4. Discussion

4.1 Water quality state

The assessment of water quality state (period 2003–2007) shows that water quality was highly variable throughout New Zealand. Median nutrient concentrations frequently exceeded the ANZECC (2000) trigger valuesand were lower than the clarity guidelines (Table 3 and Table 4). Faecal bacterial levels were also high, with E. coli numbers exceeding the MfE/MoH (2003) action value at many sites throughout the country (based on the 95thpercentiles).

Land-use impacts on water quality state were clear with the highest nutrient concentrations being associated with pastoral land cover. Nutrient concentrations were also high in urban rivers. Clarity was low in rivers whose catchments drain pastoral and urban land compared to other land-cover types. This is consistent with data reported by other authors in previous studies, (e.g., Ballantine and Davies-Colley, 2009b; Hamill and McBride, 2003; Larned et al., 2003; Larned et al., 2004; Snelder and Scarsbrook, 2002).

There are clear differences in water quality depending on where rivers have their source as shown by grouping sites by REC Source-of-flow category. NOx-N and TN concentrations were significantly higher,and clarity significantly lower,in rivers in the Low Elevation Source-of-flow category, compared to rivers in the Hill, Lake or Mountain Source-of-flow categories.

4.2 Water quality trends

The trend analyses indicate that trend strength and direction is highly variable across the country. There were also considerable differences in trend strength and direction between the time periods. We used the binomial test to indicate whether there were “overall trends” in sites grouped in several ways. We deemed that there was an overall trend in a certain direction for a grouping if the number of sites that exhibited that trend were greater than could be expected if increasing and decreasing trends were equally likely. In this manner we found overall decreasing trends in clarity and increasing trends in conductivity, TN and TPat the national scale for the 1998 to 2007 period, all of which indicate degrading water quality. We note that the number of stable trends was high relative to the total number of sites for several analytes, particularly DRP, NH4-N and TP. This probably reduces the certainty with which we can conclude therewereoverall trends for these analytes.

When sites were grouped by region for the 1998–2007 period we found the following overall trends, which all indicate deteriorating water quality:

  • decreasing overall trends in clarity in the Waikato, Wellington, Hawke’s Bay and Manawatu-Wanganui regions
  • increasing overall trends in conductivity in the Canterbury, Southland, Northland and Waikato regions
  • overall increasing trends in oxidised nitrogen in the Canterbury and Waikato regions
  • overall increasing trend in TN in the Waikato region
  • overall increasing trends in both dissolved reactive phosphorus and TP in the Hawke’s Bay and Otago regions.

However, we also found overall trends which are improvements in water quality. These trends in improving water quality make it difficult to conclude that there are strong regional patterns in water quality degradation. The improving overall trends include:

  • decreasing trends in conductivity in Gisborne and Wellington regions
  • decreasing trends in oxidised nitrogen in the Auckland, Wellington and Northland regions
  • decreasing trends in both TN and dissolved reactive phosphorus in the Southland and Northland regions
  • decreasing trends in ammoniacal nitrogen in Auckland, Canterbury and Northland regions
  • decreasing overall trends in bacterial indicators (faecal coliforms and/or Escherichia coli) in Southland, Otago and Hawke’s Bay

The strongest groupings in terms of identifying overall trends for the 1998–2007 period were the REC Source-of-flow and Land-cover categories. We found overall:

  • decreasing trends in clarity in Hill and Low-Elevation Source-of-flow categories and Pasture and Urban Land-cover categories
  • increasing trends in TP in the Low-Elevation Source-of-flow category and Pasture Land-cover category
  • increasing trends in conductivity, oxidised nitrogen and TN in the Pasture Land-cover category.

These results suggest that water quality decreased over the 1998 to 2007 period in low elevation areas and in catchments dominated by pastoral land cover. Over the same period however, NH4N showed decreasing trends in the same categories and in the Lake Source-of-flow and Indigenous Forest Land-cover categories.

Comparison of the time periods for the NRWQN sites shows that trends tended to be stronger for the five year time period than the ten year period, with more meaningful trends (that is, greater rates of change) observed for the shorter time period at individual sampling sites than for the longer time period. For example, in the Waikato region, trends in TN and NOx-N were stronger for the 2003–2007 period than for the 1998–2007 period. TN and NOx-N trends were significant and increasing for the longer time period, but for the shorter time period, they were mostly meaningful and increasing. Also, for the longer period, the declining trends in visual clarity were mainly significant; but for 2003–2007 the declining trends in visual clarity were mainly meaningful (i.e. were stronger trends). Results from this study are mostly in agreement with an earlier study on Waikato River sites for the 1988–2007 time period by Vant (2008). Over the time period reported by Vant (2008), significant increases were observed in conductivity, TN, NOx-N, TP and E. coli. Results from the present study also show increasing trends in TN and NOx-N for both time periods. The present study found significant decreases in visual clarity for the Waikato region, which were not observed over the1988–2007 period by Vant (2008).

Overall national trends observed in nutrients for the NRWQN over both time periods contrast with those reported for the 19-year time period by Ballantine and Davies-Colley (2009). In the 19-year analysis, the median RSKSE values indicated increasing trends in nutrients and visual clarity (Ballantine and Davies-Colley, 2009b). However, in this study we found no overall trend in clarity in either time period. The present study did find an overall increasing trend in TN for both time periods in agreement with that reported for the 19-year period. However, the present study found an overall decreasing trend in DRP for the ten year time period which contrasts with the earlier study.