Publications: Research reports and publications

Instream habitat modelling and its application in the Motueka Catchment

1 January, 2007
Cawthron Report 1073. Prepared for Integrated Catchment Management for the Motueka River.


Management of water allocation is a major issue throughout New Zealand and involves consideration of instream flow needs along with the potential benefits associated with out-of-stream water uses. Managers are faced with the question of how much water can be abstracted from a river without causing significant harm to the river ecosystem, or affecting the cultural, aesthetic and recreational values associated with that river system.

A range of methods are available to help guide decisions on flow management. This report provides an overview of the available methods and then compares the output from two methods (1-dimensional habitat modelling and 2-dimensional habitat modelling) in a short reach on the Motupiko River. Two dimensional modelling is then applied to a reach on the Motueka River upstream of Tapawera to provide guidance on an appropriate minimum flow for this reach of the river.

Methods to help assess how much flow ought to be left in rivers can be divided into three general types: historic methods, hydraulic methods and habitat methods. Historic methods are the simplest and easiest to apply, but are probably most appropriate for river systems where the linkages between ecosystem integrity and flow requirements are poorly understood. Hydraulic methods require some field survey work and predict how depth, velocity and river width will change with flow. Habitat methods are the most sophisticated means to quantitatively assess the instream flow requirements of rivers and relate changes in depths and velocities at different flows with the habitat requirements of particular species. However, these methods require intensive field surveys and still attract controversy regarding the scientific defensibility of the habitat suitability criteria and model outputs. Some new methods involving models based on energetics concepts have significant potential in the near future, but also require intensive field surveys.

One-dimensional (1D) habitat methods are based on field surveys of river cross-sections, while two dimensional (2D) habitat methods are based on detailed surveys of riverbed topography throughout specified reaches. Both methods require a rating curve to determine how the water surface elevation in at least part of the reach will change with flow. Within the range of the rating curve, 1D models are easier to calibrate than 2D models. However 2D models give a better measure of the longitudinal variation in depth and velocity than 1D models and probably predict changes in velocity distribution outside the calibration range more accurately than 1D models. 2D models are also able to predict how braiding patterns will change with flow in braided rivers.

A comparison of 1D and 2D habitat modelling was conducted in a short reach in the Motupiko River to demonstrate the processes involved and the type of outputs from each approach. The surveys involved were not as extensive as would be required for the actual application of these techniques, so the results should be interpreted with caution and not used as the basis of flow management decisions. The 2D approach has the advantage that the distribution of suitable habitat throughout a particular reach can be plotted in a plan view (map of the river reach) and the changes related to flow clearly shown. However, it is important to use an appropriate metric of habitat suitability when producing these plots. Due to differences in computational mesh density within the 2D model in different parts of the reach, it is not appropriate to plot plan views of area weighted habitat suitability (WUA). Unweighted suitability scores should be used in such plots.

In general the response of predicted habitat suitability to flow for most species and life stages was reasonably comparable between the 1D and 2D approaches, although there were some differences in the flows that were predicted to result in peak habitat availability for juvenile brown trout using one set of habitat suitability criteria. Binary suitability criteria that simply distinguish between suitable and unsuitable habitat were also trialled. The advantage of using this binary approach is that the area weighted sum of its values actually represents an area of suitable habitat, rather than just a dimensionless index. However, in this reach and using the range of habitat suitability criteria that were applied, the binary approach essentially reduced the habitat method to an index of wetted area rather than an index of quality habitat. A binary threshold approach was also trialled where only habitat scoring above a certain threshold was considered suitable. The remaining habitat was considered unsuitable. The binary threshold approach provides an indication of how the area of optimum habitat changes with flow, however there is a degree of subjectivity in deciding what threshold level to use in distinguishing between suitable and unsuitable habitat. The other major issue with applying these binary systems is that the suitability criteria that are available have not been developed with this type of binary suitability in mind.

The 2D habitat modelling approach was applied to a 400 m reach in the Motueka River upstream of Tapawera with the aim of providing guidance on a minimum flow requirement for this section of the river. There are substantial losses of surface water to the aquifer in this reach of the river and thus this reach was considered to represent the area that will first experience the impacts of low flow. Flow management decisions need to consider the critical values of the reach of interest and provide flows capable of retaining a percentage of the habitat for the critical value that would be available at the mean annual low flow (MALF). The concept of critical values is based on the premise that if sufficient flow is provided to sustain the most flow sensitive, important value, then the other significant values will also be sustained. Using adult brown trout as the critical value in this reach of the river, a flow of 1.2 m 3s-1 is predicted to maintain 90% of the habitat available at the MALF.However, it is generally recognised that minimum flows must be set in conjunction with appropriate allocation rules to ensure that a degree of the natural flow variability is maintained. We suggest that a factor that could be considered in this process is to ensure that the invertebrate habitat at the median flow is not reduced excessively by water allocation. This would provide a biological rationale for the level of allocation in addition to that underpinning the setting of the minimum flow. Invertebrate habitat at the median flow is relevant to maintenance of the productivity of invertebrate populations, which provide the food base for fish.

All of these modelling efforts are based at a reach scale and therefore the information is really only applicable to the parts of the Motueka River Catchment where the field surveys and modelling have been conducted. Ideally, it would be useful to look at flow management in the catchment as a whole, rather than on a reach by reach basis. Over the next few years we will be attempting to scale-up some of this reach scale information to better understand how habitat availability changes with flow throughout the catchment.