Report: TR 2004/02
Author: Bill Vant, Paul Smith
We analysed trends in river water quality at 110 sites in the Waikato region using non-parametric statistical methods (seasonal Kendall slope estimator and trend test). At 10 Waikato River sites we analysed records of 19 water quality variables that began in 1987 or later and ended in 2002. At the 100 other river sites records of 14 variables beginning in 1990 or later and ending in 2002 were analysed. The data were generally obtained at monthly intervals, but some records were based on quarterly sampling. Most of the records were adjusted to remove the effects of flow, and both raw and flow-adjusted records were analysed for trends.
A total of 188 Waikato River water quality records were considered. Significant trends (p < 5 percent) were found in 92 (49 per cent) of these. Variables for which significant trends were found at five or more of the ten Waikato River sites were pH, dissolved colour, biochemical oxygen demand, arsenic, boron, ammonia and faecal coliforms. Apart from faecal coliforms, all these showed decreasing trends. The decreases in dissolved colour, biochemical oxygen demand, arsenic, boron and ammonia all represent improvements in water quality, and mostly result from improved wastewater management at known point source discharges (e.g. Kinleith mill, Wairakei power station). The decrease in pH, however, represents a deterioration. The cause of this decrease is not clear. Some of the trends in faecal coliform levels resulted from a probably harmless, non-sewage discharge to the lower river that went un-noticed until 2002, but has subsequently ceased. A more reliable measure of faecal contamination of the lower river is probably provided by the 1988-2002 records of enterococci bacteria. None of the enterococci records at the five lower river sites have shown increasing trends.
A total of 1334 water quality records from the other rivers and streams were considered. Significant trends were found in 589 (44%) of these. Across the region as a whole, the following overall patterns were apparent: (1) significant increases have occurred in conductivity, visual clarity, total nitrogen and total phosphorus; and (2) significant decreases have occurred in dissolved oxygen, pH, turbidity, dissolved colour and ammonia. While some of these overall trends were improvements (increases in clarity and decreases in turbidity, decreases in ammonia), we consider that many of the other trends were deteriorations. The magnitudes of the trends in conductivity, visual clarity, dissolved colour, total nitrogen, ammonia, total phosphorus and dissolved reactive phosphorus were significantly correlated with the proportion of the catchment area that was in pasture (i.e. the trend slopes were correlated with land use).
For a small number of the trends at the non-Waikato River sites we can identify probable causes. However, the processes that are likely to have resulted in many of the other trends are less obvious. Some of these changes have also been observed in rivers throughout New Zealand (e.g. decreases in pH, increases in conductivity), although often at slower rates than those observed in the Waikato Region. It is therefore possible that the processes responsible are operating at a national rather than a regional scale. We consider that a concerted effort needs to be made to investigate this.
Trends in river water quality in the Waikato region, 1987 - 2002
(434 kb, 62 seconds to download, 56k modem)
|2.1 Datasets analysed|
|2.2 Statistical analyses|
|2.3 Seasonal Kendall trend slope|
|2.4 Seasonal Kendall trend test|
|2.5 Flow adjustment|
|2.6 Flow records|
|2.7 Effect of land use|
|3 Results and Discussion||9|
|3.1 Waikato River||9|
|3.2 Other rivers and streams|