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

Spatial patterns in state and trends of environmental quality in New Zealand rivers: an analysis for State of Environment reporting

Executive Summary

1. Introduction

2. Methods

3. Results of analysis of physico-chemical and biological state

4. Results of analysis of trends in water quality and biological variables

5. Discussion

6. Conclusions

7. Acknowledgements

8. References

Appendix 1

Appendix 2: State analysis results summary table

3. Results of analysis of physico-chemical and biological state

3.1. Analysis of national water quality

We used the NRWQN data to characterise the state of water quality nation-wide (i.e. by aggregating all sites) and at two finer scales by using the first two levels of the REC (Climate and Source of Flow). The aggregated NRWQN site data is shown as box and whisker plots, which show medians and the variation among sites in each class. The central small square is the median value. The upper and lower boundaries of the larger boxes are the 25th and 75th percentile of the statistics, and the ends of the whiskers are the minimum and maximum values. Where a single site represents a class the data is shown as a single small box. The red line is the relevant guideline. The median value for each class has been compared to the values recommended in Table 5 to determine if the class 'exceeds' or 'within' each guideline. Summary tables of the data, which also indicate whether each class exceeds or is within the relevant guideline, are provided in Appendix 2.

Water clarity in each class at the Climate and Source of Flow levels are shown in Figure 7. Clarity increases as catchment rainfall increases. The Cool Extremely Wet class (CX) has the highest median clarity and is the only climate class that does not exceed the guideline (i.e. median clarity is greater than 1.6 m).

Classification at the Source of Flow level indicates that, generally, rivers in the Low Elevation classes had the lowest clarity. This result reflects the dominant land cover and land use in Low Elevation catchments, which would have the greatest input of organic material and sediment from both natural processes and agricultural activities. The greatest range in clarity was in rivers in the CW/Lk class. This class includes rivers that have a broad range of catchment geology and land cover. For example the Waikato River is classified CW/Lk and its catchment is dominated by volcanic geology and pastoral land cover. In contrast to the Waikato, the South Island CW/Lk river sites include the Kawarau, Clutha and Monwai rivers, whose catchments are dominated by hard sedimentary geology and indigenous forest or tussock land cover. These South Island CW/Lk rivers, therefore, belong to different lower level REC classes. Further subdivision of the sites at the REC Land Cover level would produce a reduced the range in clarity to that shown for the CW/Lk class in Figure 7. The WW/Lk class had the lowest clarity. This class is represented by a single site, the Tarawera at Awakaponga. This site is heavily impacted by an upstream industrial discharge and is not representative of other sites in this class.

Inorganic nutrient concentrations (SRP and SIN) in each class at the Climate and Source of Flow level are shown graphically in Figure 8 and results are tabulated in Appendix 2. Nutrient concentrations generally decreased as catchment rainfall increases (median concentrations increased from Extremely Wet (X) to Dry (D) climate classes on Figure 8). The Cool Extremely Wet rivers (CX) had the lowest median concentrations of SRP and SIN and was the only climate class for which median SRP and SIN were within the guideline. Rivers in the Cool Wet and Cool Dry climate classes greatly exceeded the SIN guideline but were close to the SRP guideline. Rivers in these classes may be phosphorus-limited, which means that additional phosphorus could produce an undesirable eutrophic (high algal biomass) state.

In general, rivers in the Mountain and Lake Source of Flow classes had the lowest nutrient concentrations and rivers in the Low Elevation classes had the highest nutrient concentrations, followed by Hill then Mountain source of flow classes. This result reflects a gradient in land use intensity that coincides with the three Source of Flow classes.

With the exception of rivers in the CX climate class, Low Elevation Source of Flow rivers exceeded both the SRP and SIN guidelines. Given adequate light, substrate and time for algal growth, biomass could exceed acceptable levels in rivers in these classes. Cool Wet Hill Source of Flow rivers were within SRP guidelines but exceeded the SIN guidelines. Rivers in the Hill Source of Flow class exhibited relatively wide variation in SRP concentrations (see Figure 8). This reflects significant heterogeneity in the smaller scale factors (Geology and Land Cover) that produce differences in nutrient concentrations. For example, previous studies using REC (Snelder et al. 2000) showed that rivers with volcanic geology are generally higher in phosphorous than rivers with hard sedimentary geology. The variation in water quality variables within the Source of Flow class can be further reduced by subdivision at the Geology and Land Cover levels of the REC, as discussed in Section 3.2.

BOD₅ at the Climate and Source of Flow levels are shown on Figure 9 and are tabulated in Appendix 2. The BOD₅ guideline was rarely exceeded. Only the WW/Lk class exceeded the guideline, but the site is probably not representative due to the industrial discharge at the single monitoring site in this class. There were also sites within the Cool Wet and Warm Wet Low Elevation classes (CW/L and WW/L) that failed, suggesting that organic mater inputs were high into some Low Elevation Source of Flow rivers.

Total ammonia in each class at the Climate and Source of Flow level is shown on Figure 10 and tabulated in Appendix 2. The analysis indicates that Total Ammonia concentrations that exceed the guideline are very rare. Unionised ammonia is acutely toxic but is also rapidly converted to other nitrogen forms in the natural environment. Given this, anything other than rare exceedance of the guideline by monthly sampling programs would be a cause for concern.

3.2. Water quality at the Regional scale

The increased replication achieved by incorporating the ESWQN data allowed us to 'zoom' the state analysis to smaller spatial scales in Southland than those used for the national analysis. The small scale classes were REC Land Cover level classes (see Table 4), which include both catchment geology and land cover as factors. We have analysed and present results for the state of each class for five water quality variables (clarity, SRP, SIN, Total Ammonia, and E. coli) using the same methodology as the national scale analysis.

Water clarity results are shown Figures 11, 12 and 13 and tabulated in Appendix 2. Figures 11 and 12 show the ESWQN data, classified at the Climate and Source of Flow levels, superimposed on the NRWQN data in order to compare Southland's rivers with similar rivers throughout New Zealand. Figure 13 shows the ESWQN data only, but classifies the sites at the Land Cover level to show variation in river water quality within smaller groups of rivers in the region. The level of replication of the national and regional analyses is low. For example, a single site represents the CD/H class in Southland. Differences in the range in water quality variables between ESWQN and NRWQN sites may, therefore, not be an accurate representation of differences between national and Southland rivers.

Based on the limited dataset, Figures 11 and 12 indicate that clarity in Southland's rivers was similar to rivers in the equivalent REC classes elsewhere in New Zealand. Figure 12 shows that the ESWQN has sites in the CW/H and CW/Lk classes that had greater clarity than sites in the NRWQN and the CD/L class had sites with poorer clarity than sites in the NRWQN. Wet and Extremely Wet Climate classes in Southland (i.e. CX and CW) have generally higher clarity than the Dry class (CD). This is consistent with the large-scale patterns in water clarity that were shown by the NRWQN analysis. Including the Land Cover classes in the regional analysis reduced the variability within the Source of Flow classes. The effect of land use is illustrated by comparing the median clarity for CW/H/HS/B and CW/H/HS/P classes. The baseline Land Cover category (B) has greater clarity than the sites whose catchment is dominated by a Pastoral (P) land cover. Some sites in the CW/H/HS/P class are close to the clarity guideline, whereas all the CW/H/HS/B sites are well above the guideline.

All the Low Elevation classes are below the clarity guideline. The effect of catchment geology is demonstrated by the difference between clarity data for Cool Dry Low Elevation Hard Sedimentary and Soft Sedimentary classes (CD/L/HS/P and CD/L/SS/P). There is little overlap in clarity between these two classes despite similar Land Cover category. This result suggests there is a need to adjust expectations for water clarity according to natural factors such as geology.

Figure 14 and 15 indicate that nutrient concentrations Southland's rivers are generally similar to those in equivalent REC classes nation-wide. The exception is the CD/L class where median SRP and SIN concentrations for the ESWQN sites are higher than for the NRWQN data and the range is greater Note, however, that there are more CD/L sites for the ESWQN (10) than for the NRWQN (3) (see tables 6 and 7).

The analysis of nutrient concentration by REC Land Cover classes, shown in Figure 16, reduces the variability in nutrient concentrations evident at higher-level REC classes. This result confirms that differences in land cover causes predictable differences in nutrient concentrations. For example, the national analysis (Figure 8) indicates that rivers in the Cool Wet Hill Source of Flow class (CW/H),generally exceeded the SIN guideline, but 25% of sites are below the guideline. This result is confirmed by the regional analysis. All sites in the Cool Wet/Hill/Hard Sedimentary/Baseline class (CW/H/HS/B) are within the guideline and all in the Pastoral Land Cover class (CW/H/HS/P) exceed the SIN guideline. SRP concentrations are similar in both Pastoral and Baseline classes and while both are within the guideline, these concentrations for Southland sites are above the 75th percentile of mean concentrations for CW/H shown by the national analysis.

Total ammonia concentrations in rivers from each REC Land Cover class are shown in Figure 17 and tabulated in Appendix 2. The data presented in Figure 17 indicate that Total Ammonia concentrations are higher in the rivers in the ESWQN than those in the NRWQN (see Figure 10). Note that one Land Cover class CD/L/SS/P has some sites where the 99th percentile of Total Ammonia concentrations exceeds the guideline, whereas none of the NRWQN sites exceeded the guideline. The results suggest that ammonia toxicity is not a problem at larger scales but ammonia toxicity may occur at small spatial scales.

Figure 18 and 19 shows concentrations of E. coli that have been calculated from the mean of monthly samples for the year 2000/2001. The E. coli results show a similar pattern to the other water quality variables with the lowest concentrations in rivers with wet and mountain dominated catchments (i.e. the CX and CW/M) classes, and highest concentrations in the dry, low elevation catchments (CD/L). These patterns reflect differences in the sources of feacal contamination, which are lower in mountainous areas than in low elevation catchments where pastoral land use dominates as well as gradients in dilution capacity from high to low.

The analysis of the E. coli concentrations was zoomed to the REC Land Cover level. In order to increase replication within the REC classes we simplified classes at the Land Cover level into the following classes (number of replicate sites shown in parenthesises; CD/L/HS/P(6), CD/L/SS/P (4), CW/H/HS/B (3), CW/H/HS/P (6), CW/L/HS/P (2), CW/L/SS/P (5), CX/Lk/Pl/B (2). Figure 20 indicates that feacal contamination is highest in rivers whose catchments are dominated by pastoral land cover and lowest in the simplified 'baseline' land cover categories with indigenous forest, tussock land cover.

3.3. Biological variables

The analysis of the biological state of rivers is based on a reduced set of sites because biological data is unavailable for some of the NRWQN sites. The NRWQN periphyton data for filamentous and mat cover data is shown as box and whisker plots in Figures 21 and 22. The analysis of periphyton cover at the REC Climate level indicates that there is a general tendency for cover to decrease as catchment rainfall increases and temperature decreases, which is shown by the classification. There is a weak relationship with REC Source of Flow level were periphyton cover is generally lower in Mountain and Hill rivers than rivers in the Low Elevation classes. The Warm Wet and Cool Dry Low Elevation and Cool Dry Hill classes have the poorest state with the median exceeding the guideline for filamentous algal cover. The likely reason for the large variability for each class, and the general weakness of the relationships between REC classes and periphyton cover is the importance of antecedent conditions in determining periphyton biomass. The time since the last flood is a strong determinant of periphyton biomass on any sampling occasion.

MCI scores were calculated for each site and year. The resulting score were aggregated by REC class and are shown in Figures 23. There is a general tendency for scores to decrease as catchment rainfall reduces and temperature decreases (Figure 23). At the Source of Flow level, MCI scores were generally higher in Mountain and Hill rivers than in Low Elevation rivers. The Warm Wet Low Elevation rivers had the poorest MCI scores; the median was below 80, indicating severe pollution. The relationship between climate and MCI scores raises the possibility that the MCI score interpretation may need to be adjusted by REC classes in order to account for natural reduction in some taxa with decreased flows or increasing temperature.

Higher replication could be introduced by incorporating regional council datasets. This would allow us to make comparisons at the Geology and Land Cover levels of the REC. These REC levels have been shown to account for significant variation in invertebrate community, including MCI scores (Suren et al. 2000).

3.4. Summarising the relevant information for reporting purpose

'Sun ray' diagrams were used to summarise the water quality data from the NRWQN monitoring sites. Data from the 70 sites was aggregated by REC classes at the Climate and Source of Flow levels. We used the median of clarity, SIN, SRP, BOD₅ to represent water quality in each class. We used the 99th percentile of Total Ammonia because the guideline we use is based on acute toxicity. We removed one of the 70 main stem NRWQN sites used in the previous analyses (site RO2 Tarawera at Awakaponga) from this analysis, because it was impacted by a point source industrial discharge.

The diagrams (Figures 24 and 25) were constructed by calculating the mean value of the five water quality variables for all monthly samples in the period 1995 to 2001 for each sites. The means for each site are then aggregated by class and the median of these variables is then plotted on one of 5 'rays'. We used the reciprocal of clarity so that the relative plotting position is the same as for the other variable (i.e. the number increases as water quality decreases). Blue lines linking the plotted positions form a polygon. The area of the blue polygon allows us to compare the relative water quality among classes.

The recommended guidelines (see Table 5) for each water quality variable have also been plotted on each ray. Joining these points provides an 'acceptable water quality envelope' (the orange polygon). When a class complies with all guideline, the actual mean state (blue) polygon lies entirely within the red polygon. Exceedence of guidelines is seen as the actual mean state protruding from the acceptable water quality envelope. The information shown on the sun ray diagrams has also been interpolated to all rivers that share a similar class (Figure 26) The map indicates the extent of rivers that comply or exceed guidelines.

Sixty nine main stem sites from the NRWQN have been classified at the REC Climate Level (5 classes).