Boosting the assessment of ungauged small hydro sites

Small hydro sites located in mountain environments are often on ungauged streams. A software package called the Integrated Method for Power (IMP 4.0) was designed to assist inexperienced, potential developers of small hydro sites in situations where the cost of professional engineering advice would be prohibitive. The goal was to use an appropriate methodology to enhance the hydrologic information for ungauged small hydro sites and to provide a screening tool that would help to estimate the appropriate level of investment. The basic approach is to use topographic and weather data to compute the time series of streamflow.

IMP was developed in North America over a period of 20 years, and was funded at various stages by BC Hydro, the Washington State Department of Ecology, City of Seattle and Alberta Environment. Natural Resources Canada provided the funding to bring the various independent computer programs into an integrated Windows environment.

IMP considers the time series hydrograph flow, which is usually unknown for small mountain streams, and uses it to:

• Develop a flood frequency curve.

• Select an appropriate penstock-turbine-generator set.

• Calculate the energy capability and economic optimisation of the installed capacity.

• Assess the impact of a development on instream habitat for fish.

The features provided by IMP include those which are common to many resource engineering situations. These include:

• Hydrology model.

• Flood frequency model.

• Power study model.

• River hydraulics model.

• Fish habitat model.

• Instream flow model.

IMP’s basic assessment of a small hydro site begins with an analysis of the hydrology, usually without access to streamflow data. In mountain environments, the local climate, and hence the flow, can vary significantly with elevation and over short distances. One side of a valley often experiences different precipitation than the other side, and snow accumulation and melt are elevation dependant because of the vertical lapse rate for temperature and precipitation. The weather data that are inputs to IMP must be carefully interpreted to be representative of the actual weather conditions above the site. Some of this interpretation is considered automatically by elevation dependant routines in IMP.

The software uses a conceptual model to break down the hydrological analysis into two components. The first deals with the regional climate, and how it can be applied to estimate average precipitation and temperatures in the vicinity of the site. This usually requires advice from a meteorologist familiar with local differences in the climate of the region. The second component is a model of the effect of elevation on these regional climate parameters.

The conceptual hydrologic model generates daily and hourly streamflow on an ungauged watershed from topographic data and a time series of daily precipitation and maximum and minimum temperature. Data may be in US or Canadian weather services format, or may be imported from text files.

Adapted from the University of British Columbia (UBC) watershed model developed for BC Hydro, the hydrologic model is well suited to climatologically heterogeneous mountainous areas like British Columbia in Canada. The model includes quick and slow groundwater routing; surface runoff from impervious areas; snow and glacier accumulation and melt; and surface and ground water storage. Water losses from tree cover are included. The watershed can be divided into sub-basins, each with distinct topography and hydrological parameters.

The UBC model was modified by removing computational hydrologic routines that are not relevant to small hydro sites. For this reason, caution should be exercised in using the hydrologic model in IMP for large watersheds.

Hydrologic simulation

Data describing the watershed above the power site are readily obtained from topographic maps and aerial photographs. The IMP package includes default values of hydrologic parameters that have been calibrated for six regions in British Columbia, and for streams in Manitoba, Newfoundland and Yukon. The regions cover a range of conditions including glaciated mountains on the Pacific coast, in the interior, and the arid interior plateau. The calibration was achieved by adjusting model parameters until a good agreement was reached between the computer output and the observed flows of small gauged watersheds in each region.

Hydrologic simulation will produce daily flows from daily precipitation and maximum and minimum temperatures. In some cases, daily average flows are inadequate for small flashy streams. In other cases, the hydrograph varies diurnally in response to the daily cycle of melting during the day and re-freezing at night. IMP provides a facility for using hourly precipitation patterns to prorate daily runoff to hourly values. Daily snowmelt is distributed hourly based on a diurnal temperature cycle. The hydrologic simulation will then produce hourly flows for analysis of hourly energy production and spill.

The watershed model provides opportunities to explore the sensitivity of hydrologic sequences to factors such as tree cover, soil types, land slopes, and the range of elevation. A particular watershed can also be transposed to a different climate to show the relative importance of watershed and climate parameters.

As there are no observed flows, and the simulated record normally would be too short, the hydrologic simulation model is not used to estimate flows for flood frequency analysis. Instead, a completely different approach is used, which is ideal for quantifying judgements about the climate extremes encountered on an ungauged watershed.

A theory for the closed form mathematical model of flood frequency was developed in 1972. The model can be thought of as an application of the technique of derived probability distributions.

The flood frequency model comprises two components. The first defines the transformation from rainfall (and snow melt) to direct runoff by providing overland and channel routing equations for the time of concentration as a function of the watershed shape, size, and slope and the precipitation intensity. These data are similar to the watershed description used for conceptual hydrologic simulation. Infiltration and groundwater return flows are not included in this model. Water that does not contribute to the flood peak is eliminated by two coefficients: one represents water losses, the other is the ratio of direct runoff to total precipitation. IMP provides advice for selecting these coefficients.

A joint probability distribution function for precipitation events is the second component of the flood frequency model. This provides the flood probability distribution function by numerical integration with the routing equations in the limits of integration. The parameters of the joint probability distribution function for precipitation are easily estimated, even for an ungauged site. They are the reciprocals of the average intensity and average duration of precipitation events. The flood frequency curve is then obtained from the average annual number of events, which can also be estimated.

The method separates the watershed properties from the climatological parameters that influence flood frequency. To assist in transposing precipitation statistical parameters to ungauged areas, IMP compares the mathematically derived flood frequency curves with flood frequency curves developed from recorded discharges on other streams in the region.

Large watersheds

The method has proved to be useful for very large watersheds, as well as those of interest to IMP. On large watersheds, snowmelt may contribute significantly to flood frequency. The contribution of snowmelt to flood frequency is included by providing equivalent parameters developed from analysis of simulated daily snowmelt. IMP will then identify the separate contributions of rain and snow, where the curves cross, and their combined effect.

The flood frequency method provides an opportunity to develop the unsteady flow routing equations from first principles, to apply them to real watersheds, and to gain experience rapidly in understanding the practical role played by fundamental theoretical approximations.

The power study model provides simulation for a power station with one generating unit and with the option of having one reservoir. IMP provides efficiency curves for several types of hydraulic turbine and rough coefficients for a selection of penstock materials. Inflows may be the output from the watershed model (simulated hydrographs), or actual recorded streamflow (in US Geological Survey or Canadian format). The outputs are the firm energy, average annual energy, and the time series of energy, spill and reservoir storage. An optimisation routine conducts a post-analysis around the simulated design and recommends the installed capacity at which the marginal value of the energy generated equals the marginal cost of additional capacity.

The power study model provides opportunities to develop an understanding of the significance of the hydrologic sequence in evaluating a power site. The model can also be used to develop the data for a trade-off curve between reservoir capacity and costs, and generating capacity and costs. The model clearly shows the relationship between energy output and the penstock capacity and identifies the maximum energy that can be developed from a given penstock diameter and material.

The purpose of this model is to provide a quantitative description of the velocity and depth regimes that provide fish habitat. The analysis is for a single cross-section. The results are interpreted in the fish habitat model as ‘habitat area’, described by a strip of unit width extending across the river at a particular location. At each point along this strip the hydraulic model determines the velocity and depth.

The river hydraulics model uses the Manning equation and a stage-discharge rating curve to determine variations in velocity and depth along a cross-section of the river, as a function of discharge. Manning’s ‘n’ can be different at each point across the river, reflecting changes in substrate materials. Data are provided for several river locations. The model does not deal with back eddies. This model provides an understanding of how habitat hydraulic conditions vary as the stage and discharge change at a particular location.

Fish habitat

Using fish habitat preference curves and the output from the hydraulic model of the river, the fish habitat model determines the portion of the river cross-section area that provides the most suitable fish habitat.

Fish habitat preference curves relate the velocity and depth at specific locations to the purported preferences of fish. For each species, and for each life stage, the curves are different. In practice the curves ideally would be developed from a subsurface snorkel survey in the river to identify the locations, velocity and depth preferred by the fish of interest. IMP provides default data for several species and life stages, which can be adequate for a preliminary assessment.

IMP illustrates how the river cross-sectional shape affects its suitability as habitat for specific types of fish and life stages at the specific flows that are likely to be experienced below the site.

Instream flow

The instream flow model determines the time series sequence of habitat from the time series sequence of flows in the river. The calculation of habitat uses the functional relationship between discharge and habitat that was developed by the habitat model. The instream flows can be the natural flow sequence, the regulated power house flow sequence, or the flows in the bypass reach that receives only spills. Frequency analysis provides a convenient method for comparing the habitat impacts of alternative design concepts.

There is often more than one species and life stage present in the river at the same time. This complicates water management decisions that protect fish. IMP provides a hypothetical ‘equivalent fish’ that reflects the combined preferences of a number of species and life stages. This highlights the multiple objective nature of habitat protection and enhancement.

On-line help

The on-line help module of IMP 4.0 is a hyper-textbook containing tips on how to use the software, charts, equations and documentation describing details of the methods used in the modules.

IMP was developed for the Hydraulic Energy Programme of CANMET-Natural resources Canada. Version 4.0 is available free of charge and version 5.0 is currently being developed. For more information contact the authors.