Standard Design Flood (SDF) Method

Open Design Flood Estimation

Estimate design peak floods for South African catchments using Prof. Will Alexander’s Standard Design Flood — a deterministic method calibrated regionally to 29 hydrologically homogeneous zones covering the entire country. This guide explains the method, the regional K-value parameters, and a worked example for SANRAL-style projects.

Overview

The Standard Design Flood (SDF) method is a South African deterministic flood estimation procedure developed by Professor W.J.R. Alexander and published as the Standard Design Flood — a new design philosophy (University of Pretoria, 2002) and the Flood Risk Reduction Measures manual (2003). It was developed to provide a nationally consistent, transparent approach to design flood estimation for the SANRAL road network and has become the method of choice for most SANRAL drainage designs in the small- to medium-catchment range.

The fundamental idea is simple: divide South Africa into hydrologically homogeneous regions, calibrate a pair of region-specific parameters (the 2-year and 100-year runoff coefficients, $C_2$ and $C_{100}$) against observed flood records, and combine them with Alexander’s modified Rational-form equation to produce design floods for any return period up to 1:100 years (and by extrapolation, 1:200).

The method divides the country into 29 SDF regions. Each region has its own pair of calibration coefficients, summarised as K-values. The calibration was originally performed against all available gauged flood records at the time of publication (long-term records from the Department of Water Affairs gauging network), making the SDF one of the best empirically-anchored flood estimation procedures in South Africa.

When SDF is required

SANRAL’s current drainage policy requires the SDF to be included in every design flood estimate for national roads within the calibrated catchment range (up to ~500 km²). If your project involves SANRAL infrastructure, running SDF is effectively mandatory.

What the SDF method does

Given a catchment’s area, time of concentration and design rainfall depth, the SDF method produces a peak design discharge for a chosen return period. Unlike the DRH method or Unit Hydrograph method, the SDF does not produce a full hydrograph — it produces a single peak value. Where a hydrograph is needed, the SDF is commonly combined with a standard hydrograph shape (SCS triangular) scaled to match the SDF peak and the estimated runoff volume.

The SDF is applicable for catchments approximately 10 – 500 km². Outside this range it should be used only as a cross-check against more appropriate methods: the Rational Method for smaller catchments, and regional flood frequency analysis or the RMF for very large catchments.

The SDF formula

The SDF equation is structurally similar to the Rational formula — peak discharge equals a runoff coefficient times rainfall intensity times area — but the runoff coefficient is magnitude-dependent, varying with return period between calibrated values at 2 years ($C_2$) and 100 years ($C_{100}$). The general form is:

$$Q_T \;=\; \frac{C_T \cdot I_T \cdot A}{3.6}$$

SDF peak flow

Where $Q_T$ is the peak discharge (m³/s) for return period $T$, $C_T$ is the SDF runoff coefficient for that return period, $I_T$ is the design rainfall intensity (mm/hr) at duration $T_c$ and return period $T$, and $A$ is the catchment area (km²).

The runoff coefficient $C_T$

$C_T$ is interpolated between the regional 2-year and 100-year calibration values using the log-normal frequency factor:

$$C_T \;=\; \frac{C_2 \;+\; (Y_T - Y_2)\,(C_{100} - C_2)/(Y_{100} - Y_2)}{100}$$

SDF runoff coefficient by return period

Where $Y_T$ is the standard normal variate (reduced log-normal deviate) for return period $T$: $Y_2 = 0.000$, $Y_{10} = 1.282$, $Y_{20} = 1.645$, $Y_{50} = 2.054$, $Y_{100} = 2.326$, $Y_{200} = 2.576$. $C_2$ and $C_{100}$ are tabulated by region (see table below). The /100 accounts for the coefficients being published as percentages.

Because $C_T$ grows with return period, the SDF explicitly captures the empirical observation that rare floods have higher runoff coefficients than frequent ones — soils are more likely to be saturated in the storms that produce the largest floods.

Design rainfall intensity $I_T$

The design rainfall intensity is read from a point depth-duration-frequency curve at duration $T_c$ (from the Tc Calculator) and return period $T$ (from Design Rainfall). An areal reduction factor is applied to point rainfall for catchments larger than roughly 30 km² to reflect the fact that point rainfall overstates the spatial average during a real storm.

SDF regions and parameters

South Africa is divided into 29 SDF regions based on clustering of gauge-derived flood statistics. The regions correspond approximately to the drainage regions of the original Kovács (1988) regional maximum flood analysis, but with additional subdivisions reflecting flood-frequency rather than flood-maximum behaviour.

The following table gives the SDF runoff coefficients for each region (representative values; the authoritative table is the SDF Regions Map appendix of Alexander, 2002):

Region	Description	$C_2$ (%)	$C_{100}$ (%)
1	Northern Cape arid interior	5	60
2	Upper Orange basin — semi-arid	10	65
3	Highveld grassland	15	60
4	Lowveld / eastern escarpment front	15	40
5	Zululand coastal	30	70
6	KZN midlands	20	50
7	Drakensberg foothills	25	55
8	Mpumalanga escarpment	15	40
9	Gauteng — urbanising	20	55
10	Bushveld / Limpopo	10	45
11	Waterberg	10	40
12	Soutpansberg	15	50
13	Limpopo valley	5	45
14	Northern Cape — lower Orange	5	55
15	Namaqualand coast	10	60
16	Karoo — southern	10	60
17	Karoo — central	10	55
18	Karoo — upper	10	55
19	Southern Cape coast (fynbos)	15	50
20	Tsitsikamma / Outeniqua	25	55
21	Eastern Cape — Algoa	20	55
22	Eastern Cape — coastal belt	25	60
23	Transkei coastal	25	60
24	Cape Winelands	15	50
25	West Coast	10	55
26	Boland mountain	25	55
27	Cape Flats / Cape Peninsula	15	45
28	Free State central	15	60
29	Northern Province transition	10	45

The values above are representative — always consult the authoritative regional map in SANRAL Drainage Manual Figure 3.13 (or Alexander 2002 Appendix A) for the definitive $C_2$ and $C_{100}$ for a given location. HydroDesign looks up the SDF region and its calibration parameters automatically from the catchment centroid.

Regional boundaries are approximate

SDF region boundaries are smoothed contours of the underlying flood-frequency statistics, not sharp lines. Catchments straddling two regions should be checked with both parameter sets, and if results differ markedly, the engineer should use judgement — often weighted by the proportion of area in each region — and should document the choice.

Runoff coefficient versus flood magnitude

A key conceptual feature of the SDF is the strong dependence of the runoff coefficient on return period:

Return period $T$	$Y_T$	$C_T$ (Region 1)	$C_T$ (Region 5)	$C_T$ (Region 14)
2 yr	0.000	5%	30%	5%
5 yr	0.842	25%	44%	23%
10 yr	1.282	35%	52%	33%
20 yr	1.645	44%	58%	41%
50 yr	2.054	54%	65%	49%
100 yr	2.326	60%	70%	55%
200 yr	2.576	66%	74%	60%

Note how Region 1 (Northern Cape arid) has a very large relative change from $C_2 = 5\%$ to $C_{100} = 60\%$ — a factor of 12 — reflecting the strongly non-linear rainfall-runoff behaviour of dry catchments. Region 5 (Zululand coastal) has a smaller relative change because the catchments are already wet at the start of most storms.

Worked example

A 220 km² catchment near Ermelo (Mpumalanga highveld — SDF Region 3) is analysed for the 1:50 year design flood for a national road culvert.

Given information

• Catchment area: $A = 220$ km²
• SDF region: 3 — Highveld grassland → $C_2 = 15\%$, $C_{100} = 60\%$
• Time of concentration: $T_c = 3.2$ hr (from Kirpich)
• Design rainfall (point, 1:50 yr, 3-hr): 85 mm → intensity 28.3 mm/hr
• Areal reduction factor (ARF) at A = 220 km², D = 3 hr: 0.85
• Adjusted design intensity: $I_{50} = 28.3 \times 0.85 = 24.1$ mm/hr
• Return period: $T = 50$ yr → $Y_{50} = 2.054$

Step 1 — Interpolate the runoff coefficient

$$C_{50} \;=\; \frac{C_2 + (Y_{50} - Y_2)(C_{100} - C_2)/(Y_{100} - Y_2)}{100}$$ $$C_{50} \;=\; \frac{15 + (2.054 - 0)(60 - 15)/(2.326 - 0)}{100} \;=\; \frac{15 + 2.054 \times 19.35}{100} \;=\; \frac{15 + 39.74}{100} \;=\; 0.547$$

Step 2 — Apply the SDF formula

$$\begin{aligned} Q_{50} &= \frac{C_{50} \cdot I_{50} \cdot A}{3.6} \\[6pt] &= \frac{0.547 \times 24.1 \times 220}{3.6} \\[6pt] &= \frac{2\,900}{3.6} \\[6pt] &= \boxed{805 \text{ m}^3/\text{s}} \end{aligned}$$

Step 3 — Cross-check against other methods

The 1:50 year peak discharge of ~805 m³/s should be compared against:

• The Rational Method (with an SA C₁ built from slope, permeability, and vegetation), ideally within ±40%.
• A Unit Hydrograph result from HRU 1/72 for the appropriate veld zone.
• The Regional Maximum Flood (RMF) envelope — the 1:50 yr should be a fraction of the RMF, typically 15 – 40%.

If the SDF result is the highest of the three, and the others differ by more than a factor of ~2, the engineer should investigate before adopting a design value.

Adopt the SDF as the design value for SANRAL projects

Current SANRAL policy (Drainage Manual 6th ed.) is to adopt the SDF as the design value for national road drainage wherever the catchment is within the calibrated range, with other methods as cross-checks. If the engineer elects to adopt a different method’s result, the rationale must be documented in the design report.

Limitations

• SA-specific. The SDF is calibrated only against South African flood data and should not be applied elsewhere.
• Catchment size range. Calibration is reliable for ~10 – 500 km². Outside this range the method is an extrapolation — use with extreme caution for A > 500 km² and not at all for A > 1000 km².
• No land-cover sensitivity. Unlike the Rational or SCS methods, SDF does not explicitly respond to land cover. Significant urbanisation or deforestation since the calibration period (~1990 and earlier) may not be reflected. Supplement with a CN-based analysis for urbanising catchments.
• Regional boundary sharpness. The regional $C_2$ and $C_{100}$ values change sharply across region boundaries. For catchments straddling a boundary, results can be ambiguous.
• No hydrograph. SDF gives peak only; where a hydrograph is needed, combine with a standard triangular UH scaled to match the SDF peak and runoff volume.
• Calibration age. The SDF calibration was performed on records available up to ~1990. Trends in rainfall or land cover since then are not reflected. This is a general limitation of regional flood frequency methods.
• Log-normal assumption. The $Y_T$ interpolation assumes a log-normal flood distribution. For regions where the true distribution is strongly skewed (e.g. arid regions with bimodal behaviour), very high return periods may be biased.

Related tools and guides

• Rational Method — SA Rational (C₁/C₂) for small catchments
• SCS Method — Curve Number based runoff estimation
• Unit Hydrograph — HRU 1/72 synthetic UH for SA
• DRH Method — full hydrograph with channel routing
• Tc Calculator — time of concentration (Kirpich, SANRAL, etc.)
• Design Rainfall — point rainfall depths and intensities

References

• Alexander, W.J.R. (1990). Flood hydrology for Southern Africa. South African National Committee on Large Dams (SANCOLD), Pretoria.
• Alexander, W.J.R. (2002). The Standard Design Flood — A new design philosophy. Department of Civil Engineering, University of Pretoria. (The original SDF publication, including the 29-region map and calibrated parameters.)
• Alexander, W.J.R. (2003). Flood Risk Reduction Measures — Incorporating the Standard Design Flood. Department of Civil Engineering, University of Pretoria.
• Van der Spuy, D. & Rademeyer, P.F. (2018). Flood Frequency Estimation Methods as Applied in the Department of Water and Sanitation (4th rev.). DWS, Pretoria.
• Kovács, Z. (1988). Regional Maximum Flood Peaks in Southern Africa. Technical Report TR 137. Department of Water Affairs, Pretoria.
• SANRAL. (2013). Drainage Manual (6th ed.). South African National Roads Agency, Pretoria. Chapter 3 — Hydrology; Section 3.7: Standard Design Flood.
• Smithers, J.C. (2012). Methods for design flood estimation in South Africa. Water SA, 38(4), 633 – 646.

---

Open Design Flood Estimation