Table 2.7: Pros and cons of flow assessment methods

Historical flow method 1

Proportion of recorded or estimated flows (eg, retain at least 90% of natural flow). Can be adjusted seasonally.

Quick and easy, uses existing data, results in flow variability without going to detailed level of analysis. Some abstraction allowed during times of low flow.

Historical flow method 2

Minimum flow based on a proportion of a flow statistic (eg, minimum flow 90% of seven-day mean annual low flow (MALF)). Can be adjusted seasonally.

Expert panel

Panel of experts to advise on flow requirements based on bankside inspection.

Quick, cheap, has credibility (dependent on experts), useful political tool to help overcome mistrust if well managed. Can be used to support other methods.

Not predictive, it is difficult for experts to determine how character of river changes with flow. Consensus can lead to poor ecological outcomes.

Range of variability approach (RVA)

Based on flows not exceeding the natural range of variation of a number of hydrological parameters that characterise minimum flows and flow variability. The RVA method recommends that flows be maintained within flow-management targets at the same frequency that would have occurred naturally. Flow-management targets can be within percentiles of the natural range or within standard deviations The RVA does not recommend maintaining flows exclusively at or near the level of the lower limit.

Allows interim flow targets and river management strategies to be developed where there is long-term hydrological data. Requires no field work or ecological evaluation. A conservative method suitable for rivers for which the objective is to keep them in their natural state.

Tends to be conservative in its allocation. More difficult to apply to systems with no historical flow records. Natural variation in parameters has not been related to biology, water quality or geomorphology and the method relies solely on maintaining the natural flow hydrograph. As the method stands, it is not clear how it could be applied to flow abstractions. May not be applicable where the sediment regime has also been altered. Not yet used in New Zealand for setting ecological flows.

Generalised habitat models

Describes relationship between habitat and flow, simplified versions of detailed 1D habitat models.

Don’t require full instream habitat surveys, could be used more widely.

Models lack information that could be gathered using a full 1D habitat survey; not as precise, relatively new technique, some restrictions to stream types they can be applied to (eg, braided rivers, spring fed streams).

Periphyton biomass model

Prediction of maximum biomass as a function of time between floods and nutrient concentrations.

Most relevant for rivers with major impoundments or large diversions where there is opportunity to manage flood flows.

Need accurate estimates of average soluble nutrient loads, needs more validation of accuracy of predictions.

1D hydraulic model and habitat evaluation

Predicts water depth, velocity, and habitat suitability as a function of flow.

Widely used and understood, relatively easy modelling, gives a specific relationship, most closely links hydraulic habitat availability with a range of flows. Co‑operative (multi-agency) studies can lead to better buy-in of results.

Interpretation of results is variable, modelling can be applied without consideration of biology and context. Some uncertainty in use of habitat suitability curves as predictors – particularly around invertebrates. Habitat analyses rely on good information on habitat suitability.

2D hydraulic model and habitat evaluation

Predicts water depth, velocity, and habitat suitability as a function of flow.

When working outside boundaries of current wetted channel, extrapolating beyond calibration data provides good 2D graphics for visualisation of predictions. Can extrapolate further than 1D model, especially suitable for braided rivers and rivers where flow patterns change significantly with flow (eg, overbank flows).

Requires significant and expert data inputs and analysis, difficult and expensive to apply on shallow boulder rivers. Interpretation of results is variable, modelling can be applied without consideration of biology and context. Habitat analyses rely on good information on habitat suitability, and limitations of habitat model are not always understood. Some uncertainty in use of habitat suitability curves as predictors – particularly around invertebrates.

Connectivity/fish passage assessment

Habitat method applied in a critical reach, identified by survey. See 1D and 2D hydraulic models above.

Addresses specific issue at specific downstream locations.

Need to survey entire river to identify critical reach for modelling, requires significant field work input, biological interpretation can be difficult, don’t know what length of time is sufficient for fish passage nor how length of critical reach (eg, critical riffle) interacts with critical passage depth.

Entrainment model

Predicts water levels for high flows. Critical flows for moving bed sediments based on 1D or 2D hydraulic models (above).

Most relevant for rivers with major impoundments or large diversions where there is opportunity to manage flows, provides flushing flow requirements, stability of habitat, important base for ecosystem analysis.

Need accurate estimates of bed sediment size, internationally science is not very accurate and still evolving, not a precise science, needs more validation of accuracy of predictions.

Quite specialised, requiring hydraulic expertise. Validation important but rarely carried out.

Sediment transport

See entrainment model

Bank stability

See entrainment model

Specific assessment of bank stability or erosion complex and difficult.

Suspended sediment

Predicts how downstream sediment concentrations vary with flow. Particle settling model using 1D hydraulic model above.

May need suspended sediment measurements for calibration and verification.

Seston flux

Applied in special circumstances, dispersion settling model used in lake outlets to determine how flow change will affect seston distribution. See 1D hydraulic model above.

Specifically takes account of lake outlet circumstances, which can support highly valued fisheries, demonstrates value of water itself (ie, it carries in suspension plankton etc, which is utilised by filter feeding invertebrates, which in turn are food of fishes).

May need biological data for calibration and verification.

Inundation model

Predicts area of inundation and time, in relation to flood flows. See entrainment model above.

Identify critical bands of inundation flows/levels, allows protection of wetland habitat (interpretation is difficult), protects property.

Data requirements are extensive (surface roughness, survey), quite specialised, biological interpretation is difficult.

Flow variation analysis

Identification of critical features of flow variability to maintain ecosystem functions, often based on results of other models (eg, entrainment, fish passage).

Allows for broad ecosystem requirements that may not be picked up in specific habitat analysis.

Science is still evolving, difficult to set flow thresholds except for some ecosystem functions (eg, river mouth opening for fish recruitment).

Fish bioenergetics model

Spatially explicit predictions of net rate of energy intake fish growth potential fish positions, and overall carrying capacity of a function of flow. Uses 1D representative reach or 2D hydraulic models (see above).

Provide biologically meaningful predictions that user groups can relate to (good 2D graphics for visualising predictions), outputs useful educationally, and quantitatively links hydraulics, invertebrate drift (fish food) and salmonid foraging processes.

Complicated and expensive. Available for only one species (brown trout), but could be parameterised for other salmonids.

Water quality models

Includes temperature and dissolved oxygen. Uses generalised habitat or 1D hydraulic model above.

Requires some data and links flow to critical parameters (temperature and dissolved oxygen). Application is relatively simple (eg, WAIORA).

Complicated to calibrate model, requires training in application of principles.

Groundwater models

Predicts effect of groundwater abstraction on surface water flow.

More holistic evaluation of water movement and effects.

Currently a research area, needs verification, specific to porous alluvial rivers, very detailed assessments required, data requirements are high/ expensive.