1. Introduction
1.1. Objectives of the study
River water quality is a major component of national sustainable development plans and regional water management in New Zealand (e.g., ORC 2000, MfE 2003). Effective management and sustainable development both require assessments of water quality state and trends. Water quality assessments in New Zealand are regularly made for individual streams (e.g., Wilcock et al. 1999), multiple streams within catchments (e.g., Quinn and Stroud 2002), multiple catchments within regions (e.g., WRC 2001, Meredith and Hayward 2002), and selected large rivers across the country (e.g., Smith et al. 1996). Each of these cases involves a trade-off between generality and resolution. The site-specific and fine-scaled studies cannot be used to make generalizations about larger areas, while the large-scaled studies have insufficient resolution for identifying relationships between water quality and fine-scaled environmental factors. These tradeoffs could be reduced if assessments included a large number of monitoring sites distributed across the nation, and these sites represented a wide range of river sizes and environmental conditions. Assessments of this scope have several advantages:
(1) catchment land-uses associated with impaired water quality state or negative trends can be identified after removing the confounding effects of larger-scaled variation;
(2) comparisons of water quality can be made among regions, and between regions and the national average, and
(3) the effects of intrinsic properties such as river size on water quality can be distinguished from the effects of extrinsic factors such as catchment topography and land use.
To date, no such nation-wide assessment has been conducted, primarily due to the lack of a suitable dataset.
The general objectives of this study were to acquire an extensive, up-to-date water quality dataset, and assess patterns in the recent state and trends in water quality in rivers across New Zealand. Assessments of water quality across large areas require a spatial framework to partition variability, because monitoring sites are likely to encompass a wide range of environments. Jurisdictional boundaries do not provide a useful spatial framework, as they are rarely congruent with the boundaries of environmental units such as climate zones and geological formations. Instead, we used the GIS-based River Environment Classification (REC) to organise the rivers in the New Zealand dataset according to the environmental factors that strongly influence water quality (see Overview of the REC).
This report expands on an earlier study of water quality state and trends in New Zealand that used a smaller dataset (Snelder and Scarsbrook 2002). In the earlier study, water quality in contrasting REC classes was compared using maps and graphs, but low site replication within classes prevented the authors from making statistical comparisons. While it could not be categorically stated that some REC classes had better or worse water quality than others, the earlier study revealed several patterns: rivers in dry climates and at low elevations had higher dissolved nutrient and bacteria concentrations that those in wet climates and hill or mountain elevation classes; pastoral catchments had lower water clarity and higher dissolved nitrogen and bacteria concentrations than undeveloped or "baseline" catchments. In the present study, we extend the analyses of the earlier study using a larger dataset. The new dataset allowed us to examine water quality patterns in more REC classes, and make statistical comparisons among and between classes. These advances increase our confidence that the spatial patterns delineated by the REC framework accurately represent geographic patterns in water quality in New Zealand. More generally, both the current and previous studies used the REC framework to organise monitoring sites into REC classes, then extrapolate water quality state and trends from the site-specific data to the classes. This approach provides information about geographic patterns in water quality, the geographic extent of water quality problems, and helps to identify the factors that may cause or contribute to those problems.
There were four specific objectives for the study.
- Identify differences in water quality state among REC classes, and compare state to guidelines recommended for the protection of river ecosystems.
- Identify temporal trends in water quality and compare trends among classes.
- Compare water quality state and trends among regions, and between regions and the national average, to determine whether and how water quality within REC classes differs among regions.
- Determine whether stream size is a significant source of variability in water quality within REC classes.
In addition to analyses of state and trends, several recommendations are made concerning water quality monitoring programmes in New Zealand rivers. In particular, we make recommendations that may improve the representativeness of sampling sites, and the consistency and precision of field and laboratory methods.
1.2. Overview of the REC
The REC characterises river environments at six hierarchical levels, each corresponding to a controlling environmental factor. The factors, in order from largest spatial scale to smallest, are climate, source-of-flow, geology, land cover, network position and valley landform (Table 1). In this study, valley landform was not considered, as fine-scaled subdivision results in low replication in most REC classes. Network position was also not considered. Instead, we used stream order to determine effects of stream size on water quality patters. Each factor is associated with a suite of physical processes that influence water quality, and vary at approximately the same scale. For example, the climate level of the REC is associated with precipitation and thermal regimes that vary at scales of 103 - 104 km2. Each REC factor is composed of 4 - 8 categories that differentiate all New Zealand rivers. Categories at each classification level and their abbreviations are shown in Table 1. The number of possible classes at any level is equal to the number of categories at that level multiplied by the number of classes at the preceding level. For example, the source-of-flow level has 24 possible classes (6 climate classes X 4 source-of-flow classes). At the geology level there are 144 possible classes, and 1152 at the land-cover level. Not all classes are represented in the datasets used for this study, and in fact, some do not occur in New Zealand. In general, however, classifying a given group of rivers to increasingly fine-scaled levels results in an increase in spatial resolution, and a decrease in the number of rivers per class. Figure 1 shows the spatial distribution of REC classes in New Zealand, at the source-of-flow level.
The REC class of a particular river segment is given by stringing together the categories in each factor that correspond to the segment. The general form is: climate/source-of-flow/geology/land-cover/network position/valley landform. As an example, river segments classified to the land-cover level may belong to the Cool Dry/Hill/Hard Sedimentary/Pastoral class, denoted CD/H/HS/P.
