dam breach condition and downstream inundation

Introduction

This document provides an overview of consequence classification for dams in British
Columbia. It outlines a rough method for assessing consequence and some key concepts
that require consideration in assessing consequence. If the method provides a clearly
defined consequence classification then a consequence classification can be assigned. If
the results are uncertain, use of the higher possible consequence classification is
appropriate or a more detailed assessment method should be used. For larger structures or
complicated downstream channel conditions more detailed procedures may be required

These guidelines are only intended for consequence of failure classification. They are not adequate for the preparation of inundation mapping for Emergency Preparedness Planning (EPP), or for the assessment of hazards and risk analysis.

Consequence Classification Guide

The BC Dam Safety Regulation - Schedule 1 “Downstream Consequence Classification Guide” outlines a classification guide for all dams in British Columbia. The consequence classification (very high, high, low, or very low) identifies the potential for damage and loss in the unlikely event of a dam failure. The consequence classification is not a reflection on how safe the dam is; thus age and condition of the dam are not reflected in the Consequence classification.

The consequence classification is used to determine the design requirements for a particular dam, with dams of higher downstream consequence having higher design standards. Suggested design requirements for dams falling under the various consequence classifications are identified in the “Dam Safety Guidelines” published by the Canadian Dam Association.

Dam Breach Flood Determination

The flood hydrograph resulting from a dam breach is dependent on many factors. The primary factors are the physical characteristics of the dam, the volume of the reservoir, and the mode of failure. The dam characteristics such as dam geometry, construction materials, and mode of failure; determine the dimensions and timing of breach formation. Breach formation, volume of reservoir storage, and reservoir inflow at the time of failure determine the peak discharge and the shape of the flood hydrograph.

The following sections provide a method for estimating dam breach parameters and peak flow discharges for earthfill dams. Earthfill dams are focused on because the great majority of small dams are earthfill. When estimating concrete gravity dam breach parameters, a complete failure of a discrete number of monoliths is considered. For concrete arch dams a complete dam failure is considered. Breach times for concrete gravity dams generally fall between 0.1 and 0.5 hours and for concrete arch dams they generally fall between instantaneous and 0.1 hours.

Estimation of Dam Breach Parameters

Work by MacDonald and Landridge-Monopolis (MacDonald, 1984) were successful in relating breaching characteristics of earthfill dams to measurable characteristics of the dam and reservoir. Specifically, a relationship exists between the volume of material eroded in the breach and the Breach Formation Factor (BFF):

BFF = Vw (H)

where:

Vw = Volume of water stored in the reservoir (acre-ft) at the water surface

elevation under consideration

H = Height of water (feet) over the base elevation of the breach

Interpretation of data (MacDonald, 1984) suggests that the estimates of material eroded from earthfill dams may be taken to be:

Vm = 3.75 (BFF)0.77 for Cohesionless Embankment Materials; and

Vm = 2.50 (BFF)0.77 for Erosion Resistant Embankment Materials

where:

Vm = Volume of material in breach (yds3) which is eroded

Using the geometry of the dam and assuming a trapezoidal breach with sideslopes of (Zb:1) the base width of the breach can be computed (MacDonald, 1984) as a function of the eroded volume of material as:

Wb = [27Vm – H2 (CZb + HZbZ3/3)] / [H (C + HZ3/2)]

where:

Wb = Width of breach (feet) at base elevation of breach

C = Crest Width of dam (feet)

Z3 = Z1 + Z2

Z1 = Slope (Z1:1) of upstream face of dam

Z2 = Slope (Z2:1) of downstream face of dam

If the calculated breach width is negative then the reservoir volume is not large enough to fully breach the dam and a partial breach will result. In this case the head of water (H) needs to be adjusted to estimate the breach depth and peak discharge. Maximum breach

widths have historically been limited to breach widths less than 3 times dam height (Fread, 1981). In addition site geometry often limits breach width.

The time of breach development (τ) in hours, has been related to the volume of eroded material (MacDonald, 1984). Interpretation of data suggests that the time for breach development can be estimated by:

τ = 0.028 Vm0.36 for Cohesionless Embankment Materials; and

τ = 0.042 Vm0.36 for Erosion Resistant Embankment Materials

There is a large uncertainty in the eyewitness accounts for many of these failures; thus these equations may tend to overestimate breach times. In addition, these equations appear to produce unrealistically short breach development times in the case of small dams. A lower limit for the breach development time of perhaps 10 minutes for dams constructed of cohesionless materials and 15 minutes for dams constructed of erosion resistant materials seems reasonable.

Due to the uncertainties in breach development parameters, a range of values should be used to assess the computed dam break flood peak discharges. There is a range of alternative procedures for estimating dam break parameters. An example is the computer program BREACH, developed by Fread (1987) which is used for larger complex dams.

Estimation of Dam Breach Peak Discharge

A number of computer programs, such as DAMBRK (Fread, 1988), have been developed for estimating dam break peak discharge. This computer model, and others, utilises unsteady flow conditions in combination with user selected breach parameters to compute the breach flood hydrograph.

Fread (1981) gives an alternative method suitable for many planning purposes. He developed an empirical equation based on numerous simulations with the DAMBRK model. Estimation of the peak discharge from a dam breach is computed as:

Qp = 3.1 W H1.5 [ A / (A + τ H0.5]3

where:

Qp = Dam breach discharge (cfs)

W = Average breach width (feet) W = Wb + ZbH

H = Initial height of water (feet) over the base elevation of the breach

τ = Elapsed time for breach development (hours)

A = 23.4 Sa / W

Sa = Surface area of reservoir (acres) at level corresponding to depth H

The following Tables 1 & 2 contain estimates of dam breach peak flows for overtopping induced failures of earthfill dams based on Fread’s equation. The values used in developing these estimates are presented after the Tables.

Selection of Reservoir Conditions for Breach Analysis

The selected reservoir storage is an important consideration in dam breach analysis. Normally a couple of reservoir conditions, normal pool and maximum storage elevation during floods are considered. For smaller unattended structures usually only the case of dam failure during overtopping needs to be considered. Overtopping could result from a debris blockage, or a beaver dam constructed, in overflow spillway channel.

In evaluating the overtopping dam breach it needs to be remembered that the reservoir storage and head on the dam are greater than for normal pool levels.

Downstream Routing of Dam Breach Flood

As the dam breach flood wave travels downstream there is a reduction in the peak flow. This effect is governed by factors such as:

the channel bedslope,

the cross-sectional area and geometry of the channel and overbank areas,

the roughness of the main channel and overbank,

the existage of storage for floodwaters in off-channel areas, and

the shape of the flood hydrograph.

Small attenuation is associated with:

large reservoir volume,

small confining channel,

steep channel slopes, and

little frictional resistance in channel and overbank areas.

Large attenuation is associated with:

small reservoir volume,

broad floodplain and/or off-channel storage areas,

mild channel slopes, and

large frictional resistance in channel and overbank areas.

There are a number of methods for modelling the attenuation of peak flow as the breach flood wave travels downstream. For consequence classification a simplified procedure based on generalised flood attenuation curves developed by the USBR (1982) is often adequate. The curves presented in Figure 1 should be used conservatively as they utilize generalised solutions to approximate the reduction of flood peak discharge with distance downstream of the dam. For example the attenuation would be much smaller for a dam breach flow travelling down a steep narrow valley.

Downstream Hazard Classification

Once the dam breach flood inundation path has been determined, the resulting consequence of failure classification can be determined. For BC, the classification system is outlined in Schedule 1 “Downstream Consequence Classification Guide” of the British Columbia Dam Safety Regulation. Refer to the Regulation for Schedule 1. The highest consequence rating in one of the three categories; loss of life, economic and social loss, and environmental and cultural losses is the consequence rating for the dam.

In estimating loss of life in a dam breach one needs to consider:

Time of day of failure

Number of homes in inundation area

Flood depth and velocity

3 people per home (USBR, 1988)

Highways

Recreation

Warning time

Sources of uncertainty

For further information on this topic the “Downstream Hazard Classification Guidelines” produced by the US Bureau of Reclamation (USBR, 1988) are a good starting point.

Other Considerations

There are many other factors that can influence the consequence of failure classification. They include:

Debris build-up and sediment transport can increase floodwave size and its destructive power,

Channel avulsions especially on alluvial fans,

Multiple dams on a river system, and

Current and potential future downstream development,

Warning systems can be effective in reducing loss of life in the event of a dam failure. Thus they are effective risk management tools, however they do not change the consequence of failure classification.