View all publications

8 Hauraki Gulf

8.1 Classification definition decisions

8.1.1 Pilot classifications

The ranking developed by the validation analyses was used to subjectively select a reduced set of variables for subsequent development of pilot classifications (Snelder et al. 2004). In order to simplify the classification, variables ranked lower than 9 in Table 6, except freshwater fraction, were excluded because they made only very small contributions to the statistical models. Although freshwater fraction was also ranked very low by the validation analysis, it was considered that it is likely to be important and the reasons for its low ranking in the validation analysis may have been associated with a lack of data representing areas with high values of freshwater fraction. We have a high level of confidence in the method used to derive the freshwater fraction and, therefore, considered that it should be included in the classification. Thus, the following variables were excluded: seabed rate of change of slope (profile), seabed curvature, seabed planform curvature, mean annual SST.

One of each pair of highly correlated (r > 0.95) variables was removed for the same reasons as outlined for the EEZ variables. Thus, SST annual amplitude was excluded because it was highly correlated with SST annual phase (r = 0.97) and because SST annual phase was the higher ranking of the two in the validation analysis. In addition, extreme orbital velocity was removed because it was highly correlated with mean orbital velocity (r = 0.96) and was ranked lower in the validation analysis. In addition, for the same reasons as outlined for the EEZ scale classification, we concluded that the existing sediment data layer is too coarse and that, until there is a better source of data, sediment should be excluded from the classification.

Various pilot classifications of the Hauraki Gulf were developed based on the following eight variable: depth, slope, tidal current, freshwater fraction, mean orbital velocity, SST annual phase, SST monthly standard deviation and SST semi-annual amplitude (Snelder et al. 2004). In addition, Snelder et al. (2004) suggested transformations and weighting of some variables in the definition of the pilot classifications based on subjective decisions that were guided by inspection of the mapped classification.

8.1.2 Tuning

Leathwick et al. (2004) used similar analyses, based on Mantel tests, to those used to tune the EEZ classification to help tune the definition of the Hauraki classification. The analyses performed by Leathwick et al. (2004) indicated that transformation and weighting could do little to improve correlation of environmental and biological space for most datasets. A log transformation of depth, which makes intuitive sense, improved correlation for two datasets (fish and pelagic) but decreased correlation for the core (benthic) dataset. A subjective decision was, therefore, made not to transform depth. However, it was decided that tidal current and mean orbital velocity should be transformed. This decision was supported mainly by inspection of the pilot classifications (Snelder et al. 2004) which indicated that some compression of tidal current and mean orbital velocity improved the definition of environments. The final decisions for definition of the Hauraki classification are shown in Table 9.

Table 9: The variables and transformations used to define the Hauraki classification

Variable

Transformation

Depth

 

Freshwater fraction

 

SST annual phase

 

SST monthly standard deviation

 

SST semi-annual amplitude

 

Mean orbital velocity

Log10

Tidal current

Cube root

Slope

 

8.2 Classification

Table 10 shows the within-class average value of each of the variables used to define the Hauraki Gulf classification. Each class is labelled by a number which has no specific meaning but is associated with the order in which groups of cells are agglomerated by the clustering procedure. Inspection of this table indicates that classes are distinctive from one another with respect to at least one variable. Table 10 also shows how the classification has differentiated environmental variation at the 2, 4, 6 and 9-class levels. The division at the two-class level (bold line on Table 10) predominantly subdivides the inner gulf from the mid to outer gulf (see Figure 18). Within the inner and mid to outer gulf environments further divisions occur at the four and six-class levels (thin solid and dashed lines on Table 10). These subdivisions are predominantly associated with differences in depth and separate the coastal and deeper environments (see Figure 18).

Table 10: Average value of each of the eight defining environmental variables in each class of the 20-class level of the Hauraki Gulf classification

View average value of each of the eight defining environmental variables in each class of the 20-class level of the Hauraki Gulf classification (large table)

Figure 18: Classification of the Hauraki Gulf mapped at the 2, 4, 6 and 11-class levels

Maps of the Hauraki Gulf modelling environmental classes (differentiated by colour) at four different levels of classification: 2, 4, 6 and 11 classes.

The nine-class level differentiates the coastal areas of the inner gulf into those areas with high freshwater fraction (Firth of Thames) and similarly shallow areas with lower freshwater influence (South-Eastern Bays). The nine-class level also further subdivides the deeper coastal environments of the mid and outer gulf. Environments with high tidal currents and steep (probably rocky) seabed are differentiated (see Figure 18).

Figure 19 shows the Hauraki Gulf classification at the 20-class level. Figure 20 shows the classification using the continuously varying colour scheme that reflects environmental distances between the maximum number of classes defined by the classification (i.e. 280 classes).

The relationships between classes are described in greater detail by the dendrogram shown in Appendix 1, Figure A1.2. The dendrogram shows how the classes are progressively amalgamated to form a single large group. Note that the class numbers are assigned during the clustering procedure and are derived from the order in which amalgamation of the groups occur.

8.3 Classification strength of the Hauraki Gulf classification

A full description of the biological testing of the Hauraki Gulf classification is contained within Leathwick et al. (2004). Because large parts of the environmental domain were not represented by the biological datasets, not all the classes that are defined at any given level of the Hauraki Gulf classification could be tested. ANOSIM analyses were performed on classes at each level of the classification provided classes had at least four biological samples. The testing was limited because of uneven distribution of biological sample points across the classes. Thus for the fish dataset only seven classes could be tested at the 20-class level and only 11 groups at a 50-class level of classification. A new invertebrate dataset comprising 50 sample sites was used to test the classification. Three classes from this dataset could be tested at the 20-class level and seven classes had adequate biological data at a 50-class level. An ANOSIM analysis was attempted using a pelagic dataset containing 34 sample points. However, even when the minimum number of sites per class was reduced to three, only two classes had sufficient biological samples at a 20-class level of classification and biological differences between these two environments were non-significant.

Results of the ANOSIM analysis of classification strength for the fish dataset showed the r-values initially rose sharply with progression from a two-class to a five-class level of the classification, but beyond this remained relatively invariant with increasing numbers of classes. At the 20-class level, all individual pair-wise comparisons between classes were significantly different (p < 0.01) in their biological composition, indicating that all classes that were distinctive with respect to their fish assemblage.

Figure 19: Hauraki Gulf classification at the 20-class level

A map of the Hauraki Gulf modelling 20 environmental classes (differentiated by colour).

Figure 20: Classification of the Hauraki Gulf using the continuous colour scheme based on the principal components of the eight variables used to define the classification

A map of the Hauraki Gulf displays a colour scheme based on the results of principal component analysis on the 280 classes generated by non-hierarchical classification. Each class was assigned varying levels of red, green or blue colour based on the PCA analysis. Bluer areas are deeper, with lower mean orbital velocity. Redder areas have higher values of tidal current. Greener areas are associated with higher freshwater fraction and lower SST phase.

Note: Bluer areas are deeper, with lower mean orbital velocity. Redder areas have higher values of tidal current. Greener areas are associated with higher freshwater fraction and lower SST phase.

Results from the ANOSIM analysis of the invertebrate dataset indicate a steady increase in r-values for successive levels of the classification. This indicates that the strength of the classification for invertebrates increases at lower levels. However, examination of biological similarities between the three classes with adequate data at the 20-class level indicated that the groups are not biologically distinguishable from each other (i.e. p > 0.1).

8.4 Biological characteristics of Hauraki Gulf classes

Descriptions of the biological character could only be produced for those classes with adequate biological samples. Information about fish assemblages at the 20-class level was available for seven environments while information about invertebrates was only available for three environments. In the following descriptions, classes are ordered according to the dendrogram (Appendix 1, Figure A1.2) rather than in strict numerical order so that closely related classes are more likely to be located in proximity to each other. Tables showing the frequency of occurrence of various fish and benthic invertebrates species are appended (Appendix 3, Figure A3.1 and Appendix 3, Figure A3.2).

8.4.1 Deeper water classes of the middle to outer gulf

A map of the Hauraki Gulf showing the distribution of class 6 of the 20 class level Hauraki Gulf Marine Environment Classification.

Class 6 - has the highest average depths (mean = 112 m) and occurs mostly north of Great Barrier Island. Commonly occurring species caught in trawls (occurrence > 50%) include snapper, red gurnard, john dory, scaly gurnard, leather jacket and arrow squid.

A map of the Hauraki Gulf showing the distribution of class 5 of the 20 class level Hauraki Gulf Marine Environment Classification.

Class 5 - occurs in moderately deep water (mean = 65 m) from Great Barrier Island west to Bream Head. Fish species occurring commonly in this class are snapper, red gurnard, john dory and leather jacket - scaly gurnard are less common than in the previous class while skates are more common.

A map of the Hauraki Gulf showing the distribution of class 4 of the 20 class level Hauraki Gulf Marine Environment Classification.

Class 4 - is the most extensive class at this classification level, occurring in water of moderate depth (mean = 45 m) south from about Little Barrier Island to occupy much of the Colville Channel and the mid gulf south to about Waiheke Island. Snapper, red gurnard and john dory are the most commonly occurring species, with moderate occurrences of leatherjacket, arrow squid and sand flounder. Brittle stars are by far the most commonly occurring species recorded from the invertebrate dataset (six sites).

8.4.2 Shallower water classes of the inner gulf

A map of the Hauraki Gulf showing the distribution of class 1 of the 20 class level Hauraki Gulf Marine Environment Classification.

Class 1 - occurs in inshore waters of intermediate depth (mean = 21 m) from about Bream Bay south to Ponui Island, much of it lying to the north of Waiheke and Rangitoto Islands. Snapper, red gurnard and john dory are the most commonly caught fish species with sand flounder occurring in approximately half of the trawls. Brittle stars are the most commonly occurring species recorded from the invertebrate dataset (26 sites).

A map of the Hauraki Gulf showing the distribution of class 196 of the 20 class level Hauraki Gulf Marine Environment Classification.

Class 196 - occurs at the northern end of the Firth of Thames, occupying sites of similar depth (mean = 24 m) to the previous class, but with higher tidal currents. Characteristic fish species include snapper, red gurnard, john dory and sand flounder, with moderately frequent catches of yellow-belly flounder, spotted stargazer, rig and barracouta.

A map of the Hauraki Gulf showing the distribution of class 205 of the 20 class level Hauraki Gulf Marine Environment Classification.

Class 205 - occurs mostly in protected, shallow waters (mean = 6 m) between the mainland North Island and Rangitoto and Waiheke Islands. Snapper are the most frequently caught species in trawls, followed by john dory, spotty, trevally and kahawai. The most commonly occurring benthic invertebrates are Helice crassa, Lumbrinerid spp., Siglanoidea and brittle stars (16 sites).

A map of the Hauraki Gulf showing the distribution of class 254 of the 20 class level Hauraki Gulf Marine Environment Classification.

Class 254 - occurs in the southern half of the Firth of Thames where average water depths are shallow (mean = 5 m) but tidal currents are moderately strong. Snapper are again the most common species caught in trawls along with red gurnard, rig, rays, sand and yellow-bellied flounder, kahawai and yellow-eyed mullet.