River biology

Samples containing invertebrates are collected according to standard protocols for aquatic plants, fish, and macroinvertebrates and habitat depending on whether the stream has riffles and is stony, or is slow-flowing and has sufficient aquatic plants and submerged wood for sampling. The numbers and species of invertebrates in a sample are used to calculate different 'metrics' that reflect the health of the site at that time. 

The monitoring sites are wadeable which means they can be safely accessed to sample a representative range of habitats, and they are spread across the region. They include long-term monitoring sites (those sampled for more than ten years), reference sites unmodified by human activities, and a set of randomly selected sites of different size on developed land. We also have several sites on streams where riparian restoration has taken place or is planned. Reference sites are sampled annually, long-term monitoring sites are each sampled in two out of every three years, and random sites are each sampled once every three years.

More information is available on sampling design and the history of the invertebrate monitoring programme.

Monitoring sites

Over 520 sites have been sampled since 1994. Between 2012 and 2014 we monitored 252 sites. These sites were spread over eight river management zones and grouped int three high level categories.

Table 1: Invertebrate monitoring by river management zones

Monitoring frequency

Waikato Regional Council monitors ecological health of streams and rivers annually during the summer months. Sampling originally spanned mid-November through to March, but since 2002 all sampling has been conducted over January-March. 

Data history

By 2014, over 2,500 samples had been collected at more than 520 river and stream sites. From its inception, various changes to the monitoring programme have occurred, including the development of improved habitat assessment methods in 1999 and 2005, and changes in invertebrate collection methods to bring them in line with national protocols. From 2005-08, the network investigated relationships between stream health and different levels of upstream development for dominant stream types relative to a network of representative reference sites in unmodified catchments.

In 2009, the network was redesigned to:

• maintain key long-term monitoring sites and reference sites
• provide an unbiased estimate of ecological condition of non-tidal wadeable streams using a random site selection process incorporating streams of different size. 

Measurement technique

A site is defined as a 100 metre stretch of a stream or river that is representative of the general landscape setting. For the random sites the reach usually extends 50 m above and below a randomly selected point. 

Since 2002 we have used sampling methods consistent with national protocols for streams with riffles and stony bottoms (‘hard-bottom’), and slower-flowing streams with streambeds dominated by fine material but with sufficient aquatic plants and submerged wood suitable for sampling (‘soft-bottom’). We collect invertebrates using a D-frame kicknet with 500-micron mesh. 

Samples are stored in alcohol and taken to the laboratory. There each sample is spread over a gridded tray and random grids are sorted until at least 200 identifiable specimens have been found. The rest of the sample is then searched for species that were not found in the 200-invertebrate sample. These specimens/individuals are then identified under a dissecting microscope to the accepted level of taxonomy for calculating the Macroinvertebrate Community Index. Prior to 2002, the required sample size was 100 organisms. 

Further information about the accepted level of taxonomy and scores for calculating the Macroinvertebrate Community Index can be found in Appendix 2 of technical report Regional Guidelines for Ecological Assessments of Freshwater Environments.

Quality control procedures

Qualified and experienced personnel carry out all sorting and identification. An independent person re-identifies the invertebrates in 10 per cent of all samples for quality assurance and control every second year. All data entered into Waikato Regional Council’s database are checked for accuracy.

Data storage

The raw data is stored in Waikato Regional Council’s Ecobase database.

Waikato Regional Council uses the Macroinvertebrate Community Index (MCI) and a composite condition score called the Average Score Per Metric (ASPM) to report on trends over time, and the Quantitative MCI (QMCI) and the ASPM to report on state. 

The graphs presented here are the ASPM; both the ASPM and the QMCI provide similar conclusions for state but show different patterns for trends. 

Average Score Per Metric

We have developed a composite condition score called the Average Score Per Metric (ASPM) for Waikato streams. The ASPM integrates information from three metrics that reflect the sensitivity and diversity of the invertebrate community at a site. 

The ASPM is comprised of the following metrics which were found to have low variability among undeveloped reference sites in native forest:

• number of sensitive taxa ('species'): mayflies+stoneflies+caddisflies(EPT)
• percentage of sensitive taxa - %EPT
• tolerance of taxa to pollution: Macroinvertebrate Community Index (MCI).

These metrics were standardised by the maximum value from reference sites sampled in the same year to generate values potentially ranging between 0 and 1.0. The average of these standardised scores across the three metrics provided the ASPM.

When calculating these biological metrics for the entire dataset to examine trends over time, staff were faced with the dilemma of how to deal with the change in sampling protocols that took place in 2002 to bring sampling in line with national guidelines. One of the main issues was how to deal with the change from 100 to 200 sample counts recommended by the national protocols. To address this problem and enable comparisons of the data over time, a computer program called ECOSIM was used to randomly calculate 100-count data for all samples collected up to 2008 using the same level of taxonomy throughout (for example, all chironomids were combined and given a MCI score of five). If a sample had less than 100 invertebrates in it, it was not included in subsequent analyses. Comparisons showed that 100 and 200 sample counts provided similar values for %EPT, MCI and ASPM, so these can be compared over time. However, the number of sensitive taxa (EPT richness) was affected by the different sample counts and could not be compared over time. 

Interim condition classes for the ASPM have been developed by taking even splits between the worst case data point and the lower standard deviation of the mean of all reference site samples to define ‘poor’, ‘fair’, and ’good’. Samples above the reference site lower standard deviation were classified as ‘excellent’. These classes are considered interim because almost all available reference sites are in hard-bottom streams, and we do not know what to expect in unmodified low-gradient soft-bottom streams which now all occur in highly developed landscapes. Waikato Regional Council ecologists are currently working on ways to deal with this issue.

Habitat quality assessment

Habitat quality has been assessed at our monitoring sites annually since 1999. Habitat assessments take into consideration things like the amount of erosion and extent and quality of riparian vegetation. Find out more about our habitat quality assessment method. We do not currently have habitat quality classes for reporting purposes.

Over 2009 to 2011 we sampled 60 different sites each year using a probability survey design to provide an unbiased estimate of wadeable stream health around the region, the same 180 sites were sampled again over 2012 to 2014. Probability survey designs are widely used in the USA to extrapolate results from randomly selected sites to provide estimates the ecological health by stream channel length. By knowing the probability of site selection we can calculate stream kilometres and percentage of stream length using statistical methods in the computer program spsurvey. We used an unequal probability of selection by aiming for an even number of 1^st, 2^nd, 3^rd and ≥4^th order sites in our initial sample to cover a range of stream sizes, and from these selected ‘target’ sites for sampling. Our target population is perennial, non-tidal, wadeable streams on developed land, and we used the River Environment Classification (REC) drainage layer as the data frame for site selection. Because the REC does not accurately capture very small headwater streams, our results will underestimate actual stream length. We will repeat sampling of these sites every three years, and over time this will provide a picture of regional trends.

Samples are collected, processed and quality assured according to standard Ministry for the Environment protocols. There are no formal guidelines or standards for assessing invertebrate monitoring data in New Zealand. A quality assurance check is made of all data entered into Waikato Regional Council's Ecobase.

The ASPM has been developed  by Waikato Regional Council for use in this region; it has been peer reviewed and published in the New Zealand Journal of Marine and Freshwater Research (34: 367-378).  The ASPM provides a more holistic overview of stream health than single metrics by integrating diversity, compositional and tolerance metrics which may respond differently to various stressors. It is used in association with the more widely used pollution tolerance metrics the MCI and QMCI. Because almost all available reference sites are in hard-bottom streams, we do not know what to expect in unmodified low-gradient soft-bottom streams and therefore the condition classes developed for the ASPM are considered interim.

•  The ecological health of streams based on macroinvertebrates is grouped into the interim classes of 'excellent', 'good', ‘fair’, and 'poor'¹.
• Graphs represent the cumulative percent of wadeable stream length in the region shown by increasing percentage values, based on random site selection of different sizes of streams. 
• Higher macroinvertebrate and habitat indicators represent better ecological health while very high aquatic plant indicators are indicative of poorer health.
• The ecological health data for different parts of the region are expressed as the percentage of randomly selected sites sampled during summer over 2012 to 2014, and not as percentage of wadeable stream length (as explained in the technical report- The ecological condition of Waikato wadeable streams based on the Regional Ecological Monitoring of Streams (REMS) Programme)

Ecological health indicators, showing interim condition classes for macroinvertebrate health¹, over 2012-14 for Waikato wadeable streams expressed as a percent of regional stream length on developed land:

Interim macroinvertebrate health classes over 2012-14 for sites sampled in different parts of the Waikato region (for ASPM):

• 'Undeveloped' sites are those with catchments that have more than 90 per cent of their area in unmodified vegetation.
• 'Moderately developed' sites have 51-90 per cent of their upstream catchment area remaining in native forest. 
• 'Highly developed' sites have 11-50 per cent remaining in forest cover. 
• 'Very highly developed' sites have less then 10 per cent of their catchment remaining in forest cover. 
• Urban sites are all located within Hamilton City and were sampled once in 2006. 
• The ecological health data for other land use classes are from sites sampled over 2005-2008 (see the technical report: Spatial and Temporal Patterns in the Condition of Waikato Streams Based on the Regional Ecological Monitoring of Streams (REMS) Programme). 

Ecological health of the region’s streams declines progressively with increasing levels of catchment development, and is lowest in urban streams influenced by stormwater.

Habitat quality declines markedly with catchment development and is lowest in streams draining catchments almost entirely developed for agriculture. Lower habitat quality is mainly because of the lack of riparian forest cover and increased bank erosion along streams in developed pastoral catchments, although it can also partly reflect the low slope of many lowland stream channels. Habitat quality is rated higher in urban streams than in highly-developed pasture streams because many streams in Hamilton city flow through vegetated gullies.

Long-term trends in macroinvertebrate metrics show that some are declining and some are increasing over time at particular sampling sites. Because the long-term sites have not been randomly selected, conclusions cannot be drawn about regional trends in wadeable stream health. However, across the 37 sites on developed land that we have sampled for at least 10 years, there is an overall decline in MCI which was developed as an indicator of water quality. There was no change over this time in ASPM which integrates a wider range of effects. Overall, across the 5 long-term monitoring sites where riparian management has been carried out there is an overall improving trend in both stream health metrics.