Skip to main content.

Water-quality Baseline

The following section presents a summary of baseline water quality in the Tier 1 and Tier 2 dairy catchments. Data for each of the water quality parameters listed in table 2 is presented in graphical format. However, a summary of the statistics underpinning the graphs, and details on sampling regimes and data set size, are provided in tables A2-B to A2-N in appendix 2.

The same tables also provide the time periods of monitoring for each catchment, and water quality parameter, from which medians have been derived. Time periods varied between catchments and parameters but were generally 2001/02–2006 for the Tier 1 catchments and two Tier 2 catchments (Enaki and Washpool), and 2006–2007 for the remaining Tier 2 catchments (with the exception of the Taharua: 2000–2008).

The section finishes with a brief account of water-quality changes (trends) that have been observed in the monitored catchments (where information is available to do this).

Presentation of data

Results for different water-quality and stream health variables are not presented in any particular order of importance. For freshwater managers, the relative importance of nutrients versus physical and chemical stressors versus measures of ecosystem health will vary according to local management priorities.

Unless otherwise stated the statistics presented in each figure and referred to in the text are for the most downstream monitoring site in the catchment.

The order of catchments along the horizontal axes of the graphs is consistent throughout this section and is based on the amount of catchment area under dairy farming: dairy area as a percentage of total catchment area decreases from left to right in each graph.

It is important to note that although catchment results are presented together, drawing conclusions about the overall condition of catchments relative to one another should be done with caution for two reasons: firstly, because of the abovementioned variability in time periods of monitoring between catchments, and secondly, because the actual consequences of disturbances to water and habitat quality are catchment specific. The primary value of water-quality indicators is to track change at one site (or catchment) over time, rather than to make comparisons with other sites, although the latter can be informative.

Many of the graphs in this section have logarithmic scales on the vertical axes to enable large spreads of data from different sites to be presented together.

The Australia and New Zealand Environment and Conservation Council guidelines for fresh water quality (ANZECC, 2000) have been used in several cases to benchmark findings. It is important to note that the ANZECC guideline values are default ‘trigger values’ intended primarily to assess the risk of adverse effects to aquatic ecosystems, rather than thresholds to report against. They should therefore be interpreted in this report as indicative, rather than absolute, thresholds. More information about specific guidelines is included in text for each water-quality variable.

Faecal contaminants

Faecal contamination of waterways poses a public health risk. Illness may be contracted as a direct result of ingesting bacterial, viral and protozoal pathogens that occur in faecal material. Faecal material reaches streams in run-off from the land through effluent pond discharges (eg, Smith et al, 1993) and from cows defecating directly into the water (eg, Davies-Colley et al, 2004).

Risk of illness is primarily associated with recreational activities where water may be ingested (including harvesting fish and other aquatic food for consumption). The indicator commonly used to assess this risk is E. coli, a faecal coliform bacterium that originates in the gut of warm-blooded animals and indicates the presence of other potentially harmful microbes. There are several reference values and guidelines used for interpreting E. coli data (Table 5).

Table 5: Reference values and guidelines used for interpreting E. coli data in Figure 3

Values and guidelines used E. coli per 100 mL Description and source
Reference values   Based on assessments of regional council state-of-environment monitoring data (1996–2002) for different land-use categories. Data held by the Ministry for the Environment.
Natural catchments 50 Based on median value of 42 E. coli per 100 mL.
Data from 75 sites and 1,572 samples.
Pastoral catchments (both high- and low-intensity pastoral use) 200 Based on a median value of 200 E. coli per 100 mL.
Data from 259 sites and 6,330 samples.
Contact recreation   Microbiological Water Quality Guidelines for Marine and Freshwater Recreational Areas (Ministry for the Environment and Ministry of Health, 2003).
Single sample ‘Alert’ 260 Single sample thresholds indicating elevated health risk (Alert) and unsafe concentrations of pathogens (Action).
Single sample ‘Action’ 550
Stock drinking 100* ANZECC, 2000. However, ANZECC also recommends that “investigations of likely causes (of contamination) are warranted when 20% of results exceed four times the trigger value”.

* Faecal coliforms per 100mL (not E. coli).

Figure 3 shows the median and range of E. coli measurements in the monitored dairy catchments in relation to several reference values that are summarised in table 5. The dashed lines are based on median values for catchments in New Zealand with predominantly natural and pastoral land uses and provide an indication of the relative extent to which the monitored dairy catchments are polluted with faecal matter. The solid lines are ‘single sample’ contact recreation thresholds and provide a further indication of the extent to which water quality in the catchments could be considered hazardous to humans.

Figure 3: E. coli medians and ranges for the most downstream site for each catchment during the monitoring period

View graph|View text description
Figure 3: E. coli medians and ranges for the most downstream site for each catchment during the monitoring period

Notes:

Dashed lines are reference medians for pastoral (upper) and natural (lower) catchments.

Data, details on monitoring periods, guidelines, reference values and sample numbers are provided in appendix 2, table A2-B.

The vertical axis has a logarithmic scale.

T1 = Tier 1 catchment and T2 = Tier 2 catchment.

The majority (10) of the monitored catchments had median E. coli concentrations that exceeded the overall median for pastoral catchments in New Zealand (200 E. coli per 100 mL). This indicates that these monitored catchments are more heavily affected by faecal contamination than the ‘average’ pastoral farming catchment. Three catchments (the Rhodes, Rai and Enaki) had median E. coli concentrations just below 200 per 100 mL, indicating faecal pollution exists but could be considered moderate in the context of all pastoral catchments. One catchment, the Taharua, had a median E. coli concentration of well below 50 per 100 mL, which compares with values found for unmodified (natural) catchments.

All catchments except the Taharua had peak (sampled) bacteria counts of well over 1000 per 100 mL (and seven catchments had peaks of well over 10,000 per 100 mL) suggesting microbiological pollution is a substantial, if occasional, problem in all cases. However, such peaks are not uncommon in other (non-dairy) farmed and urban catchments and generally correspond with high rainfall run-off (eg, Greater Wellington Regional Council, 2006).

Dry-weather peaks in E. coli are arguably of more concern in a dairy environment because they may relate to poor farm practice such as overflowing effluent ponds or stock defecating directly into the stream. It would be useful in future reports to include a stream flow-related analysis of E. coli concentrations to determine whether there is any change in the magnitude and frequency of dry-weather peaks.

Human health risk – contact recreation

Some of the catchments are recreational areas in their own right, such as the Rai River, where swimming and kayaking are common, while many of the other catchment streams drain into larger water bodies where contact recreation or food harvesting takes place (see appendix 5, table A5-A for a description of recreational values in each catchment). Examples include swimming and whitebait fishing areas on the Motupipi River, into which Powell Creek drains; and swimming, trout fishing and eeling on the Pomahaka River, into which the Washpool Creek drains.6

Although only the Rai River is routinely monitored for recreational water-quality purposes, it is informative to consider the bacterial water quality in all of the dairy catchments in relation to recreational water-quality guidelines (table 5).

Three catchments – Waiokura, Inchbonnie and Washpool – had median E. coli concentrations in excess of the Action level of 550 per 100 mL. This means that more than half of the samples taken in each of these catchments had concentrations of E. coli that exceeded 550 per 100 mL. This is indicative of an unacceptable health risk, although it should be noted that activities such as swimming are not known (by council staff) to occur in the streams in these catchments (see appendix 5, table A5-A). A further six catchments had median E. coli concentrations that exceeded the Alert threshold level of 260 per 100 mL. This means that more than half of the samples taken in each of these catchments had concentrations of E. coli that would, if the sites were recognised and monitored for contact recreation, trigger follow-up sampling and investigations.

The remaining five catchments had median E. coli concentrations below 260 per 100 mL, meaning that on more than half of the occasions these sites were sampled, water quality was acceptable for contact recreation.

Stock drinking-water quality

The microbiological quality of stream water is also important in farmed catchments for stock health reasons, because water is sometimes drawn untreated from this source for a stock drinking supply. However, there appears to be very little, if any, reliable data on acceptable numerical limits for microbes (Sinton and Weaver, 2008; Ministry of Agriculture and Forestry, 2004).

In the absence of an adequate pool of scientific data, the ANZECC (2000) guideline ‘trigger value’ for stock drinking water quality is often referred to in New Zealand and has been used in this report as a reference point. ANZECC recommends a trigger value of 100 faecal coliforms per 100 mL, but also recommends taking action to remediate a stock drinking supply once 20 per cent of samples taken over an extended period exceed four times this value (ie, 400 per 100 mL). Although 20th percentile data was not available for this report, five of the monitored farm catchments had median E. coli7 concentrations exceeding 400 per 100 mL (ie, at least 50 per cent of samples exceeded 400 per 100 mL).

Nutrients (nitrogen and phosphorus)

Nitrogen and phosphorus are essential nutrients for the growth of aquatic plants and algae that form an important part of any healthy stream ecosystem. However, excessive in-stream concentrations can lead to proliferations of algae (eg, rock slime) and macrophtyes (aquatic plants), which in turn may compromise a range of in-stream values such as amenity, native fish conservation and recreation (Biggs, 2000).

Measuring nitrogen and phosphorus concentrations therefore provides an important indication of the potential for proliferations to occur in the monitored catchments (all other factors such as light, stream bed and temperature conditions being equal). These measurements also provide an indication of the contribution of nutrients from the monitored catchments to downstream receiving waters such as larger rivers, lakes and estuaries.

The main routes for nitrogen loss from pasture to streams are (1) leaching through soils from livestock urine patches and applied fertiliser, (2) direct input of livestock excreta (dung and urine) to streams and excreta run-off from paddocks, and (3) soil erosion (eg, McKergow et al, 2007). Phosphorus is generally lost to pastoral waterways from paddock run-off of eroded soil and fertiliser as well as effluent pond discharges (either directly or through a land application) (eg, McDowell et al, 2008).

Box 3: Nutrient guidelines

The determination of guidelines relating to nutrient concentrations in rivers and streams is a scientifically complex issue. The concentrations at which nitrogen or phosphorus actually begin to have an adverse effect on ecosystem health or amenity values is highly site- and catchment-specific and depend on many factors. For example, a stream with relatively fast, variable flow that discharges to an open coastline may be able to support high nutrient concentrations that do not have an observable impact (eg, nuisance growths). However, if that stream discharges to a lake or an estuary where nutrients are likely to accumulate and boost plant growth, then in-stream concentrations become much more important to control. Likewise, a stream with primarily sandy substrate may be more resistant to nuisance blooms than a rock- or cobble-bottomed stream (given similar concentrations of nutrients).

In New Zealand, two national guidelines are commonly used to assess nutrient concentrations, and they have been used in this report.

1. Periphyton guidelines (Biggs, 2000). These guidelines provide suggested thresholds for the dissolved nitrogen and phosphorus concentrations required to control periphyton growth. A range of thresholds are provided that are related to flow conditions (high flow events tend to scour out periphyton growth). For this report the upper guideline values suggested by Biggs and relating to “20 mean days of accrual” have been adopted: 0.295 mg/L soluble inorganic nitrogen and 0.026 mg/L dissolved reactive phosphorus.

2. ANZECC guidelines (ANZECC, 2000). These guidelines provide default trigger values for total and dissolved nitrogen and phosphorus for assessing the risk of adverse effects in slightly disturbed ecosystems. These trigger values are based on the 80th percentile of a distribution of reference data and have the following values for lowland rivers: 0.614 mg/L for total nitrogen; 0.444 mg/L for oxides of nitrogen (nitrate–nitrite nitrogen); 0.033 mg/L for total phosphorus; and 0.01 mg/L for dissolved reactive phosphorus.

Arguably, both the periphyton and ANZECC guidelines could be considered environmentally conservative for highly modified dairy catchments. However, in the absence of more appropriate reference data for these modified systems, both guidelines still provide useful contexts for the data in this report.

The medians and ranges of nitrogen and phosphorus measurements in the monitored dairy catchments are presented in figure 4. The commentary below is focused on dissolved nitrogen (nitrate–nitrite nitrogen and soluble inorganic nitrogen) and phosphorus (dissolved reactive phosphorus) as these forms of each nutrient are most readily available for plant uptake in running stream waters. However, total concentrations are also important, particularly as they contribute to the overall nutrient loading in downstream water bodies.

Figure 4: Medians (crosses) and measured ranges (min, max) for total nitrogen (top), nitrate–nitrite nitrogen (middle), and soluble inorganic nitrogen (bottom) for the most downstream site during monitoring period

View graph|View text description
Figure 4: Medians (crosses) and measured ranges (min, max) for total nitrogen (top), nitrate–nitrite nitrogen (middle), and soluble inorganic nitrogen (bottom) for the most downstream site during monitoring period

Notes:

A log scale is used on the y axis.

Data, guideline values (dashed lines), sample number and monitoring period for individual catchments are provided in tables A2-C to A2-E in appendix 2.

Where bars go off the edge of the y axis scale the minimum or maximum is noted next to the bar.

T1 = Tier 1 catchment and T2 = Tier 2 catchment.

With few exceptions, total nitrogen and nitrate–nitrite nitrogen concentrations exceeded the ANZECC guidelines for protection of ecosystem health (see figure 4). Soluble inorganic nitrogen (nitrate–nitrite nitrogen + ammoniacal nitrogen) exceeded the periphyton guidelines in all catchments except for Puwera (see figure 4, bottom).8 (See box 3 for an explanation of guideline values.)

Nitrate–nitrite nitrogen concentrations in the Rhodes and Petrie catchments are particularly high (between two to four times the concentrations found in catchments with the next highest values). This probably reflects the relative importance of groundwater as a source of stream recharge in these Canterbury Plains catchments: groundwater samples taken upgradient of the Rhodes and Petries catchments are also high in nitrate–nitrite nitrogen and conductivity (Environment Canterbury, 2007).

Total phosphorus and dissolved reactive phosphorous concentrations were generally lower relative to the guidelines than nitrogen, as is typical in New Zealand freshwaters (see figure 5). However, most catchments (nine) exceeded the ANZECC guideline for total phosphorus and half (seven) exceeded the periphyton guideline for dissolved reactive phosphorus.8

Figure 5: Medians (crosses) and measured ranges (min, max) for total phosphorus (top), and dissolved reactive phosphorus (bottom) for the most downstream site during the monitoring period

View graph|View text description
Figure 5: Medians (crosses) and measured ranges (min, max) for total phosphorus (top), and dissolved reactive phosphorus (bottom) for the most downstream site during the monitoring period

Notes:

A log scale is used on the y axis.

Data, guideline values (dashed lines), sample number and monitoring period for individual catchments are provided in tables A2-F and A2-G in appendix 2.

T1 = Tier 1 catchment and T2 = Tier 2 catchment.

As is clear from figure 5, total and dissolved phosphorus concentrations are noticeably lower in most of the Tier 2 catchments compared with the Tier 1 catchments. Phosphorus commonly enters waterways bound to soil particles, and so the relatively high concentrations of this nutrient in the Tier 1 catchments may be an indication of the importance of soil loss and erosion as management issues in these catchments. RJ Wilcock (pers. comm., 2009) advises that the Tier 1 catchments generally have high Olsen phosphorus soil concentrations9 with high run-off potential to waterways.

The results of linear regression between nitrogen and phosphorus and three catchment variables –percentage catchment area under dairy (%Dairy), average annual rainfall and mean stream flow – do not reveal any particularly strong relationships (see table A3-A in appendix 3). However, it is worth noting the apparent influence both increasing %Dairy and decreasing rainfall have on increasing nitrogen concentrations (see figure 6).

Figure 6: Regression of %Dairy (left) and average annual rainfall (right) with median total nitrogen for all monitored catchments (n = 14)

View graph|View text description

Comparison with reference conditions and other pastoral catchments

Figures 4 and 5 show that concentrations of both nitrate–nitrite nitrogen and dissolved reactive phosphorus are well above the medians reported for reference (ie, relatively undisturbed) catchments in New Zealand by Scarsbrook (2008) (with the exception of the Powell and Rhodes catchments for dissolved reactive phosphorus). Furthermore, in all but the Inchbonnie and Puwera catchments, nitrate–nitrite nitrogen concentrations exceed the median for lowland pastoral catchments10 of 0.55 mg/L reported by Scarsbrook (see table A2-D in appendix 2). Dissolved reactive phosphorus concentrations for the dairy catchments are more comparable with the median of 0.016 mg/L for lowland pastoral catchments reported by Scarsbrook (see table A2-G in appendix 2); nine catchments exceeded and five were below.

Nitrogen:phosphorus ratios

Exceedances of guidelines for either nitrogen or phosphorus do not necessarily lead to algal proliferation. Given sufficient light, and suitable water temperatures and substrate conditions, the extent to which nutrient concentrations will lead to nuisance plant growth is controlled largely by the relative abundance of dissolved nitrogen to phosphorus (ie, the SIN:DRP ratio).

Wilcock, Biggs et al (2007) suggest that SIN:DRP ratios of 50:1 and 3:1 probably indicate phosphorus-limitation and nitrogen-limitation of algae growth, respectively, unless concentrations of both SIN and DRP are well above those expected to saturate growth (in which case neither nutrient is limiting).

Table 6 shows the SIN:DRP ratios for each monitored dairy catchment (where data was available), derived from medians for each of the variables.

Table 6: SIN:DRP ratios for each monitored dairy catchment (based on data for the most downstream site)

Catchment Median concentrations (mg/L) SIN:DRP ratio
SIN DRP
Inchbonnie 0.388 0.059 7
Waiokura 2.846 0.032 89
Rhodes 9.374 0.008 1,172
Petrie 4.627 0.015 308
Toenepi 1.212 0.089 14
Puwera 0.276 0.048 6
Washpool 0.824 0.037 22
Waikakahi 1.782 0.075 24
Powell 1.098 0.007 157
Enaki 1.135 0.026 44
Bog Burn 0.775 0.023 34
Taharua 1.135 0.011 103
Mangapapa 0.86 0.017 51
Rai 0.63 0.011 57
Guideline (Biggs, 2000) for 20-day accrual period 0.295 0.026  

Note: Bolded ratios indicate strongly p-limiting conditions.

It is not possible from the data presented in table 6 to make conclusive statements about the extent to which nutrients are limiting growth in each of the catchments. However, four catchments, including three from the South Island, have SIN:DRP ratios exceeding 100 as well as relatively low median DRP concentrations. These catchments are likely to be strongly p‑limited (on average). Two catchments, Inchbonnie and Puwera, have considerably lower SIN:DRP ratios than the other catchments, and nitrogen is likely to be the limiting nutrient (on average) in these catchments.

The interpretation above is based on a simplistic analysis and is intended only to provide an indication of average nutrient conditions in relation to nuisance growth potential. SIN:DRP ratios can fluctuate considerably within catchments and across seasons. For example, although the Toenepi Stream is, on average, tending towards p-limitation, with a SIN:DRP ratio of 25, during the summer low-flow period the stream is known to be N-limited (Wilcock, Biggs et al, 2007). Phosphorus inputs were clearly variable in all of the catchments (see the ranges in table 4), providing further indication that there are likely to be times when this nutrient is not limiting.

Nuisance weed and algal growth

The type of plant or algal growth that occurs in a stream is largely dependent on bed conditions. In hard-bottomed (rock- or gravel-bedded) streams, periphyton in the form of algal slime on the rocks (diatom algae) and filamentous strands of greenish algae attached to rocks, is the most common form of nuisance growth. In soft, sandy-bottomed streams, some filamentous periphyton may establish, but submerged and emerging macrophytic ‘weeds’ are generally more prevalent. Given sufficient nutrients, excessive algal and weed growths are most common in late summer or early autumn, when stream flows are lowest and water temperatures are highest. However, nuisance growth can occur in winter (eg, Horizons Regional Council, 2007).

Although comparing nuisance growth between sites and over time is difficult, because measures are generally subjective (eg, observer estimates of bed coverage), some published information is available for some of the monitored catchments to help characterise this aspect of water quality.

Wilcock et al (2006) summarised the weed and periphyton growth in the five Tier 1 catchments qualitatively as follows.

  • Pigeon Creek in the Inchbonnie catchment is a stony, hard-bottomed stream and has summer blooms of periphyton and filamentous green algae.

  • The Toenepi Stream is soft-bottomed and at times has a high biomass of emergent macrophytes, notably swamp willow weed, as well as submerged macrophytes and filamentous green algae.

  • The Waiokura Stream is a soft-bottomed stream that is reasonably well shaded and periphyton blooms are not considered a problem (by the author).

  • The Bog Burn has a gravelly substrate but with some fine sediments. Macrophyte cover is limited, but periphyton mats and filamentous green algae do occur in summer.

With respect to the Tier 2 catchments, the following records have been made.

  • The Rhodes and Petrie Streams both have submerged and emerging macrophyte growth, often to the extent that the channels are choked during summer months and require weed spraying and mechanical clearing (Environment Canterbury, 2007). Periphyton and macrophyte growth has been classed by Environment Canterbury during their habitat assessments as “marginal” to “poor” (on a qualitative scale) over all years of the monitoring programme (2000–2007).

  • Horizons Regional Council (2007) report that in the Mangapapa catchment, periphyton biomass and coverage were measured in 2007 at up to 12 sites. Mean periphyton biomass ranged from 8.7 mg/m2 at the uppermost reference site to 152.5 mg/m2 in one of the tributaries of the Mangapapa Stream, lower in the catchment. Five of the nine sites at which biomass measurements were made exceeded the national guideline value for the protection of benthic biodiversity of 50 mg/m2 chlorophyll a (Biggs, 2000). With respect to bed coverage, half of the 12 sites exceeded 60 per cent for diatom algae – the national guideline threshold for aesthetic quality (Biggs, 2000) – and five of these exceeded 80 per cent coverage. Only one site exceeded the national aesthetic guideline of 30 per cent (Biggs, 2000) for filamentous green algae.

  • Summer periphyton (Rapid Cover Assessment) scores from 2006 to 2009 in the lower Powell catchment are typically below 5, indicating excessive periphyton growth (scores close to zero indicate periphyton is abundant, close to 10 indicate periphyton absence). Across the rest of the Powell catchment in the summer of 2007 only one other site produced a score less than 7, indicating low quantities of algae (T James, Tasman District Council, pers. comm., 2009).

  • Sampling in the Enaki catchment between 2002 and 2007 shows that excessive periphyton growth has occurred on occasion. Biomass (chlorophyll a) ranged from 4.3 to 45.3 mg/m2 for most years (below the 50 mg/m2 guideline for protecting benthic biodiversity) but reached 801 mg/m2 in 2003, indicating bloom conditions. The aesthetic guideline for filamentous cover (30 per cent) was exceeded on three out of 52 monthly sampling occasions (Greater Wellington Regional Council, 2008).

  • Periphyton biomass was measured on six occasions at each of three monitoring sites on the Taharua River between 2000 and 2008. All three sites had average measurements of less than 20 mg/m2, while the most downstream site had a maximum measurement that exceeded 50 mg/m2 but was below the 120 mg/m2 guideline being applied by Hawke’s Bay Regional Council (Hawke’s Bay Regional Council, in press).

To summarise the above results: of the nine catchments (combining results for Rhodes–Petrie) for which measurements or observations have been made, all but two have exhibited excessive weed or algal growth on occasion. However, the subjective elements of this assessment need to be restated and point towards a need to improve the way quantitative periphyton and macrophyte data is collected and assembled for the purpose of national-scale reporting (see ‘Summary and recommendations’).

Case Study 1: Removal of dairy herd stream crossings on the Rai River, Marlborough

The adverse effects of dairy herds crossing the Rai River and its tributaries have been identified as a priority for action by the Marlborough District Council (2003). In addition to elevated bacterial loadings at crossing points, nutrient pulses coinciding with herd crossing times (ie, for milking) are obvious from in-stream monitoring data collected by the council (see figure 7).

Figure 7: Nitrogen pulses during herd crossing on a tributary of Rai River over two days in September 2001

View graph|View text description
Figure 7: Nitrogen pulses during herd crossing on a tributary of Rai River over two days in September 2001

Source: Marlborough District Council.

Note: A logarithmic scale has been used.

A stream crossing survey by the Council in 2003, identified 112 crossings in the catchment. Forty-eight of these crossings were considered a high priority for elimination (ie, replacement with a bridge or culvert). Since then, each farmer in the catchment has entered into an agreement with the Council to eliminate the highest-priority crossings on their land. By 2007, a total of 28 high-priority crossings, and a further 28 less significant crossings, in the Rai catchment had been eliminated (Marlborough District Council, 2007).

Although quantitative data on the individual effect of each crossing removal is not available, some qualitative observations can be made. For example, visual assessments indicate marked improvements in water and habitat quality immediately downstream of crossing points after a bridge or culvert has been put in place (see figure 8). Marlborough District Council report that where crossings have been eliminated, the stream substrate has in all cases changed from a clogged, fibrous bed to open cobbles with clear water and abundant macroinvertebrates (Marlborough District Council, 2008). Furthermore, it is logical to expect that the spikes in nutrient (and other pollutant) concentrations associated with herd crossings have been significantly reduced.

Figure 8: Example of stream bed condition before (left) and after (right) herd crossing removal on a tributary stream in the Rai catchment

Figure 8: Example of stream bed condition before (left) and after (right) herd crossing removal on a tributary stream in the Rai catchment

Source: Marlborough District Council, 2008.

Although local-scale improvements are obvious, the collective effect on catchment water quality of all crossing eliminations so far has not been quantitatively assessed. It is the intent of the ongoing catchment-wide water-quality monitoring programme to do this over time (Marlborough District Council, 2008).

Stressors and toxicants

Figures 9 to 12 show medians and ranges for several water-quality measurements that indicate toxic or otherwise stressful conditions for aquatic organisms. Electrical conductivity is included in this section because it can indicate the presence of pollutants, although it is not a stressor per se. Some water-quality measurements described below (dissolved oxygen and water temperature) often fluctuate considerably during the course of a normal day. The implications of such daily variations for interpreting the measurement statistics are discussed in the ‘Data constraints and limitations’ section.

Ammoniacal nitrogen

Ammoniacal nitrogen can, at sufficiently high concentrations, be toxic to fish and other aquatic life (in addition to contributing to eutrophication). In farmed catchments, elevated concentrations of this compound generally arise from stock effluent reaching the streams via direct discharge, paddock run-off or direct stock access to stream banks and beds. This is most likely to be exacerbated when stream flows are low (eg, in late summer), when cattle are often near waterways and when dilution rates are low. Run-off and leaching of urea fertiliser can also contribute.

The concentration at which ammoniacal nitrogen becomes toxic is dependent on stream water temperature and pH (ANZECC, 2000). ANZECC recommends adopting a trigger value of 0.9 mg/L ammonia nitrogen for pH 8 and 20oC to adequately protect 95 per cent of species. The average of measured catchment maximums was just over 8 for pH (provided in table A2-I but not graphically presented) and just over 20oC for water temperature; this indicates that a 0.9 mg/L trigger value for ammoniacal nitrogen is a reasonable guideline to report against for these catchments (ANZECC, 2000). Maximum measured water temperature was just over 25oC (Enaki) and maximum measured pH was 9 (Bog Burn), indicating conditions in some of the catchments where ammonia may become toxic at lower concentrations of around 0.4 mg/L.

Figure 9 shows that four catchments – Inchbonnie, Toenepi, Puwera and Washpool – had peak ammoniacal nitrogen concentrations that came very close to, or exceeded, 0.4 mg/L. Two of these catchments had peak concentrations that exceeded 0.9 mg/L. All catchments except Taharua had median concentrations of ammoniacal nitrogen that were elevated above concentrations found in predominantly natural catchments (see figure 9).

While there has been a relatively high rate of progress in the Washpool catchment towards Accord targets aimed at excluding stock from waterways (ie, 94 per cent of Accord waterways fenced and only one stream crossing still existing, see table 4), this catchment has the highest median concentration of ammoniacal nitrogen. Sampling from multiple sites throughout this catchment (reported by Otago Regional Council, 2007) reveals that a high proportion of ammonia in the main stem of the Washpool Stream originates from tributary mole-and-tile drains (artificial channels used to drain paddocks, which are particularly common in Otago). Although mole-and-tile drains have been specifically identified for management action through an Accord-related memorandum of understanding between Otago Regional Council and Fonterra, there is some progress yet to be made to meet targets.11

Figure 9: Medians (crosses) and ranges for ammoniacal nitrogen for the most downstream sites during the monitoring period

View graph|View text description
Figure 9: Medians (crosses) and ranges for ammoniacal nitrogen for the most downstream sites during the monitoring period

Notes:

A log scale has been used on the y axis.

The upper dashed lines are based on the ANZECC toxicant trigger values of 0.4 mg/L for 95% ecosystem protection (assuming a water temperature of 25oC and pH 9), and 0.9 mg/L for 95% ecosystem protection (assuming a water temperature of 20oC and pH 8). The lower dashed line is the median for predominantly natural catchments.

Data and further information on guidelines are provided in table A2-H in appendix 2.

Ammoniacal nitrogen concentrations at the Rai catchment site were reported as median of <0.01 mg/L and a maximum of 0.01 mg/L.

T1 = Tier 1 catchment and T2 = Tier 2 catchment.

Dissolved oxygen

The concentration of oxygen that is dissolved in water and available for respiration by aquatic animals is a crucial indicator of the life-supporting capacity of a stream. Dissolved oxygen concentrations are commonly increased during the daytime by the photosynthesis of periphyton and aquatic plants and reduced at night by periphyton and microbial respiration and the decay of organic matter. As a result, streams with undesirable organic waste discharges or proliferations of algal and weed growth often exhibit large fluctuations in dissolved oxygen, including dangerous depletions at night.

The depletion concentrations at which aquatic species show signs of impairment are variable. However, a study by Deans and Richardson (1999) indicates a drop below 5 mg/L (equating roughly to 50–60 per cent saturation) may begin to affect some less tolerant fish species in New Zealand. Another study by Landman et al (2005) indicates prolonged (ie, 48-hour) concentrations of less than about 3 mg/L start to have lethal effects on inanga (whitebait). The Resource Management Act 1991 suggests a minimum threshold of 80 per cent saturation12 to protect aquatic life, while ANZECC (2000) suggests saturation levels between 98 and 105 per cent are optimal for lowland reference sites (ie, only slightly disturbed ecosystems).

Dissolved oxygen statistics (mostly from monthly spot measurements) are shown in two plots in figure 10. This is because both methods of measurement (per cent saturation and concentration) were employed by the monitoring agencies who provided the data and there is no ready conversion from one method to the other (without accounting for temperature). The five Tier 1, as well as Taharua, Enaki, Rai and Powell, catchment results are given as per cent saturation, and medians were all above 80 per cent, indicating that daytime oxygen concentrations were generally adequate to support aquatic life. However, six of these nine catchments had daytime minimums that indicated oxygen concentrations fell below the guideline on occasion, and the Toenepi daytime median of 80.7 percent saturation indicates more frequent oxygen sags below the guideline.

Figure 10: Medians (crosses) and ranges for dissolved oxygen for the most downstream site (unless specified otherwise) during the monitoring period

View graph|View text description
Figure 10: Medians (crosses) and ranges for dissolved oxygen for the most downstream site (unless specified otherwise) during the monitoring period

Notes:

Powell data in the left figure (%SAT) is from monthly daytime spot measurements over one year, and in the right figure (mg/L) is from two-week summer deployment of an oxygen meter. (Note the units differ from data for Powell in the left graph.)

The y axes are different in the two graphs.

The guideline is the RMA guideline for ecosystem protection (80 per cent).

T1 = Tier 1 catchment and T2 = Tier 2 catchment.

Furthermore, several catchments had maximums well over 120 per cent, or 12 mg/L, indicating occasional super-saturation with oxygen during the daytime (eg, Mangapapa, Toenepi, Washpool). This is indicative of high rates of photosynthesis during the day (probably related to eutrophication) but also suggests, correspondingly, that there may have been high rates of respiration depleting oxygen at night (in the absence of photosynthesis). While two of the catchments mentioned above did indeed exhibit relatively low dissolved oxygen minimums, giving an inkling of the overall fluctuation, daytime measurements are unable to pick up the true bottom of the oxygen profile (which often occurs around dawn).

A much better representation of dissolved oxygen concentrations is gained from continuous 24-hour measurements (ie, measurements that include night-time measurements). Such measurements have been made in four of the Tier 1 catchments and indicate that actual oxygen minimums were about 10 percentage points lower (indicating greater oxygen depletion) in these catchments than the routine monthly monitoring results suggest (with minimums ranging from about 25 to 60 per cent saturation in the Toenepi, Waikakahi and Bog Burn catchments).13 Only the Waiokura catchment remained relatively well aerated throughout 24-hour periods and across seasons. Night-time dissolved oxygen concentrations as low as 6 per cent saturation have been measured at Tahuroa Road monitoring site in the Toenepi Stream (Wilcock et al, 2006).

Continuous measurements have also been made in the Powell catchment. These are shown, in addition to the monthly results for the Powell catchment, in figure 10 (right graph). These results are from the short-term (two-week) summer deployment of an oxygen logger. The short-term summer median (5.35 mg/L) is much lower than the medians from monthly measurements for the other Tier 2 catchments (9.3–11.2 mg/L), while the range is much greater. This highlights the need to consider seasonal continuous data (especially in summer low-flow conditions), in addition to monthly daytime measurements, when making assessments of water-quality change.

Water temperature

Stream water temperatures that vary too much from the natural range, or climb too high, can be detrimental to aquatic life. For example, temperatures exceeding 22oC begin to have lethal effects on some mayfly insects (Quinn et al, 1994), while temperatures over about 30oC may be lethal to some fish, such as inanga/whitebait (eg, Richardson et al, 1994). Generally, pastoral streams are susceptible to warm spikes in temperature as a result of riparian vegetation (shade) removal and channel disturbance reducing natural flows.

Figure 11 shows that most individual catchments had a seasonal range (as defined by single daytime readings) of at least 10oC. Two catchments, the Powell and Enaki, had measured peak temperatures reaching, or in excess of, 25oC – a threshold that is considered to be the upper tolerance limit for trout (Schedule 3 of RMA, 1991). In the case of the Powell catchment, this measurement is an actual peak as it comes from a continuous temperature sensor record (as opposed to the highest single daytime measurement).

Figure 11: Medians and ranges for the most downstream site during the monitoring period for water temperature

View graph|View text description

Notes:

Powell data from two-week summer deployment of an oxygen meter.

Data, details on monitoring periods and sample numbers are provided in table A2-K in appendix 2.

T1 = Tier 1 catchment and T2 = Tier 2 catchment.

Electrical conductivity

Electrical conductivity is a measure of the total dissolved salts or ions in the water. Elevated concentrations may indicate the presence of point-source discharges (eg, effluent) or diffuse nutrient inputs, but can also be a naturally occurring result of catchment geology.

Median electrical conductivity generally lies in the range of 100–300 microS/cm across catchments (see figure 12). The Mangapapa catchment shows the greatest range in monthly measurements (as it does for many of the other measurements) discussed in this report, including some of the nutrient forms. The pattern across catchments is broadly correlated with nitrate–nitrite nitrogen (eg, the Rhodes–Petrie and Waiokura catchments register high values for both measurements).14 The most notable exception is the Puwera catchment, where conductivity is high but nitrate–nitrite nitrogen is very low.

Figure 12: Medians (crosses) and ranges for electrical conductivity for the most downstream site during the monitoring period

View graph|View text description
Figure 12: Medians (crosses) and ranges for electrical conductivity for the most downstream site during the monitoring period

Notes:

A log scale is used on the y axis.

Data and details on monitoring period and sample number are provided in table A2-L in appendix 2.

Where bars go off the edge of the y axis scale the minimum or maximum is noted next to the bar.

T1 = Tier 1 catchment and T2 = Tier 2 catchment.

Sediments and visual quality

Suspended solids and turbidity are important indicators of aquatic habitat and visual quality and affect human values such as fishing, swimming and amenity. If concentrations of suspended solids are too high for prolonged periods, mobile species (eg, fish) may not have sufficient light to navigate and feed effectively, and juvenile recruitment or passage of fish into catchments may be limited (Richardson and Jowett, 2001). As fine sediments settle out of the water column, benthic habitats may be smothered.

High suspended solids concentrations are commonly associated with higher flows and are also naturally elevated in catchments with soft (erosion-prone) geology or sandy-bottomed streams. However, high suspended solids and turbidity (which generally result in low visual clarity (ANZECC, 2000) may also indicate stream bank and paddock erosion associated with poor land management.

Suspended solids and turbidity are generally correlated for each catchment,15 with Waiokura and Washpool having particularly high medians for both measurements (see figure 13). Four of the catchments have median turbidity levels that are at, or in excess of, ANZECC guidelines for ecosystem protection (5.6 NTU). Cawthron Institute research on trout fisheries indicates that turbidity in excess of the ANZECC guideline may result in a reduction in visual foraging area of drift-feeding trout of about 60 per cent (from clean water conditions), even for small fish (< 10 cm) (John Hayes, Cawthron Institute, pers. comm., 2009). A similar relationship is likely for drift-feeding native fish such as inanga and smelt, with a proportional decrease in energy (food) intake and adverse consequences for fish growth, condition and possibly survival (John Hayes, Cawthron Institute, pers, comm., 2009).

Figure 13: Medians (crosses) and ranges for the most downstream site during the monitoring period for suspended solids and turbidity

View graph|View text description
Figure 13: Medians (crosses) and ranges for the most downstream site during the monitoring period for suspended solids and turbidity

Notes:

A log scale is used on the y axes.

The dashed line in the lower graph is the guideline threshold for turbidity (5.6 NTU, from ANZECC trigger values for ecosystem protection in lowland waterways).

Data, details on monitoring period and sample number are provided in tables A2-M and A2-N in appendix 2.

T1 = Tier 1 catchment and T2 = Tier 2 catchment.

Although concentrations of suspended solids indicate potential habitat impacts, measurements of deposited fine sediment (as percentage sediment cover on the stream bed) would add important additional information. Deposited fine sediment has been identified as a key stressor for agricultural stream ecosystems in New Zealand in recent years (eg, Niyogi et al, 2007) because it tends to accumulate during periods of normal and low stream flows. As a result, it affects stream organisms for longer periods of time than do suspended solids. This measurement could be considered for inclusion in ongoing catchment monitoring programmes.

Visual clarity data was not sought for this report,16 but a summary for the period 2001 to 2006 has been published for the five Tier 1 catchments by Wilcock, Monaghan et al (2007). Median visual clarity was closely correlated with suspended solids (r2 = 0.75) and turbidity (r2 = 0.69) and ranged between 0.38 metres (Waiokura) and 1.4 metres (Toenepi). For reference, the ANZECC (2000) guideline (trigger value) for ecosystem protection in lowland waterways is 0.6 m.

Benthic macroinvertebrates

In addition to the measurements of water quality already described, benthic macroinvertebrates (ie, the insects, worms and snails that live on stream beds) are also good indicators of stream water quality, habitat quality and overall ecosystem health. Samples of macroinvertebrate population numbers and species types can be used to calculate biotic indices, such as the Macroinvertebrate Community Index (MCI),17 which reflect the range of water quality and habitat conditions experienced in a stream over time. In a degraded stream, for example, the macroinvertebrate community will be dominated by pollution-tolerant species such as snails, worms and midge larvae. In a more pristine stream, larvae of insects such as mayflies and caddisflies will predominate.

Table 7 provides a summary of the average MCI and Quantitative MCI (QMCI) scores from the monitored catchments for which data is available (10 out of 14 – considering Rhodes–Petrie catchments together). Where multiple sampling sites existed within catchments, scores have been averaged. However, upper catchment sites (ie, those that could be considered unaffected) have been excluded.

One way of interpreting macroinvertebrate scores is to apply the water quality categories defined by Wright-Stow and Winterbourn (2003) and presented in table 8. Since there are some differences between catchments in terms of the type of metric available, it is not appropriate to apply these categories in a rigid manner, but they can be used to broadly characterise the catchments: five catchments had average QMCI scores above 5.0 (or a high MCI in the case of Taharua), indicating relatively clean water or mild degradation, while the other five had average QMCI scores below 5.0, indicating moderate to severe pollution.

Table 7: Average MCI and QMCI scores

Catchment Method of score calculation and data source Substrate Average MCI score
(range between site averages in brackets, or standard deviation)		 Average QMCI score
(range between site averages in brackets, or standard deviation)
Taharua (T2) 3 sites sampled 4–6 times each between 1999 and 2005. Average for each site used to calculate overall average
(Hawke’s Bay Regional Council, 2006)		 Predominantly mobile pumice sands 122
(110–147)		  
Waiokura (T1) 3 sites sampled on 3 occasions between 2001 and 2003 (2 summer and 1 winter). Average for each site used to calculate overall average
(Scarsbrook et al, 2005)		 Dominated by fine silt and sand upstream, becoming coarser downstream 111
(SD = 10.7)		 6.2a
(SD = 0.3)
Enaki (T2) 1 site, 3 replicate samples taken annually between 2002 and 2007. Mean scores given in columns to the right
(Greater Wellington Regional Council, 2008)		   104
(SD = 11.4)		 5.6a
(SD = 1.25)
Bog Burn (T1) 3 sites sampled on 3 occasions between 2001 and 2003 (2 summer and 1 winter). Average for each site used to calculate overall average
(Scarsbrook et al, 2005)		 Variable. Becoming progressively finer downstream. Moderate concentrations of deposited fine sediment have been observed on occasion 100
(SD = 6.5)		 5.3a
(SD = 0.58)
Waikakahi (T1) 3 sites sampled on 3 occasions between 2001 and 2003 (2 summer and 1 winter). Average for each site used to calculate overall average
(Scarsbrook et al, 2005)		 Gravel and cobble bed. Fine sediment deposits are typically a minor component 99
(SD = 8.2)		 5.3a
(SD = 1.13)
Mangapapa (T2) Average of 6 sites on the Mangapapa and tributaries sampled on 1 occasion each during low flow in 2007b
(Horizons Regional Council, 2007)		 Predominantly gravels and small cobbles with 3 sites dominated by sands and silts 89
(73–105)c		 4.2
(2.9-5.9)c
Powell (T2) Average of 5 sites on the Powell and tributaries sampled on 2 occasions each during 2006/2007b
(Tasman District Council, 2007)		 Variable. Some sites dominated by silts and muds, others gravel and cobble 81
(56–95)		 4.0
(2.75–4.49)
Puwera (T2) 2 sites on the main stem sampled 4 times each between 2006 and 2007. Average for each site used to derive overall average
(Northland Regional Council, 2007)		   82
(74–89)		 3.6
(2.76–4.05)
Rhodes (T2) 2 sites sampled 12 times between 1999 and 2006. Average of each site used to derive overall average
(Environment Canterbury, 2007)		 Variable along length from heavily sedimented to gravelly   3.5
(1.76–4.74)
Petrie (T2) 1 site sampled 13 times between 1999 and 2006. Average score taken
(Environment Canterbury, 2007)		 Variable along length from heavily sedimented to gravelly   3.3
(1.97–4.59)
Rai (T2) No data available for this report      
Toenepi (T1)      
Inchbonnie (T1)      
Washpool (T2)      

Notes:

a SQMCI (not QMCI).

b The most upstream site (considered reference site) has been excluded.

c Ranges in brackets are for site averages (actual sample data was not available).

Numbers in brackets indicate ranges in metric scores: the minimum and maximum actual sample scores across all monitoring sites in each catchment, except for the Tier 1 and Enaki catchments where the average standard deviation is given and the Mangapapa catchment where the range of average site scores is given.

T1 = Tier 1 catchment and T2 = Tier 2 catchment.

SD = Standard deviation.

Table 8: MCI and QMCI degradation categories suggested by Wright-Stow and Winterbourn (2003)

MCI range QMCI range Degradation category
125–200 6.2–10.0 Clean (water)
105–115 5.2–5.7 Mild
85–95 4.2–4.7 Moderate
< 75 0–3.7 Severe

Although the MCI may be influenced by water quality, it may also be influenced to greater or lesser degrees by other catchment factors such as land and channel disturbance, stream substrate and flow conditions. For example, Environment Canterbury (2007) has attributed the low MCI scores in the Rhodes and Petrie catchments more to channel alteration and significant habitat degradation than to poor water quality per se. Conversely, the relatively high MCI scores in the Waiokura catchment are probably more a reflection of favourable substrate (gravel and rock) and flow conditions (eg, see analysis below) than good quality water.

Results of regression analyses indicate a relatively strong relationship between stream flow and average MCI (r2 = 0.89) and QMCI (r2 = 0.79) scores, although this analysis is based on only a small number of data points (see table A3-A in appendix 3). There is also an apparent, but weaker, linear relationship between each metric and percentage catchment area under dairy (r2 = 0.27 for MCI and 0.35 for the QMCI); note that the Waiokura catchment is excluded from this analysis as an outlier.

With respect to limitations, it should be noted that the MCI scores summarised in table 7 are averages from the results of all available monitoring sites in each catchment. The ranges of site results (given in brackets in table 7) indicate significant variability within catchments.

It should also be noted that MCI scores have been derived using the approach designed for hard-bottomed streams. There is now a soft-bottomed variant of the MCI (Stark and Maxted, 2007). The suitability of this variant for use in any of the monitored catchments should be determined.

Upstream vs downstream measurements

One of the methods for assessing the effects of best practice management in the monitored catchments over time is to compare water quality from upstream of the main dairy farming area with downstream water quality. Assuming water quality is higher upstream, if management interventions on the dairy farmland are effective, the gradient (ie, degree of difference) between upstream and downstream water quality would be expected to reduce over time.

Tables A2-A to A2-N in appendix 2 present median and range data for upstream sites (where available) in each monitored catchment. In some cases, such as in the Mangapapa catchment, these sites represent true reference conditions (ie, are upstream of all significant modified land use), while in other cases, such as in the Powell catchment, they are as far upstream as is practicable for the monitoring agency to visit but may still have significant modified land use above them.

Table 9 shows the percentage change between median values for nitrate–nitrite nitrogen, dissolved reactive phosphorus, E. coli and suspended solids at upstream sites and downstream sites for each catchment (where suitable comparative data existed).

Table 9: Comparison of upstream and downstream measurements

Catchment Percentage change in measurement between upstream and downstream monitoring sites
Nitrate–nitrite nitrogen Dissolved reactive phosphorus E. coli Suspended solids
Inchbonnie (T1) ­ (+42) ­ (+51) ­ (+24) ­ (+130)
Waiokura (T1) ­ (+22) ¯ (–55) ­ (+168) ¯ (–2)
Rhodes (T2) ­ (+28) ¯ (–38) ¯ (–67) ­ (+86)
Petrie (T2) ¯ (–13) ­ (+88) ­ (+92) ­ (+400)
Toenepi (T1) ND ND ND ND
Puwera (T2) ¯ (–38) ¯ (–53) ­ (+7) ­ (+33)
Washpool (T2) ­ (+1144) ND ¯ (–24) ­ (+159)
Waikakahi (T1) ­ (+9) ¯ (–23) ¯ (–76) ¯ (–71)
Powell (T2) ¯ (–21) ­ (+133) ¯ (–59) 0
Enaki (T2) ND ND ND ND
Bog Burn (T1) ­ (+529) ­ (+5) ­ (+2550) ¯ (–43)
Taharua (T2) ¯ (–55) ¯ (–39) ¯ (–18) ­ (+233)
Mangapapa (T2) ­ (+580) ­ (+240) ­ (+1367) ­ (+100)
Rai (T2) ­ (+1160) ­ (+10) ¯ (–60) ND
Total positive (­) gradients 8 6 6 7
Total negative (¯) gradients 4 5 6 3

Notes:

Positive numbers and upward arrows indicate a downstream increase (deterioration) in measurement value; negative numbers and downward arrows indicate a decrease (improvement). Numbers in bold indicate the magnitude of the increase or decrease was greater than 100 per cent of the upstream value.

Derived from data in tables A2-B to A2-N in appendix 2.

ND = No data, which in most cases means there is no data for both an upstream and downstream site for this catchment.

T1 = Tier 1 catchment and T2 = Tier 2 catchment.

The expected gradients in measurements – indicating water quality deterioration from upstream to downstream– are found in most of the catchments for nitrate–nitrite nitrogen, dissolved reactive phosphorus and suspended solids/turbidity. However, there are exceptions to this. In addition, gradients of E. coli increase in the downstream direction in half of the catchments, but decrease in the downstream direction in the other half. Of note, all of the largest changes, for all four variables (ie, those that are greater than 100 per cent of the upstream median value – see bolded numbers in Table 9), represent deteriorations in the downstream direction.

It is difficult to speculate about the reasons for variable patterns in the downstream gradient of water quality. The nature of land-use activities, stream inputs, and riparian disturbance and protection differ within, and between, catchments, highlighting the complex nature of the relationships between land use and water quality. For example, access to waterways may be easier for stock in upstream reaches where the channel is smaller than in downstream reaches where it may be more incised and/or protected by fences. This effect may more than offset the normal downstream accumulation of pollutants that results from other, less direct, sources along the stream length.

Land use above the upstream monitoring sites in the catchments under discussion can also account for ‘reverse’ gradients. For example, in the Powell catchment, both nitrate–nitrite nitrogen and E. coli concentrations are higher at the upstream monitoring site. Tasman District Council (2008) notes that there is a considerable amount of sheep and beef farming and unfenced waterway in the headwaters of the Powell catchment that may account for this loading.

The analysis presented in table 9 is descriptive only and should be validated in future reports with more in-depth assessment of the raw data for each site. For example, understanding the frequency with which gradients strengthen, weaken or reverse would reveal much more information about land-use influences and water-quality change through the dairy areas. It should also be noted that in those catchments where intensive dairying is continuous throughout the catchment (eg, the Rhodes and Petrie), the comparison of upstream and downstream sites is less useful for interpreting the effect of land-use interventions.

Observed temporal changes

Information on trends and changes in water quality over various time periods is available for some of the catchments. This information is summarised here because it provides a useful context for the baseline data and any future assessments of long-term trends in the monitored catchments.

However, it is important to note that the results of statistical analyses are only available for four catchments (Enaki, Taharua, Toenepi and Waiokura), and these should be considered preliminary findings because they are based on relatively short time series (five to seven years). All other results, unless otherwise stated, are based on comparative observations rather than formal statistical tests. It should also be noted that trends are not necessarily related to existing or recent land-use activities and may reflect historical farming activities.

In the Taharua River, statistically significant increasing concentrations of nitrogen (in both dissolved and total forms) have been measured at all three monitoring sites over the period 2001–2005, but the trends are strongest (up to 0.366 mg/L per year) at the closest downstream monitoring site to the farming activity in the catchment (Twin Culverts). Monitoring does not show any other significant trends in water quality (Hawke’s Bay Regional Council, 2006).

In the Enaki catchment, statistically significant decreasing trends in water temperature
(0.8oC per year) and total phosphorus (0.004 mg/L per year) were found for the period 2002 to 2006. Although the phosphorus result is not thought to be particularly ecologically significant, the decrease in water temperature may indicate that fencing and planting carried out by Greater Wellington Regional Council in recent years is starting to have some beneficial effect
(Perrie, 2008).

A comparison of sampling results between 2003 and 2006 for the Washpool catchment indicates that water quality is deteriorating (Otago Regional Council, 2007). Annual mean values of E. coli, nitrate–nitrite nitrogen, turbidity, ammoniacal nitrogen and total phosphorus were all higher in 2006 than in 2003. Only dissolved oxygen showed some improvement over this period (Otago Regional Council, 2007).

In the Rhodes–Petrie catchment, there is not yet enough water-quality chemistry data to perform robust statistical analysis. However, some observations have been made by comparing measurements made by Environment Canterbury (2007). The main long-term trends appear to be an increase in total nitrogen since the late 1990s and a decrease in habitat quality since 2000. Reduction in habitat quality is said to be related to siltation, removal of riparian vegetation and channel alteration (Environment Canterbury, 2007).

There also appear to have been some water-quality improvements in the Rhodes–Petrie catchment. There have been slight reductions in phosphorus and turbidity since 1999/2000, and a noticeable elimination of large spikes in these measurements. This may indicate improved bank stability and stock control in some places (Environment Canterbury, 2007).

Descriptive assessments of water-quality changes in several Tier 1 catchments were summarised three years ago by Wilcock et al (2006); briefly, the Bog Burn and Inchbonnie streams showed little change over the five years of monitoring from 2001 to 2006. The Waikakahi catchment, which also has a longer monitoring history, showed some improvement in suspended solids as a result of better riparian management, but changed little in other respects.

The Toenepi Stream, which has been monitored since the mid-1990s, has undergone more formal trend assessment. Wilcock et al (2006) report that, for the period 1995 to 2004, average water quality changed little but that there were some notable improvements, including statistically significant decreases in total nitrogen, ammoniacal nitrogen and suspended solids, and an improvement in visual clarity.

Recent analyses of data from the Waiokura catchment have revealed that some changes in water quality have occurred during the period 2001–2007 (Wilcock et al, in press). These include statistically decreasing concentrations of total and dissolved phosphorus, suspended solids and E. coli that have been attributed to improvements in point-source discharges, permanent stock exclusion and riparian planting. Concentrations of total and nitrate–nitrite nitrogen increased significantly over the period, reportedly because of intensification of land use and increased nitrogen cycling (Wilcock et al, in press).

No trend information is yet available for the Mangapapa, Puwera, Powell or Rai catchments.

 

6 It should be noted that some regional plans designate all water bodies as having recreational value. For example, the Manawatu Catchment Water Quality Regional Plan sets contact recreation standards that apply to all surface waters at flows under half median flow.

7 E. coli are a subset of the faecal coliform bacteria family. Therefore, a count of 400 E. coli per 100 mL will equate to a higher count of faecal coliforms (eg, a conversion of 200 faecal coliforms per 100 mL to 126 E. coli in marine waters is given in Ministry for the Environment and Ministry of Health, 2003).

8 Note that the periphyton guidelines are based on mean concentrations, which are being compared with median data in this report.

9 Olsen phosphorus: the plant-available (bicarbonate extractable) form of phosphorus in soil.

10 As defined by the land-cover layer in the River Environment Classification. Catchments are defined as ‘pastoral’ once pasture exceeds 25% of the catchment area (but may be a mix of dry stock cattle, dairy, sheep and other pastoral land uses).

11 The target agreed by Fonterra and Otago Regional Council is that “100% of dairy farms on tile and mole drained land to have the approved environmental management system section from the Best on Farm Practice Manual completed by September 2006”. Fonterra monitoring of 70 dairy farms in the Clydevale area, in which the Washpool catchment is included, indicated that, as of May 2008, approximately 35% of dairy farms had met the target completely while approximately 5% had made no progress (E Brown, Otago Regional Council, pers. comm., 2009).

12 The RMA guideline value is widely used in New Zealand to broadly benchmark dissolved oxygen concentration. However, it is recognised that councils and other agencies also develop and use other dissolved oxygen thresholds depending on their specific management aims. For example, Wilcock et al (2006) have proposed a threshold value of 40 per cent saturation for use in the Toenepi catchment in Waikato.

13 Estimated from a comparison of ranges from daily (ie, 24-hour continuous) dissolved oxygen and routine monthly daytime measurements reported in Wilcock, Monaghan et al, 2007.

14 Linear regression of the catchment medians for conductivity with nitrate–nitrite nitrogen (n = 12) had an
r2 of 0.59. The Puwera catchment was excluded from this regression as an outlier.

15 Linear regression of the catchment medians for suspended solids with turbidity (n = 12) had an r2 of 0.56.

16 Although, visual clarity is measured in all but two of the monitored catchments (see table A4-A) and could be considered for inclusion in future reporting.

17 See A User Guide for the MCI (Stark and Maxted, 2007) for a description of the use of the MCI and its variants in New Zealand.

 

[ |