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

A Cultural Health Index for Streams and Waterways: A tool for nationwide use

1 Introduction

2 Development of the Cultural Health Index

3 Evolution of the Cultural Health Index for streams and rivers

4 Engagement with Māori values

5 Conclusion

Glossary

References

Acknowledgements

Appendix 1: Recording form used to assess sites in the Taieri, Kakaunui and Hakatere catchments

Appendix 2: Recording form used to assess sites in the Tukituki ctchment

Appendix 3: Table used to assign the 1-5 scores for the nmber of species present at a site

Appendix 4: Final recording form used to assess sites in any catchment

Appendix 5: Examples of site assessments completed in the four catchment studies

Appendix 6: Ministry for the Environment's Environmental Performance Indicator Programme

3 Evolution of the Cultural Health Index for streams and rivers

The Cultural Health Index is made up of three components:

1. site status, specifically the significance of the site to Māori

2. a mahinga kai measure

3. a stream health measure.

In this section we provide:

• an overview of the overall structure of the index
• a description of each of the components of the index and how they evolved as the index was applied to a new river and a new iwi.

3.1 Overall structure of the Cultural Health Index (CHI)

Component 1 - site status

This component of the CHI explores the significance of the site to Māori and distinguishes between traditional and contemporary sites.

There are two questions to establish this component:

The first question requires a site to be classified as either:

A - indicates a traditional site of significance to Māori or

B - indicating the site is not traditional but has been included because of other aspects (eg the site may be one monitored by the regional council).

The second question asks whether Māori would return to the site in the future. If the rūnanga would return, the site is awarded a 1 and, if not, a 0. When the answers to the two questions are collated there are four possible combinations:

Component 2 - Mahinga kai measure

The second component of the CHI addresses the mahinga kai values of a site. This component, in addition to encapsulating the many intangible qualities associated with the mauri of a waterway, is tangibly represented by some of the physical characteristics of a freshwater resource including: indigenous flora and fauna, water clarity, water quantity, and the mahinga kai it yields (Ministry for the Environment 1997).

There are four parts to the 'mahinga kai measure' of the Cultural Health Index.

1. The first part (a) identifiesmahinga kai species present at the site. A list of plant, bird and fish species is prepared. A score (1-5) is then assigned, depending on the number of species present.

2. The second part (b) compares the species present today and the traditional mahinga kai sourced from the site. This was deliberately factored into the design of the Cultural Health Index to recognise that maintaining cultural practices, such as the gathering of mahinga kai, is an important means of ensuring the transference of cultural values through the generations. Cultural continuity means that greater value is likely to be assigned to sites of traditional significance that continue to support the mahinga kai species sourced in the past. A single score (1-5) is assigned, based on the number of species of traditional significance that are still present:

• non-traditional site scores 1
• none of the species sourced in the past is present at the site scores 1
• at least 50percent of the species sourced in the past are still present at the site scores 3
• all species sourced in the past are still present at the site scores 5.

3. Mahinga kai gathering assumes Māori have physical and legal access to the resources that they want to gather. The third part of the mahinga kai measure (c) assesses each site based on access to the site. (No access scores 1 and unimpeded legal and physical access scores 5.)

4. The fourth part in the mahinga kai measure (d) assesses whether Māori would return to the site in the future and use it: No scores 1, Yes scores 5.

The four mahinga kai elements are then averaged to produce a single score (1-5).

Component 3 - Cultural Stream Health Measure

The third and final component of the CHI is the Cultural Stream Health Measure (CSHM). Indicators of stream health identified in Part 1 of the study have been tested and refined by different iwi assessment teams on the four rivers studied. This has resulted in a set of indicators that best reflects iwi participants' assessment of overall stream health and that can be defined objectively. Each of these eight indicators receives a score (1-5) from each rūnanga member involved in the assessment. The scores for each indicator are then averaged. The average of all indicator scores is calculated as the CSHM (1-5).

Figure 1: Example of an assessment site: Kakaunui Catchment

Application of the Cultural Health Index results in a score of A-0/ 2.56/ 1.06 (representing each of the three components).

The site at Island Stream was traditionally used by the Tipa whānau who travelled there each autumn to harvest eels during the downstream migration. Because of this traditional use, the site is classed as an 'A'. However, its degraded condition means the whānau would not return to use the site, hence component 1 scores A-0. The mahinga kai measure score is in the middle of the range (2.56 out of 5) because while two of the factors included in the measure score highly, the other two receive a low score. The stream health measure (1.06 out of 5) confirms the poor health of this site.

3.2 The evolution of Component 1 of the CHI: site status

Oral records from rūnanga members reinforced the relevance and usefulness of the overall structure of the CHI (components 1-3). There was no call for addition or deletion of any component, however, refinements of a generally minor nature were made as the research progressed.

Each of the river studies (Taieri/Kakaunui, Hakatere, and Tukituki) confirmed that the structure of Component 1 (site status) accommodated the values and concerns of Māori in different iwi (Ngāi Tahu and Ngāti Kahungunu). However, a minor refinement was made as the study progressed, eg the original definition for the Taieri/Kakaunui study, recognising culturally significant sites and ability to sustain future mahinga kai use, was refined to recognising culturally significant sites and ability to sustain future cultural use in the Hakatere study. This broader definition to cultural use was maintained in the Tukituki study.

For some sites, members of the assessment team were not unanimous about whether they would return to the site. However, the assessment of the majority was always clear and able to be used as the final determinant of the score for that dimension.

This component of the CHI enables Māori and resource managers to make comparisons between different catchments. Table 2, for example, shows how rivers vary in the proportion of sites recorded as traditional and to which rūnanga would return or not. While only 25 percent of Taieri sites would be revisited, Māori would return to more than 70 percent of sites on the rivers of the other three catchments.

Table 2: Number of traditional sites being accorded A-1 and A-0 status

Table 3 reveals the percentage of all sites, traditional and non-traditional combined, that tangata whenua would return to in future. This ranges from 23.3 percent (Taieri) to 73 percent (Tukituki).

Table 3: Number and percentage of all sites, traditional and non-traditional combined, to which Māori would return (A1 or B1 status)

In the future, a 'traffic light' system could be used to graphically display overall catchment scores - for example:

• red: all traditional sites are unable to sustain cultural use - ie would not be visited by Māori in the future
• orange: most sites (the majority) are unable to sustain cultural use - ie would not be visited by Māori in the future
• yellow: most sites (the majority) are able to sustain cultural use - ie would be visited by Māori in the future
• green: all traditional sites are able to sustain cultural use - ie would be visited by Māori in the future.

3.3 The evolution of Component 2 of the CHI: mahinga kai

This component of the Index recognises that mauri is tangibly represented, in part, by some of the physical characteristics of a freshwater resource including the mahinga kai it yields (Ministry for the Environment 1997).

In the Taieri/Kakaunui study, Component 2 focused entirely on mahinga kai values. However, rūnanga members on the Hakatere study identified the need to recognise cultural uses other than mahinga kai, and this was confirmed in the Tukituki study where it was observed that while mahinga kai is important for many sites, Component 2 needs to recognise sites that maintain other cultural uses. Thus, of the four parts making up Component 2 (refer to Section 3.1) the only refinement was to the fourth part, changing the focus from whether Māori would return and gather mahinga kai in the future to whether Māori would return and use the site in future.

In relation to assigned scores for the number of species present, concerns were raised at each stage about expressing the 1-5 score in relation to the maximum number of species at any site on the river in question. The problem was threefold. First, it may be inappropriate for a degraded site to score highly simply because it was the 'best of the worst'. Second, unlike other measures, this scoring system does not allow for comparisons between catchments. Third, such an approach to deriving a catchment-specific score may involve complex calculations. Despite these concerns, it was not considered appropriate to apply a generic 1-5 rating for a set number of species. Thus, to assign a score for the number of species present at a site, the steps are:

• collate the list of plant, fish and bird species present
• for all sites assessed, determine the maximum number present
• based on the maximum number of species present at any one site, use the table (Appendix 1) to assign a score of 1-5 based on how many species there are compared to this maximum figure.

3.4 The evolution of Component 3 of the CHI: cultural stream health measure

The cultural stream health measure (CSHM) was first developed for the Taieri and Kakaunui Rivers (that were combined because of their similar nature). To determine whether a different CSHM must be developed for every river or, alternatively, whether a generic CSHM can be used, the study was extended to the Hakatere River (a different river type) and the Tukituki River (a different iwi). In this section of the report, the CSHMs for the four rivers are compared and then the data from all four rivers are combined to derive and test a combined CSHM. Finally, a new generic CSHM is described. The development of a CSHM for each of the four rivers is documented in Tipa and Teirney (2003c [Taieri/Kakaunui], 2005a [Hakatere] and 2005b [Tukituki]).

The CSHM is calculated by averaging scores for a number of separate cultural indicators of stream health. The original 30 indicators identified during Stage 1 (Section 2.2, Table 1) were reduced to 19 after further interviews with kaumātua and others from within the Ngāi Tahu rohe. Some indicators were considered to be too subjective or unable to be clearly defined or readily measurable. Thus, only a subset was used. For the Taieri/Kakaunui and Hakatere catchments, the same recording form was used for all site assessments (Appendix 2). However, as a result of rūnanga interviews in the Tukituki study a slightly modified recording form was used (Appendix 3). The indicators included in the different studies are shown in Table 4.

Table 4: Indicators used in the different studies

Note: Some indicator descriptions were modified slightly in the Tukituki study (compare Appendices 2 and 3) and two indicators were added.

Rūnanga interviews revealed the similarity in the majority of indicators used by Māori to assess stream health and the emphasis on resource use, in particular mahinga kai. This similarity gave confidence that a CHI with a common set of indicators might be applied across the country.

3.4.1 The process

The process of deriving a CSHM for each river involved:

• identification of cultural stream health indicators by tangata whenua
• rating the importance of each indicator (on a 1-5 basis) by the tangata whenua team for selected sites along the river
• rating the overall stream health, a holistic subjective measure, of each stream site by the tangata whenua team
• identifying the relationships between each indicator and overall stream health using correlation coefficients
• setting aside indicators that are so highly correlated with overall stream health that they actually represent alternative measures of stream health rather than contributing factors
• for the remaining indicators, using multiple regression to identify those that best account for overall stream health (scores for this subset of indicators are then averaged to give the CSHM for each site)
• assessing the performance of the CSHM by comparing values for each site in relation to:
  • stream size (to ensure applicability of the measure across a range of stream sizes)
  • western stream health measures (MCI - Macroinvertebrate Community Index, SQMCI - Semi-Quantitative Macroinvertebrate Community Index)
  • land use (percentage of developed land in the catchment area of a stream site).

The process of deriving a combined CSHM involved:

• following the same steps as for individual rivers except that data from all 106 stream sites were first combined (Taieri/Kakaunui 35, Hakatere 30, and Tukituki 31)
• identifying indicators to be included in the combined CSHM using multiple regression on the combined data set
• applying the combined CSHM to each site on each river and assessing performance of the combined CSHM for individual rivers in relation to:
  • stream size
  • western stream health measures (MCI, SQMCI)
  • land use (percentage of developed land in the catchment area of a stream site).

3.4.2 Identification of alternative indicators of stream health

The indicators "would you eat fish", "would you go fishing" and "would you taste the water" were very highly correlated with overall stream health in all four rivers and for all data combined (Table 5). The consistency of responses across contrasting rivers by different rūnanga and iwi emphasises the fundamental importance of mahinga kai to a cultural evaluation of stream health by tangata whenua. The presence of food resources that would indicate stream health was highlighted by Ngāti Kahungunu, as was swimming, an important activity associated with the Tukituki River that is unlikely to be the case for colder South Island rivers.

Table 5: Correlations between the overall stream health measure and selected stream health indicators for individual rivers and all data combined

Note: The relationship between two variables can range from being perfectly correlated, correlation coefficient of 1.0, to not correlated, correlation coefficient of 0.0. For instance, a correlation of 0.97 between "would you eat fish" and the overall stream health measure for the Hakatere is exceptionally strong. The correlations between the indicators listed in table 5 and overall stream health are all highly significant.

Note: N/A - not applicable.

3.4.3 Identification of indicators that contribute to overall stream health

Two statistical methods were used to identify the indicators that contribute the most to cultural stream health, correlations and regressions. Both methods provide important insights into the make up of a generic cultural health stream measure.

Indicators evaluated by correlation coefficients

The indicators that are significantly correlated with the rūnanga assessments of overall stream health show remarkable consistency across the different rivers (Table 6). Water quality was always the most significant indicator of overall stream health, whilst water clarity, flow, riverbed condition/sediment and use of the riparian margin were also strongly correlated with overall stream health in each river. Other indicators showed some variation between rivers but even these were all positively and quite strongly related to overall stream health. These results indicate that a single set of stream health indicators may be appropriate to use in all rivers by different rūnanga and iwi. Given the consistency in pattern, it is not surprising that the combined data set also shows good levels of correlation between the indicators and overall stream health. An exception is river flow because of different patterns of flow in the contrasting rivers.

Table 6: Correlations between the overall stream health measure and stream health indicators for individual rivers and all data combined

Indicators evaluated by multiple regression

Multiple regression analysis determines the indicators that account best for variation in overall stream health. The total amounts of variation explained in the different rivers was 92.6 percent, 88.6 percent and 76.0 percent, and for all rivers combined 71.8 percent (Table 7). All these values are remarkably high and give confidence that, in each case, a set of indicators can effectively encapsulate what rūnanga members assess to be overall stream health. Not unexpectedly, water quality is of key importance in all river data sets. Flow, catchment land use and channel modification contribute to the results for two rivers whereas condition/sediment, water clarity and use of the riparian margin feature only once. River bank condition, use of the river (takes and discharges), indigenous species and riparian vegetation did not feature in the multiple regressions. In the combined data set, water quality, riparian vegetation, flow and bed condition/sediment were the factors of importance. Thus, these are the indicators whose scores will be averaged to produce the combined CSHM to be discussed in Sections 3.4.5 and 3.4.6.

Table 7: Regressions to identify the indicators that account best for variation in overall stream health of separate rivers and for all rivers combined

View Table 7: Regressions to identify the indicators that account best for variation in overall stream health of separate rivers and for all rivers combined

3.4.4 Comparing CSHMs for individual rivers with other stream health measures

In our previous reports, the CSHMs calculated for each site were compared with two existing measures of stream health, the macroinvertebrate community index (MCI) and its semi quantitative variant (SQMCI), both used extensively by researchers and water managers in New Zealand.

We also compared the CSHMs for each site with the percentage of developed land in the catchment area of the site. There were always significant positive correlations between CSHM and the invertebrate stream health measures (pollution sensitive invertebrates being less common where the CSHM was low) (see bold values in Table 8). Moreover, there was always a negative relationship between CSHM and percentage of developed land (where more of the catchment area of a site is developed, the CSHM was low).

These results are noteworthy in that they indicate the CSHM, like the other stream health measures, successfully captures aspects of stream health. Furthermore, the relationship is particularly noteworthy given MCI and SQMCI are based on measures of stream invertebrates whereas the CSHM evaluates stream health from a Māori perspective, based on a combination of catchment, river margin and in-stream characteristics.

3.4.5 Are the CSHMs from individual rivers interchangeable?

To compare the CSHMs of individual rivers, an independent measure is needed to which they all relate. As indicated in Section 3.3, MCI, SQMCI and percentage of developed land within the catchment have been shown to be significantly related to individual CSHMs (values in bold in Table 8). The other cells in Table 8 show what happens when the set of indicators used to calculate the CSHM in one river (eg Hakatere) are applied to the dataset from another case (eg the Taieri/Kakaunui - column 2 of row 1). In this case, even using the Hakatere CSHM indicator set, the new CSHMs for sites on the Taieri/Kakaunui are still very strongly correlated with MCI, SQMCI and percentage of developed land in the catchment. The same pattern is seen in every case (although relationships with percentage of developed land are less obvious when Tukituki data are involved).

The general conclusion is obvious and important: the precise set of indicators used in the CSHM is not critical to the assessment of river health. Indeed, when the combined CSHM is used to recalculate health in all sites in each river, there remain highly significant correlations with MCI (0.54***), SQMCI (0.54***) and percentage of developed land -0.44***). Accordingly, cultural stream health can be assessed using a generic group of indicators rather than different sets of indicators for individual rivers.

Table 8: Correlations of CSHMS with MCI, SQMCI and percentage of developed land, in all possible combinations of river data sets

*** < 0.001 p value - very highly significant

** < 0.01 p value - highly significant

* <0.05 p value - significant

Features of the analysis using all river data sets combined

• The combined data sets show strong correlations between overall stream health and mahinga kai cultural activities, further reinforcing the interpretation that these indicators are functionally equivalent to overall stream health (Table 5).
• Correlations between overall stream health and influential indicators reinforce the conclusion of a consistently important set of contributors to overall stream health revealed for each river (Table 6). The final choice of indicators must also take into account whether each can be assessed consistently.
• The multiple regression analysis further highlights water quality, riparian vegetation, flow and riverbed condition/sediment as major contributors to stream health (Table 7).
• Correlations between the combined CSHM and MCI, SQMCI and percentage of developed land in the catchment are all highly significant. This is an important result given the cultural stream health measure is based on cultural perceptions of aspects of the entire catchment whereas the established indicators are based on the macroinvertebrates inhabiting a stream site - very different measures.

3.4.6 A generic cultural stream health measure

We have considered using the combined CSHM, calculated according to the results of multiple regression analysis, as a generic CSHM for use in any river by any rūnanga or iwi. However, given the consistency in patterns among indicators that correlate with overall stream health, we recognised there is considerable leeway in defining a generic CSHM. Accordingly, we decided not to restrict our generic CSHM to only the four indicators from the combined analysis in Table 7. Instead, we selected eight of the indicators shown in Table 6. Three were not included for the following reasons:

• 'riverbank condition' was not sufficiently highly correlated with overall stream health to be included
• 'indigenous plant species' was highly correlated with riparian vegetation and the latter indicator was more highly correlated with overall stream health
• 'use of the river (takes and discharges)' was not included because of difficulties experienced by rūnanga members when evaluating the indicator. Whereas the indicator referred to takes and discharges that could be seen, local knowledge could influence the rating.

As different rivers were added in the course of the study, the definition of flow was refined but never satisfactorily reflected the aspect of stream health being sought. Variation in flow was a way of describing variability in stream habitats (slow to fast water, still to white water, pools, runs and rapids). Finally, in the Tukituki study, 'variety of habitats present' was added as a more satisfactory indicator and demonstrated a correlation of 0.75 with overall stream health. Accordingly, flow (visible) is replaced by variety of habitats in the generic CSHM.

With these adjustments, the generic CSHM is made up of the indicators listed in Table 9. The generic CSHM thus combines evaluations of eight indicators that cover catchment, riparian and in-stream factors. That a measure of cultural stream health is made up of indicators from the top of the catchment down to and into the river reflects the holistic ki uta ki tai (mountains to the sea) philosophy that is fundamental to tangata whenua kaitiakitanga.

Table 9: Indicators selected for the generic CSHM

Note: Correlations with overall stream health in the combined datasets are indicated.