRWQN sites have been reported more fully by Scarsbrook (2006). The locations of the sampling sites used in this report are given in Figure 1.

Figure 1: Locations of 545 sampling sites

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

Note: Map colours relate to REC land cover level following the convention used by Scarsbrook (2006).

2.2 Water quality indicators

Larned et al (2003) reported on state and trends in five commonly used water quality indicators. The parameters were nitrate + nitrite (NO_x), ammoniacal nitrogen (NH₄), dissolved reactive phosphorus (DRP), and Escherichia coli bacterial concentrations, and water clarity measured by the black disk clarity procedure (BDC: Davies-Colley 1988). In the present report, we provide statistical summaries for these same five physico-chemical indicators of water quality, and provide an assessment of percentage of sites exceeding accepted water quality guideline values (Table 1). Note that ANZECC and ARMCANZ (2000) provide separate guidelines for protection of aquatic ecosystem values in upland and lowland streams, but do not include an ‘average’. We have included an average of the upland and lowland guideline values to allow for a national perspective on water quality across all stream sites.

Table 1: Guideline water quality values for protection of New Zealand river ecosystems

Note: ANZECC and ARMCANZZ 2000: all values except recreation, and human health: MfE/MoH 2003.

MPN = Most Probable Number.

2.3 Water quality summaries

The summaries provided in this report cover the period 1997–2002. It is recognised that more up-to-date regional council data is available, but the timeframe for this study did not allow for collation of new data.

Larned et al (2003) assigned all monitoring sites to REC classes. In the present report, we use this prior classification to provide summaries across different land uses (ie, pastoral, natural and urban), and for sites in both upland and lowland source-of-flow (SOF) categories within the land cover classes (Appendix 1 gives a breakdown of monitoring sites by REC classes). The pastoral land cover class comprises sites where pastoral agriculture is the dominant land use in the catchment upstream of the monitoring site. The natural land cover class is an aggregate category, and includes sites in predominantly indigenous forest, tussock grassland, bare mountain and scrub land cover classes. Aggregate summaries are also provided for all pastoral and natural sites (ie, regardless of source of flow).

The first step in producing data summary tables was to compute annual median values at each site for NO_x-N, NH₄-N, clarity DRP, and 95th percentiles for E. coli. Summary tables were then compiled as land cover type by source of flow by year, with values from each site (ie, either annual median or 95th percentiles) used as the replicate data to calculate 5th, 50th (median) and 95th percentile values for each class. We follow Scarsbrook (2006) and interpret these percentiles as representing the best, average and worst river conditions, respectively. For example the 5th percentile values for NO_x-N, NH₄-N, DRP and E. coli represent the sites with lowest concentrations, whereas the 95th percentile represents highly enriched, and/or contaminated sites.

The percentage of sites that failed to meet the appropriate guideline values for each variable was also calculated.

[ Previous|Next