Dam Safety Evaluation

Open Dam Safety Evaluation

Determine the design flood criteria for a dam in South Africa using the SANCOLD classification system. This guide covers size-and-hazard classification, dam Category (I / II / III), the Recommended Design Discharge (RDD), the Safety Evaluation Flood (SEF), PMF and RMF-based expressions, spillway adequacy routing, freeboard requirements, and the regulatory context under the DWS Dam Safety Regulations.

Overview

Every dam must be designed and evaluated against two flood criteria:

• Recommended Design Discharge (RDD) — the flood that the spillway must pass without structural damage or loss of function. The RDD represents the day-to-day operational flood that can reasonably be expected several times in the life of the dam.
• Safety Evaluation Flood (SEF) — the extreme flood that the dam must survive without catastrophic failure (i.e., without dam-break). Passage of the SEF may involve overstressing of auxiliary spillways, parapet inundation, or minor structural damage, but the main dam body must remain intact and the reservoir must not be uncontrollably released.

The SANCOLD (South African National Committee on Large Dams) guidelines specify both criteria as a function of the dam’s Category — a combined classification of size and downstream hazard. The Category ranges from I (small, low-hazard) to III (large or high-hazard) and determines the severity of the design flood, ranging from a 1:50 year event for small low-hazard dams to a full Probable Maximum Flood (PMF) for the highest-hazard structures.

The Dam Safety Evaluation tool automates this classification, selects the RDD return period and SEF factor per SANCOLD Tables 5.1 and 5.2, and provides a framework for the spillway adequacy check when inflow hydrographs and stage-storage-discharge curves are supplied.

Regulatory context

Under the South African National Water Act (Act 36 of 1998), all dams with a wall height exceeding 5 m and a storage capacity greater than 50 000 m³ are classified as “dams with a safety risk” and fall under the DWS Dam Safety Regulations (Government Notice R139 of 24 February 2012, as amended). Owners are legally required to commission periodic safety evaluations by an Approved Professional Person (APP) — engineer registered with ECSA and accredited with the Dam Safety Office. The flood-evaluation component of those studies is what this tool supports.

Dam hazard classification

Classification follows the SANCOLD / DWS three-step procedure: determine Size, determine Hazard, then combine them via the Category matrix.

Size classification

Size is based on the maximum wall height, measured from the lowest point of the general foundation level (or the natural ground level if no foundation cut-off exists) to the non-overspill crest. SANCOLD Report 4 (1991), Table 2.1:

Size class	Maximum wall height
Small	> 5 m and < 12 m
Medium	≥ 12 m and < 30 m
Large	≥ 30 m

Storage capacity is used as a secondary discriminator when the height is borderline — a dam with height < 12 m but storage > 12 million m³ may be promoted to Medium on the strength of volumetric impact alone. The DWS Dam Safety Office retains final classification authority.

Wall height vs dam height

“Wall height” in SANCOLD usage is the structural height of the dam wall. Some older South African references used “height above riverbed” or “height to spillway crest” — these are different quantities and can give different size classes for the same structure. Always confirm the definition when comparing with historical classifications.

Hazard classification

Hazard is the potential consequence of dam failure on life and property downstream, independent of the probability of failure. SANCOLD Report 4 (1991), Table 2.2:

Hazard class	Potential loss of life	Potential economic loss	Typical downstream setting
Low	None	Minimal	Uninhabited farmland
Significant	≤ 10 lives	Significant (infrastructure)	Rural village, minor road
High	> 10 lives	Great (urban, critical)	Town, major road, industry

Loss-of-life estimation requires a dam-break inundation study. Approximate methods (e.g. Graham 1999 — fatality rates based on warning time and severity of flooding) are acceptable for screening but full 2D hydraulic modelling with HEC-RAS 2D or similar is required for Category III dams near populated areas.

Dam Category matrix

Size and Hazard combine per SANCOLD Report 4, Table 2.3:

	Low hazard	Significant hazard	High hazard
Small	I	II	II
Medium	II	II	III
Large	III	III	III

Category determines both the required RDD return period and the SEF factor, and also governs inspection frequency, instrumentation requirements, and emergency-action-plan obligations under the DWS regulations.

Why size matters when hazard is low

A large, low-hazard dam is still Category III — a large dam represents a large kinetic and potential energy release, and even if the immediate downstream reach is uninhabited, the cumulative effect of an uncontrolled release propagates far downstream. Conversely, a small dam above an urban area is Category II, not III, because its energy release is physically limited by its smaller storage.

Recommended Design Discharge (RDD)

The RDD is the peak discharge the spillway must pass while maintaining structural integrity and design freeboard. It is specified as a return-period-based event per SANCOLD Table 5.1:

	Low hazard	Significant hazard	High hazard
Small	0.5 Q₅₀ – Q₅₀	Q₁₀₀	Q₁₀₀
Medium	Q₁₀₀	Q₁₀₀	Q₂₀₀
Large	Q₂₀₀	Q₂₀₀	Q₂₀₀

The peak flow for the applicable return period is obtained from a standard design flood method — the Rational Method for catchments under about 15 km², the regionalised Unit Hydrograph or SCS Method for medium catchments, the QRT Calculator or Flood Frequency Analysis for gauged catchments, and the RMF Calculator or SDF Method for regional envelope checks.

The Dam Safety Evaluation tool accepts RDD as a direct input or as a link to any saved peak-flow calculation. At least two independent methods should be used and the results reconciled — a single-method estimate is not acceptable for dam design.

When the RDD is a range

For Category I dams the SANCOLD table gives 0.5 Q₅₀ – Q₅₀. The engineer selects a value in the range based on the downstream sensitivity and the cost differential between the bounds. Best practice is to size for Q₅₀ unless economic analysis explicitly justifies a lesser standard, because the marginal cost of moving from 0.5 Q₅₀ to Q₅₀ is typically small (one extra module of spillway width).

Safety Evaluation Flood (SEF)

The SEF is the extreme flood that the dam must survive without dam-break failure, though it may be passed with compromised freeboard or minor structural distress. Two parallel formulations exist in the SANCOLD guidelines and both should be computed.

SEF as a fraction of PMF

Per SANCOLD Report 1 (1986) Table 4.1 and Report 4 (1991) the SEF is expressed as a fraction of the Probable Maximum Flood (PMF):

	Low hazard	Significant hazard	High hazard
Small	0.2 PMF	0.5 PMF	0.7 PMF
Medium	0.5 PMF	0.7 PMF	1.0 PMF
Large	0.7 PMF	1.0 PMF	1.1 PMF

The PMF is derived from the Probable Maximum Precipitation (PMP) combined with a design-flood hydrological model. See the PMP Calculator and SCS Method — a typical workflow extracts PMP from the grid, applies an SCS-CN-like loss model, and convolves with a unit hydrograph using the Hydrograph Generator.

SEF relative to RMF

Report 4 (1991) also allows the SEF to be expressed in stepped increments of the Regional Maximum Flood (RMF) — Kovács’ regional envelope curve. SEF options are RMF, RMF-Δ, or RMF+Δ, where Δ represents one K-region step in Kovács’ classification.

	Low hazard	Significant hazard	High hazard
Small	RMF − 2Δ	RMF − 1Δ	RMF
Medium	RMF − 1Δ	RMF	RMF
Large	RMF	RMF	RMF + 1Δ

Both the PMF-fraction and RMF-stepped values should be computed. If the two converge, confidence is high; if they diverge materially (>30 %), an engineering judgement must be documented explaining which is adopted. In most Southern African hydrologies the RMF-based SEF is more conservative than the PMF-based value for small-to-medium catchments; the reverse is often true for large catchments where the PMF is driven by multi-cell storms that the RMF envelope captures only approximately.

PMF fractions greater than 1.0

For large high-hazard dams the SANCOLD SEF is 1.1 × PMF. This is NOT a statement that the 1.1 × PMF event is physically possible — PMP is by definition the theoretical upper bound. The 1.1 factor is a conservative margin on the PMF hydrograph that accounts for uncertainty in PMP estimates, in loss modelling, and in unit hydrograph regionalisation. It is applied to the peak and volume of the PMF hydrograph — the resulting flood is routed through the reservoir and the spillway checked as for any other design event.

Spillway adequacy check

The RDD and SEF are inflow peaks — they must be translated into spillway releases and reservoir stages through a flood routing analysis. The outputs of interest are:

• Maximum water surface elevation (Max WSEL) — the peak stage reached during routing.
• Spillway discharge at Max WSEL — the peak release.
• Freeboard at Max WSEL — the remaining margin below the dam crest and the parapet top.

The routing procedure is the same as for any reservoir — see the Flood Routing tool for the modified-Puls (storage-indication) algorithm. Specifically for the dam-safety check:

1. Construct (or reuse) the stage-storage curve from topographic surveys, as-built drawings, or LiDAR.
2. Construct the stage-discharge (rating) curve for every outlet: service spillway, auxiliary spillway, outlet works, low-level release.
3. Develop the inflow hydrograph for the RDD and SEF — peak scaled from a synthetic unit hydrograph (e.g., using the Hydrograph Generator) with an appropriate volume consistent with the design storm duration (24 hour is typical for small-to-medium catchments; 48 to 72 hour for large).
4. Route each hydrograph through the reservoir starting from the Full Supply Level (FSL) — or a higher initial pool if the reservoir is consistently at surcharge at the time of year of interest.
5. Compare Max WSEL against the dam crest elevation for the SEF and against the spillway design elevation plus freeboard for the RDD.

Starting pool elevation

SANCOLD guidelines assume the reservoir is at FSL at the onset of the design storm — a conservative but standard assumption that avoids arguments about antecedent conditions. For reservoirs with a formal flood-control operating rule that lowers the pool before a forecast storm, a variance to this assumption may be permitted, but must be rigorously justified and documented in the safety report.

Freeboard requirements

Freeboard is the vertical distance between the Max WSEL during the design event and the top of the dam crest (or parapet top, if present). Components:

• Normal freeboard — the static margin for design flood passage (RDD).
• Minimum freeboard — the minimum margin during SEF passage.

Three physical processes consume freeboard and must be added to any structural margin:

Wind setup

Wind blowing over the reservoir surface exerts a drag that tilts the water surface downwind, raising the elevation at the downwind shore (which is typically the dam) above the still-water surface. The classical Zuider Zee formula:

$$S \;=\; \frac{V^2 \cdot F}{63200 \cdot D}$$

Zuider Zee wind setup

where $V$ is the sustained wind velocity (m/s), $F$ is the effective fetch (m), and $D$ is the average water depth along the fetch (m). Output $S$ is in metres.

The design wind velocity is typically taken as the 1:100 year one-hour average wind from the weather service. Effective fetch accounts for the reservoir shape — a narrow, elongated reservoir aligned with the prevailing wind produces full fetch, while a circular reservoir uses an effective fetch equal to the maximum radius.

Wave run-up

Wind generates waves; when those waves meet the upstream face of the dam they run up the slope a vertical distance $R$ above the still-water level. $R$ depends on wave height, wave period, slope angle, and face roughness.

For embankment dams with rock armour, the Saville (1957) method computes significant wave height $H_s$ and period $T_s$ from wind speed and fetch, then applies empirical run-up curves as a function of slope:

$$H_s \;=\; 0.283 \cdot \tanh\!\left(0.0125 \cdot (g F / V^2)^{0.42}\right) \cdot V^2 / g$$

Saville wave height — deep water (Sverdrup-Munk-Bretschneider)

For concrete faces and smooth slopes, simplified relations apply: $R \approx 1.5 \, H_s$ for smooth concrete on a 1:2 slope; $R \approx 1.0 \, H_s$ for rough rock armour on the same slope.

The TAW (2002) procedure, increasingly adopted internationally, couples wave generation (using the Wilson fetch-limited wave-growth relations) with a single run-up formula that includes berm effects and roughness reduction. The HydroDesign Dam Safety tool implements both the Saville / SMB legacy procedure and the TAW procedure; results typically agree to within 10 percent for standard embankment profiles.

Settlement allowance

Earth and rockfill dams settle over time. Post-construction settlement typically follows a logarithmic trend in time:

$$\delta(t) \;=\; \alpha \cdot H \cdot \log_{10}(t / t_0)$$

Post-construction dam settlement (empirical)

where $\delta$ is the cumulative settlement, $H$ is the embankment height, $t$ is time since construction, $t_0$ is a reference time (typically 1 year), and $\alpha$ is an empirical rate coefficient (0.001 to 0.003 for well-compacted earth, higher for rockfill with rock-flour crusher-run).

For design purposes, a total settlement allowance of 0.5 to 1.5 percent of wall height is standard. This allowance is added to the freeboard requirement at construction so that the design freeboard is maintained throughout the design life.

Design freeboard requirements

The design freeboard is the sum of wind setup, wave run-up, and settlement, plus a safety margin. SANCOLD Report 3 (1990, revised 2011) provides the full procedure with worked examples and regional wind data.

Freeboard type	Minimum requirement (typical)
Normal (at RDD)	≥ (wind setup) + (wave run-up) + 0.5 m
Minimum (at SEF)	≥ settlement allowance (no overtopping permitted)
Absolute minimum (at SEF for earthfill)	≥ 0.3 m (per ICOLD / USBR practice)

Different rules for different dam types

An earthfill or rockfill embankment cannot tolerate any overtopping — even a few hours of flow over the crest will initiate breach formation. A concrete gravity or arch dam, by contrast, can tolerate modest parapet overtopping provided the downstream face is erosion-resistant and the foundation is not undermined. The SANCOLD freeboard tables reflect this: earthfill dams require strictly positive freeboard at SEF, while concrete dams may allow parapet inundation under SEF if the resulting energy is dissipated safely.

Outlet works and emergency release

Large dams typically have multiple release structures with distinct roles:

• Service spillway — primary route for routine flood release; typically a fixed-crest ogee or a gated structure sized to pass the RDD and partially attenuate larger events.
• Auxiliary spillway — passes flows above the service-spillway capacity; often a fuse-plug, labyrinth, or stepped-channel configuration activated only during rare events.
• Outlet works / low-level outlet — used for reservoir drawdown, sediment flushing, environmental flow release, and emergency dewatering.
• Emergency pump / siphon release — rarely-specified but increasingly considered for climate-adaptation retrofits.

The stage-discharge rating curve for the reservoir is the sum of all active outlets at each stage. The Flood Routing tool supports composite outlets — see the composite-outlet section of the Flood Routing guide.

Spillway types and discharge coefficients

Spillway type	Typical $C_d$	RDD / SEF application
Sharp-crested weir	1.8 – 1.9	Small farm dams, instrumentation use only
Broad-crested weir	1.6 – 1.7	Small to medium dams; robust, self-cleaning
Ogee (WES standard)	2.0 – 2.2	Medium to large dams; hydraulically efficient
Labyrinth / piano-key	1.8 – 2.1	Retrofits — 1.5 to 2x the capacity of straight weir
Stepped chute	effective 1.6	High-head dams with built-in energy dissipation
Radial gate (Tainter)	varies	Large dams with flood-control operating rules
Fuse-plug auxiliary	n/a (all-or-nothing)	Extreme-event relief for infrequent activation

Discharge coefficients $C_d$ are referenced to the equation $Q = C_d \, L \, H^{3/2}$ where $L$ is crest length, $H$ is head over crest. Use the Weir Analysis tool for detailed rating-curve development including submergence, approach-velocity, and end-contraction effects.

Fuse plugs and emergency overflow

A fuse plug is a purposefully-erodible section of embankment designed to wash out at a predetermined pool elevation, opening an emergency spillway channel that otherwise remains inactive. Fuse plugs are attractive retrofits for Category II and III dams that fail the SEF check because they add capacity only during rare events — routine operation is unaffected.

Design considerations:

• Trigger level — top of fuse-plug crest, typically 0.3 to 0.5 m above FSL of the adjacent service spillway.
• Erodibility — fuse-plug material is gap-graded and under-compacted relative to the main dam; must wash out predictably once overtopped.
• Channel capacity — post-washout channel must have clear capacity to pass the balance of the SEF (SEF peak minus service-spillway release at fuse-plug crest).
• Dam-break risk — the erosion must stop at the channel cross-section; uncontrolled progressive erosion into the main dam is the failure mode to guard against.

Fuse plugs are the cheapest way to upgrade an undersized Category II dam; costs are typically 5 to 15 percent of a new equivalent-capacity concrete spillway.

Inundation mapping and dam break

For Category III dams, and often for Category II dams near populated areas, a dam-break inundation study is required in addition to the spillway adequacy check. The study models the hypothetical sudden or progressive failure of the dam and produces:

• Flood wave arrival times at downstream receptors (villages, roads, critical infrastructure).
• Maximum depth and velocity at each receptor.
• A dam-break inundation map delineating the area subject to flooding.
• An Emergency Action Plan (EAP) triggered by specified pool-level thresholds.

Breach parameters

Modern dam-break analysis treats the breach as a progressively-eroding trapezoidal notch through the dam. The breach is parameterised by:

• Breach bottom width $B_b$ — typically 2 to 5 times the dam height for embankment dams.
• Breach side-slope $z$ — horizontal : vertical; typically 0.5 for embankments, 0 for concrete.
• Breach formation time $t_f$ — time from initiation of breach to full development; typically 0.5 to 2 hours for earthfill.
• Trigger level — pool elevation at which breach begins (typically crest overtopping or piping failure initiation).

Empirical breach-parameter equations from Froehlich (2008), Wahl (2014), and the USBR (2015) are commonly used. The breach outflow peak often greatly exceeds the inflow during flood passage — for a typical earthfill dam, peak breach discharge can be 5 to 20 times the inflow peak.

Inundation hydraulics

The dam-break hydrograph is propagated downstream using 1D or 2D hydraulic models. 1D models (HEC-RAS standard mode) are appropriate for deep, confined valleys; 2D (HEC-RAS 2D, TUFLOW, MIKE21) is required for floodplain spreading and urban flooding. Outputs include:

• Maximum depth at each cell / cross-section.
• Maximum velocity at each cell.
• Depth × velocity product — a standard measure of hazard intensity for stability assessment of people, vehicles, and buildings.
• Arrival time — time from breach initiation to flood arrival, used for warning-time and EAP planning.

Dam-break modelling is outside the scope of the Dam Safety Evaluation tool — it requires 2D hydraulic models (HEC-RAS 2D, MIKE21, Delft3D) and detailed downstream terrain data (typically LiDAR at 1 m grid resolution or better). See the references for standard procedures. The DWS Dam Safety Office requires the EAP to be filed for every Category III dam and for Category II dams in populated catchments.

Loss of life estimation

The Graham (1999) method gives a fatality rate as a function of warning time, flood severity, and understanding of the hazard:

$$\mathrm{LOL} \;=\; f_w \cdot \mathrm{PAR}$$

Graham LOL estimate

where $\mathrm{PAR}$ is the Population at Risk (people in the inundation zone) and $f_w$ is the fatality rate (from Graham Table 2, ranging from 0.0002 for low-severity flooding with > 1 hour warning to 0.75 for high-severity flooding with no warning). The more recent RESCDAM (Assaf 2003) and DSO-99-06 (Graham 1999) methods refine this with depth-velocity dependencies.

Climate change and non-stationarity

The SANCOLD guidelines and the regional flood-frequency methods underpinning them were developed from the 20th-century hydrological record. Climate change introduces non-stationarity — the statistical distribution of extreme rainfall and runoff is no longer constant in time. For dams being licensed in the 2020s with design lives extending to 2070 and beyond, this matters.

Regional downscaled projections for Southern Africa (IPCC AR6 Chapter 12, Atlas regions AFR-SW, AFR-SE, AFR-S) indicate:

• Mean annual rainfall is projected to decrease 10 to 30 percent in the southwestern Cape; increase modestly or stay stable in the northeast; high uncertainty in the Highveld and Karoo.
• Extreme daily rainfall (99th percentile) is projected to increase across most of the subcontinent by 5 to 25 percent by mid-century under a moderate-emissions scenario — consistent with global Clausius-Clapeyron scaling of atmospheric moisture.
• Evapotranspiration demand is projected to increase substantially, reducing net runoff even where rainfall increases.

For dam safety, the implication is that the design flood is likely higher in 2070 than today, by a factor that varies regionally but is typically 1.05 to 1.3. A simple engineering response is to scale the RDD and SEF by a regional climate-adjustment factor — the Water Research Commission has published guidance (Cullis et al., 2015; WRC TT 716/17). The DWS has published a climate-change guidance note (DWS 2019) that recommends a 20 percent increase in PMF for high-hazard dams commissioned after 2018.

Adaptive pathways

Rather than designing for a single projected 2070 flood, some SANCOLD dam owners are adopting an adaptive pathway approach: design for current SANCOLD criteria but include physical space (spillway width, parapet strengthening) to accommodate a future upgrade as climate projections firm up and observed trends emerge. This is cheaper than retrofitting a dam that was designed right to the current standard, and avoids premature commitment to an uncertain projection.

Instrumentation and surveillance

Dam safety is not only about passing a design flood on paper — it requires continuous confirmation that the dam is behaving as designed. The DWS Dam Safety Regulations prescribe minimum instrumentation by Category:

Category	Minimum instrumentation	Inspection frequency
I	Pool-level gauge; annual visual inspection	Annual
II	Pool level, seepage weir, piezometers; biannual inspection	10-year formal safety evaluation
III	Plus crest-settlement survey, inclinometers, accelerometers	5-year formal safety evaluation

Flood-specific instrumentation: a real-time pool-level gauge with telemetry is required for all Category II and III dams. Telemetered rainfall within the catchment is best practice for early warning. Automated alarm on rising pool levels during storm events is required for all Category III dams.

Real-time flood forecasting — coupling a rainfall-runoff model (e.g., a Hydrograph Generator instance) with reservoir routing to give 6 to 24 hour forecasts of pool level — is becoming standard for Category III dams and large regulated storage systems.

Uncertainty and sensitivity

The outputs of a dam-safety flood evaluation carry substantial uncertainty. A responsible report quantifies at least the following:

• Return-period flood peaks — confidence intervals from the frequency-analysis method (e.g., 95 % CI ± 30 percent at Q₁₀₀, widening to ± 60 percent at Q₁₀₀₀).
• PMP — the WMO (2009) methods provide no confidence bounds explicitly; practical uncertainty is ± 20 percent on the 1-day depth.
• PMF — combined uncertainty from PMP, loss parameters, and UH method can reach ± 40 percent.
• Stage-storage-discharge — survey errors typically < 2 percent on storage, ± 5 percent on spillway discharge coefficients.
• Inundation models — hydraulic roughness (Manning’s $n$) is typically ± 25 percent; breach parameters ± 50 percent.

Best practice is a Monte Carlo or simpler scenario sensitivity analysis over the key input distributions, producing a confidence envelope on the Max WSEL, the peak release, and the inundation area. For Category III dams this is a minimum expectation; for Category II dams, scenario analysis at the 10th / 50th / 90th percentile inputs is usually adequate.

Regulatory compliance checklist

Before a dam can be commissioned or re-licensed under the DWS regulations, the flood evaluation must demonstrate:

• Size and Hazard classification agreed with the DWS Dam Safety Office.
• Category determined per SANCOLD Table 2.3.
• RDD established using at least two independent flood-estimation methods.
• SEF established both as a PMF fraction and in RMF stepped form; divergence documented.
• RDD routed through the reservoir — spillway adequacy confirmed, freeboard satisfied.
• SEF routed through the reservoir — dam integrity confirmed; no overtopping of the main dam crest (parapet overtopping allowed for Category II and III only with engineering justification).
• Freeboard analysis per SANCOLD Report 3 including wind setup, wave run-up, and settlement.
• For Category III dams: dam-break inundation study and Emergency Action Plan.
• Safety evaluation report signed and stamped by an APP registered with ECSA and accredited by the DWS Dam Safety Office.
• Five-year inspection interval for Category III, ten-year for Category II, as applicable.

Worked example — 20 m medium earthfill dam near a rural village

The following example demonstrates a full dam safety evaluation workflow for a representative farm dam in KwaZulu-Natal.

Given information

• Dam type: Earthfill embankment, upstream rip-rap protection.
• Wall height: 20 m above foundation.
• Storage: 2.0 × 10⁶ m³ at FSL.
• Catchment area: 45 km², rolling foothills.
• Downstream setting: Rural village of ~300 residents 2.5 km downstream of the dam; provincial road crossing at 4 km.
• Downstream potential loss of life: Estimated 3 – 8 lives in the worst case (with 0.5 hr warning, depth 1.5 – 3 m through the village).
• Potential economic loss: Significant (road, houses, livestock, fencing).

Step 1 — Size class

Wall height 20 m → Medium (≥ 12 m and < 30 m).

Step 2 — Hazard class

3 – 8 lives, significant economic loss → Significant hazard.

Step 3 — Category

Medium + Significant → SANCOLD Table 2.3 → Category II.

Step 4 — RDD selection

Category II, Medium + Significant → SANCOLD Table 5.1 → RDD = Q₁₀₀.

For the 45 km² catchment the 1:100 year peak is computed using two independent methods:

• SCS method (see SCS Method) with 1:100 year design rainfall from the Design Rainfall tool: 185 m³/s.
• Regionalised unit hydrograph (HRU 1/72): 210 m³/s.

Mean of the two methods, rounded conservatively: RDD = 210 m³/s.

Step 5 — SEF selection

Category II, Medium + Significant → SANCOLD Table 5.2 → SEF = 0.7 PMF.

PMP at the site (from the PMP Calculator and the 1-day grid): 580 mm. Applying the SCS-CN method with a saturated-AMC curve number of 87 and convolving with the SCS unit hydrograph using the Hydrograph Generator gives PMF peak = 950 m³/s.

SEF peak = 0.7 × 950 = 665 m³/s.

Cross-check against the RMF for the region (Kovács K-region 4.6 with A = 45 km²): RMF ≈ 720 m³/s. SEF by the RMF formulation is RMF with no step adjustment → 720 m³/s. Mean of the two formulations: SEF = 690 m³/s (adopted).

Step 6 — Spillway adequacy

Existing service spillway: 25 m wide fixed-crest ogee, crest at FSL. Rating curve from the Weir Analysis tool.

Routing the 210 m³/s RDD hydrograph (peak 210 m³/s, volume 3.2 × 10⁶ m³, base 18 hr) through the 2 × 10⁶ m³ reservoir using the Flood Routing tool gives Max WSEL = FSL + 1.35 m, peak release = 170 m³/s. Freeboard at Max WSEL = dam crest − 1.35 m − FSL = 1.65 m remaining. Wind setup + wave run-up + settlement = 0.8 m. Net freeboard margin = 0.85 m → adequate.

Routing the 690 m³/s SEF gives Max WSEL = FSL + 2.4 m, with freeboard to crest of 0.6 m. Wind + wave + settlement = 0.8 m under extreme conditions → marginal: overtopping possible. An auxiliary spillway (fuse-plug or labyrinth retrofit) is recommended. Once fitted, re-routing gives Max WSEL = FSL + 1.9 m and a 1.1 m freeboard margin → adequate with retrofit.

Step 7 — Conclusions

The dam’s RDD is adequately handled by the existing service spillway. The SEF, however, requires either a spillway upgrade or a fuse-plug auxiliary to maintain freeboard under the 0.7 PMF event. A Dam-Break study and Emergency Action Plan are also required given the Category II classification and the proximity of the village — these must be commissioned separately.

The value of the Category matrix

Notice how the 20 m dam near a village needed substantial retrofit work, despite not being a “large” dam by any intuitive definition. The SANCOLD Category matrix deliberately elevates such structures to Category II to force a rigorous SEF check. Dam owners who classify their structures informally and work only to RDD standards routinely underestimate the safety investment required.

Limitations and caveats

• The tool does not perform flood routing itself. It determines the RDD and SEF peaks and the applicable criteria. The routing step must be done in the Flood Routing tool (or a full-feature package such as HEC-HMS or HEC-ResSim for complex multi-outlet structures).
• PMF estimation is method-dependent. The PMF is not a single number — different PMP products, different loss models, and different unit hydrograph regionalisations produce PMF peaks that can differ by 30 to 50 percent. The SEF should bracket the likely range and the final adopted value should be the more conservative.
• Regional context matters for the SA Kovács-based RMF. For catchments outside the mapped Kovács K-regions (e.g., highly disturbed mining areas, or climate-changed catchments), the RMF-based SEF may understate the true extreme.
• Climate change is not explicitly handled. The SANCOLD tables are based on the observed hydrological record. Climate-change adjustments (IPCC AR6 regional projections for Southern Africa, DWA 2013 climate-change guidance) should be applied externally for dams being licensed now with a design life extending past 2070.
• The tool gives peaks, not volumes. For spillway adequacy on narrow-crested structures, the peak is the governing variable. For spillway adequacy on broad-basin reservoirs where routing is substantial, the volume of the inflow hydrograph matters as much as the peak — always compute both and route.
• Approved Professional Person sign-off required. Outputs of this tool are indicative and do not replace the formal safety evaluation by a registered APP. DWS Dam Safety Office will only accept reports signed and stamped by an accredited practitioner.

Related tools and cross-references

• PMP Calculator — obtain PMP depth needed for PMF derivation.
• Hydrograph Generator — convolve PMP and design rainfall with a unit hydrograph to produce the PMF and RDD hydrographs.
• Flood Routing — level-pool routing of inflow hydrographs through the reservoir for the spillway adequacy check.
• RMF Calculator — compute the Kovács regional maximum flood for the RMF-stepped SEF formulation.
• Rational Method — peak-flow estimation for the RDD on small catchments.
• SCS Method, Unit Hydrograph, SDF Method — alternative peak-flow methods for the RDD and for deriving PMF on medium-to-large catchments.
• QRT Calculator, Flood Frequency Analysis — gauged flood-frequency methods for the RDD when suitable records exist.
• Weir Analysis — spillway rating curves.
• Design Rainfall, Daily Rainfall Data — design rainfall inputs.

References

• SANCOLD. (1986). Interim Guidelines on Safety in Relation to Floods. South African National Committee on Large Dams, Report No. 1, Pretoria. (Original tabulation of RDD and SEF criteria.)
• SANCOLD. (1991). Guidelines on Safety in Relation to Floods. South African National Committee on Large Dams, Report No. 4, Pretoria. (Standard reference; the basis for the Category matrix used in this tool.)
• SANCOLD. (1990, revised 2011). Guidelines on Freeboard for Dams. SANCOLD Report No. 3, Pretoria.
• SANCOLD. (2013). Interim Guidelines on the Safety of Dams. Covering procedural aspects of dam safety management in South Africa.
• Department of Water and Sanitation. (2012). Dam Safety Regulations. Government Notice R139 of 24 February 2012, in terms of the National Water Act, Act 36 of 1998. (The statutory instrument.)
• Department of Water Affairs. (1988). Regional Maximum Flood Peaks in Southern Africa. Technical Report TR 137, Pretoria.
• Kovács, Z.P. (1988). Regional maximum flood peaks in Southern Africa. Technical Report TR 137, Department of Water Affairs. (Kovács envelope methodology.)
• Hydrological Research Unit. (1972). Design Flood Determination in South Africa. HRU Report 1/72, University of the Witwatersrand, Johannesburg.
• ICOLD. (2003). Dams and Floods — Guidelines and Case Histories. International Commission on Large Dams, Bulletin 125, Paris.
• ICOLD. (2011). Role of Dams in Flood Management. ICOLD Bulletin 131, Paris.
• USBR. (1987). Design of Small Dams (3rd ed.). US Bureau of Reclamation, Denver. (Freeboard, spillway, outlet design — still a standard reference.)
• USACE. (1995). Hydrologic Engineering Requirements for Reservoirs. EM 1110-2-1420, US Army Corps of Engineers, Washington DC.
• USACE. (2002). Hydrologic Modelling System HEC-HMS — Technical Reference Manual. Hydrologic Engineering Center, Davis, California.
• FERC. (2018). Engineering Guidelines for the Evaluation of Hydropower Projects. Chapter 2 — Selecting and Accommodating Inflow Design Floods for Dams. Federal Energy Regulatory Commission, Washington DC.
• Graham, W.J. (1999). A Procedure for Estimating Loss of Life Caused by Dam Failure. US Bureau of Reclamation DSO-99-06, Denver. (The standard SA-referenced method for LOL estimation.)
• Wahl, T.L. (1998). Prediction of Embankment Dam Breach Parameters — A Literature Review and Needs Assessment. USBR DSO-98-004.
• Smithers, J.C. & Schulze, R.E. (2003). Design Rainfall and Flood Estimation in South Africa. Water Research Commission Report 1060/1/03, Pretoria.
• Van Vuuren, S.J. & van Dijk, M. (2012). Review of SANCOLD Guidelines on Safety in Relation to Floods. Civil Engineering (SAICE), 20(5), 47 – 52.
• Van Vuuren, S.J. (2016). Dam Safety Evaluation Report — Template and Guidance. ECSA CPD material, Pretoria.
• Pitman, W.V. (2011). Overview of water resource assessment in South Africa: Current state and future challenges. Water SA, 37(5), 659 – 664.

---

Open Dam Safety Evaluation