Several aspects of G. maculatus spawning site ecology are potential sources of spatiotemporal variation. The reported relationship with salinity results in horizontal structuring along the axis of waterway channels in relation to saltwater intrusion (Richardson and Taylor, 2002; Taylor, 2002). This may drive variability in the position of spawning reaches on a catchment scale when coupled with dynamism of river discharges and tidal forcing. Despite that previous studies have highlighted the use of the same spawning sites for multiple years (Taylor, 2002), this case was characterized by habitat shift in both catchments in comparison to all known records. Although the potential effects of salinity changes have seldom been highlighted in the literature, this study indicates that they may important in relation to perturbations from extreme events or to incremental changes, such as sea level rise. However, a lack of historical salinity data for the reaches of interest makes the degree of variation difficult to confirm directly in our study area, and this is generally case elsewhere.

In addition, the timing of spawning on or soon after the peak of the tide, combined with a preference for shallow water depths, leads to vertical structuring of the habitat in relation to water level heights (Benzie, 1968a; Mitchell and Eldon, 1991). Interaction between the waterline and floodplain topography also influences the distance between spawning sites and the alignment of (i.e., perpendicular to) waterway channels. This variation may be considered where the topography is relatively flat and is a further consideration for effective PA design.


There are at least three aspects of this study that are likely to be applicable to the design and evaluation of Category IV PAs elsewhere. They include the question of PA boundary setting in relation to the habitat to be protected, the need for data to inform this and monitoring strategies to support future evaluations, and practical considerations for identifying boundaries on the ground as required by stakeholders.

Clearly, accuracy is important when setting boundaries for Category IV PAs, yet spatiotemporal variation may hamper acquisition of the necessary data in practice. For G. maculatus strong temporal trends are a particular consideration. Variation has been reported in relation to the peak days of activity within a tidal sequence, the tidal sequences preferred in different parts of the country and months of most spawning activity in the year (Taylor, 2002). International studies have also reported large-scale variation in traits associated with spawning (Barbee et al., 2011). In combination, these aspects suggest that spatiotemporal variability could arise at multiple scales creating practical difficulties for both empirical data collection and model-based approaches for determining habitat distribution. In this case, the study catclunents are New Zealand’s best-studied spawning areas yet surveys have only been periodic and seldom comprised more than one month in any given year (Taylor, 2002). Consequently, the times of peak spawning activity may not have been captured in the survey record. Identification of the spawning distribution has, therefore, relied on the compilation of multi-year data despite the potential for confounding factors associated with both short- and long-term change.

Albeit that the postearthquake context represents a major perturbation, the impacts of spatiotemporal variance on PA effectiveness are clearly seen in planning methods 1 and 2. These methods were developed using the planning authority’s up to date infonnation on spawning habitat in both catclunents. Particularly in the Heathcote, earthquake-induced habitat shift rendered these methods relatively ineffective. Despite this, regular monitoring and amendment of the same protection mechanism could provide a strategy for maintaining effectiveness and addressing change. However, for method 1 the data collection requirements would be onerous to achieve this in practice. This partly reflects reliance on a network of small PAs but also that the detection of spawning sites is difficult (Orchard and Hickford, 2018). The number of PAs identified appears woefully inadequate in light of the postquake data, yet fairly represents results of the monitoring effort that was in place prequake. Increasing this to the level of a census-survey for peak spawning months represents a considerably scaling-up of the monitoring program.

In comparison, method 3 was based on considerably larger PAs and was much more resilient to earthquake change. In that case, a degree of redundancy was seen as a desirable aspect of resilience (Greer et al., 2015). However, from the perspective of PA evaluation, the thr ee PA mechanisms share similar monitoring requirements. This arises since the demonstration of PA management effectiveness requires information on the values to be protected (Stoll-Kleemann, 2010). Given that monitoring resources are inevitably limited, dynamic environments demand particular attention. In turn, this illustrates the need for research on monitoring strategies to inform priorities for data collection and frequency (Teder et al., 2007). Moreover, it exemplifies the need for more management-driven science to close the gap between conservation policy and practice (Knight et al., 2008).

Potential strategies include using abiotic proxies for conservation objectives for which data acquisition is easier thus reducing the burden of repeat measurement (Lawler et al., 2015). Method 3 provides an example of this approach, using a predictive model based on elevation above sea level (Greer et al., 2015). However, the results indicate that its efficiency as a planning method is relatively low since much of the area set aside did not help achieve the stated objectives and it could not be used as a proxy for outcomes monitoring against the relevant policy objectives. From an ecosystem-based perspective, inefficient planning methods may also hinder other potential uses of the areas involved, leading to unnecessary trade-offs (Southworth et al., 2006). The practical aspects of this relate to the rules that apply within the PA and are designed to confer protection. Where a degree of sustainable use is envisaged within PAs, the specific arrangements for management need to be well matched to intended objectives.

Efficiency may be a particularly important consideration for Category IV PA evaluation where the management context is characterized by high land-use intensity in adjacent areas (Dudley, 2008). In this case method 2 offered an alternative approach that identified the known EOO and additional areas considered be ‘potential’ habitat and included these in the areas delineated for protection (Margetts, 2016). Essentially this created a buffer around the mapped EOO that served to address limitations in the information available for quantifying known habitat, as well as a providing a degree of redundancy, to improve resilience. Although in the Heathcote the post-quake habitat was found to have shifted outside of these areas, they were effective in accommodating the smaller magnitude of change observed in the Avon (Fig. 21.2). Management effectiveness evaluation of methods 2 and 3 primarily requires information on EOO as could be obtained by regular census-surveys of spawning habitat (Orchard and Hickford, 2018). The combination of an evaluation-informed adaptive approach and degree of redundancy could offer an effective and efficient PA strategy for the dynamic habitat with regards to land use allocation.

Lastly, this case highlights some practical issues for the visualization of PA boundaries. In our evaluation, spatial co-occurrence was based on coordinates describing the upstream and downstream extent of spawning sites and polygons describing PAs. In many instances, the spawning site locations were very close to the PA boundaries as mapped. Unless they were clearly outside of the boundaries, such sites were assessed as being protected with the result being an optimistic view of the spatial coverage of the PA mechanism. In reality these boundaries may not be so clear. However, it is important that they are clear for the benefit of all stakeholders (Langhammer et al., 2007), and this depends considerably on design and communication of the planning methods. In this case the areas delineated by method 1 were interpreted by stakeholders using a location description and schedule of coordinates (Table 21.2). This is considered to offer a relatively clear mechanism for implementation of the PA requirements in practice.

Under method 2, the areas for protection were first visualized as lines in Council planning documents (Margetts, 2016) and then subsequently incorporated into ‘Sites of Ecological Significance’ (SESs) in a recent statutory plan (Christchurch City Council 2015), which is now operative. The visualization method for plan users is a set of polygons annotated on planning maps appended to the plan (Fig. 21.Sla). These SESs have, therefore, become the PAs of interest and method 2 (as assessed in this study) can be interpreted in relation to G. maculatus objectives within these larger areas. However, at the scale of the mapping provided it is difficult to see exactly where the PA boundaries lie in the riparian zone, requiring considerable guesswork by plan users (Fig. 21.Sib).

Under method 3 the situation is unproved by the provision of PA polygons as a public dataset with an online GIS viewer available, in addition to planning maps appended to the relevant plan (Environment Canterbury, 2017). Nonetheless, similar boundary issues arise with regards to the exact location of the PA in relation to the spatial extent of spawning habitat. The GIS analysis revealed a few spawning sites that were clearly outside of the PA boundary in the Avon, as reflected in effectiveness results of <100% in both years (Table 21.5) and, in general, many of the actual spawning locations were again very close to the PA boundaiy. Furthermore, the habitat may shift a considerable distance from the low flow channel on high water spawning events, and these circumstances are difficult to detect by operators (e.g., management contractors) in the field. Indeed spawning sites were found to have been destroyed by the City Council’s own reserve management contractors subsequent to notification of the relevant statutory plan (Orchard et al., 2018). This suggests that better guidance materials, such as interactive maps, may be required to improve PA effectiveness in practice, as was recommended in a recent management trial that aimed to avoid such damage to spawning sites (Orchard, 2017). These results also indicate that a buffer should be considered as an aspect of PA design.


Several assumptions have been made in this evaluation consistent with a focus of the protection of dynamic habitats and the objective of identifying

S1 Planning maps showing Sites of Ecological Significance

FIGURE 21.S1 Planning maps showing Sites of Ecological Significance (SESs) in the Christchurch City area (Christchurch City Council 2015). (a) Schedule Reference Map. (b) Example of detailed planning map. No enlargements are provided for SESs in riparian zones. For brevity, only an excerpt of the full legend is shown.

learning from the unique postearthquake situation. Most importantly, the focus has been restricted to the spatial basis of protection mechanisms for critical habitat as found in planning documents. In all cases, they were assumed to confer protection where spatial overlap occurred. In reality, this also depends considerably on the design of the rules that apply within the PA and aspects such as the provision of compliance monitoring. Also, a conservative approach has been taken in the mapping of PA boundaries and protection assumed to be effective. In the case of method 2, the width of the riparian zone protected could not be accurately identified and all spawning sites with the protected reach were assumed to be covered. Other limitations of the study include the spatial coverage of postquake surveys in relation to method 3 since the full extent of those PAs was not directly surveyed. Despite this the spatial coverage of the surveys was extensive in both catchments and the methodology was designed to capture the upstream and downstream extents of the full habitat distribution (Orchard and Hick- ford, 2018). Different evaluation results can also be expected in light of new information. In particular, the number of spawning events captured in the postquake survey record is limited. Further spatiotemporal variation may arise from effects, such as differing water heights outside of the sampled range, future vegetation change, river engineering impacts, the potential for further ground-level changes, and the ongoing influence of sea-level rise.

< Prev   CONTENTS   Source   Next >