Design Flood Estimation in South Africa: Methods, DFET and the DWS Guide

A design flood is the peak flow, and sometimes the full hydrograph, that a structure or a floodline is designed for: the 1:50-year flood at a culvert, the 1:100-year flood for a floodline, a far rarer flood for a dam spillway. South Africa has no single method for estimating it. The Department of Water and Sanitation (DWS) opens the preface of its flood frequency guide with exactly that: “different flood frequency estimation methods will give different results and there is no single calculation method that can be presumed to be better than any of the other methods.”

In practice that means running several methods suited to the catchment, comparing them, and adopting a design flow with the reasons written down. This article covers the methods in use, the free tools DWS publishes, and how to choose between results.

Downloads: DFET and the DWS guide

The DWS National Flood Studies Programme publishes the three documents most practitioners start from. All three download from the DWS website without registration.

• Design Flood Estimation Tool (DFET) Version 1.4: a macro-enabled Excel workbook (.xlsm, about 10 MB). It is also listed on the DFET page of the DWS website.
• DFET User Manual, Version 1.4: inputs, checks and equations for every worksheet (PDF, about 3.6 MB).
• A Guide to Flood Frequency Estimation Methods as applied in the DWS (July 2025): the eleventh edition of the DWS methods guide (PDF, about 2.5 MB).

For urban catchments, the Water Research Commission and IMESA’s Best Practice Guideline for Design Flood Estimation in Municipal Areas in South Africa (WRC Report TT 921/23, July 2023) covers design flood estimation and water surface profiles for municipal work.

The three families of methods

The DWS guide groups South African flood estimation into three approaches:

• Statistical methods fit a probability distribution to the annual flood peaks recorded at a flow gauge. They need a usable record at or near the site.
• Deterministic methods convert storm rainfall into a flood peak, assuming the flood has the same return period as the storm that causes it. They are the starting point at ungauged sites, which is most sites. Every one of them depends on the design rainfall and the catchment’s time of concentration.
• Empirical methods are relationships fitted to observed floods, used mainly as a quick check on the others.

The Regional Maximum Flood (RMF) sits outside all three. It is an envelope drawn over the largest floods recorded in a region, has no return period, and is used to judge whether other estimates are plausible and in dam safety work (see PMF vs RMF).

Methods in the DWS guide and their catchment-area ranges

Method	Family	Catchment area stated in the DWS guide
Rational method	Deterministic	Generally under 15 km²; used on larger catchments by experienced users
SCS-SA	Deterministic	Small to medium catchments
Synthetic unit hydrograph (HRU 1/72)	Deterministic	20 to 10 000 km²
Direct runoff hydrograph (Bauer and Midgley, 1974)	Deterministic	20 to 20 000 km²
MIPI-67 (Midgley and Pitman, 1967)	Empirical	100 to 200 000 km²
MIPI-71 (Midgley and Pitman, 1971)	Empirical	20 to 100 000 km²
CAPA (catchment parameter method)	Empirical	10 to 100 000 km²
Frequency analysis of annual flood peaks	Statistical	Wherever a flow record of adequate length exists
Regional Maximum Flood	Envelope	Any size; no return period

The ranges are guidance. A method applied well outside its range deserves less weight, and the report should say so.

The SANRAL Drainage Manual, the usual reference for road drainage, covers an overlapping set and adds the Standard Design Flood (SDF): a Rational-type method (Alexander, 2002) whose runoff coefficient rises with return period and is calibrated for each SDF drainage basin.

What changed in the July 2025 edition

The eleventh edition (July 2025, version 2) is by P Rademeyer, E Oakes and J Nathanael of DWS and Dr D van der Spuy. Its edition notes list five additions:

• Time to peak (tP) as an alternative to time of concentration when setting the critical storm duration, using equations by Gericke and Smithers (2016) and Gericke et al. (2024).
• A new approach to areal reduction factors (ARFs) from Gericke et al. (2023): regionalised ARFs that vary with return period, with a web-based tool covering five ARF regions. Under the traditional approach, DWS uses the UK Flood Studies Report ARFs, which it found more applicable in South Africa than the HRU curves.
• The Z-set plotting position, which DWS now uses as an additional technique when a flood record contains outliers. Standard plotting positions cannot give an outlier a realistic probability.
• The IPZA flood quantile technique (Van der Spuy and Du Plessis, 2024), a method-of-moments approach that the guide finds much less sensitive to high outliers than the GEV distribution.
• New RMF regions from Du Plessis and Masule (2023), alongside the Kovács (1988) regions.

Two further points matter for design work:

• DWS no longer uses TR102 storm rainfall or the HRU depth-duration-frequency graph. It derives catchment rainfall from station records and from the Smithers and Schulze regional rainfall, produced with the Design Rainfall Estimation software.
• For flood records, the guide recommends fitting the LN, LP3, GEV and IPZA and choosing the most appropriate fit, and it prefers a site-specific analysis to a regional one where possible.

The Design Flood Estimation Tool (DFET)

DFET is an Excel workbook, programmed in VBA, that puts the recognised South African methods side by side for one site. It was developed by OJ Gericke and JA du Plessis (Central University of Technology, Free State, and Stellenbosch University), copyright 2019 Stellenbosch University. Version 1.4 is the release DWS hosts; its manual was issued for the Stellenbosch University flood hydrology course in May 2023.

Worksheets	What they do
Catchment information	Drainage region, area split (rural, urban, lakes, dolomite), rainfall region, natural and artificial flow paths
Design rainfall	TR102 or RLMA-SAWS station depths (arithmetic mean or Thiessen polygons), or 1′ × 1′ RLMA&SI grid depths from the Design Rainfall Estimation software
Slopes	Average catchment slope (grid, empirical or DEM value) and main watercourse slope (10-85, Taylor-Schwarz or equal-area), with a longitudinal profile plot
Deterministic methods	Rational, Alternative Rational, SCS, Standard Design Flood, synthetic unit hydrograph, lag-routed hydrograph
Empirical methods	MIPI, CAPA and the RMF
Probabilistic methods	Annual maximum and partial duration series; normal, extreme value, log-normal, log-EV1 and LP3 distributions by the method of moments; generalised logistic by L-moments; the square-root-area method to combine records from nearby gauges; probability plots
Summary	Every method’s results in one table and one chart

Getting DFET running

1. Download the workbook and the manual.
2. Open the workbook in Excel 2007 or later on Windows. DFET does not work in Office 2003.
3. Enable macros. Current versions of Excel block macros in files downloaded from the internet: if a “macros have been blocked” banner appears, close Excel, right-click the file, choose Properties, tick Unblock and reopen it.
4. Save a copy under a new name before entering data, as the manual advises, so the original stays a clean template.
5. Enter data only in the light-green input cells, and clear the previous site’s entries before starting a new one.

What DFET leaves to you

DFET calculates; it does not measure. Catchment area, longest watercourse, slopes, centroid distance, land cover and soils come from your own GIS work or maps. Rainfall stations are entered by SAWS station number, and RLMA&SI grid depths are pasted in from the Design Rainfall Estimation software’s output. Two people running DFET for the same site can therefore get different answers from different inputs, so record where each input came from.

The manual’s disclaimer states that Stellenbosch University and the developer accept no liability for errors or omissions. The workbook is a calculation aid; the engineer who signs the report owns the adopted flow.

Choosing a design flood

A defensible design flow usually comes out of this sequence:

1. Delineate the catchment and derive its parameters from a hydrologically corrected DEM.
2. Set the design rainfall for the critical storm duration (time of concentration or, now, time to peak).
3. Run every method whose range covers the catchment, plus the empirical methods as a check.
4. Where a flow gauge exists on the river, run a frequency analysis and weight it according to the record’s length and quality.
5. Compare the results. Agreement between methods built on different data is the evidence for a design value. A method far from the rest needs an explanation: urban cover, a dam upstream, dolomite, or a catchment outside the method’s range.
6. Check the result against the RMF, adopt a flow and record why.

Return period choice is a separate decision, set by the structure and the approving authority; return periods explained covers what a 1:100-year flood does and does not mean.

Running the methods in Delta Hydro Studio

DFET needs the catchment parameters prepared by hand. Delta Hydro Studio, our engineering platform, starts from an outlet point instead. Its Design Flood Estimation tool delineates the catchment on a hydrologically corrected Copernicus GLO-30 DEM, derives the area, 10-85 slope, time of concentration, MAP, land cover, soils, veld zone and regional classifications, pulls the design rainfall and runs up to eleven methods for the 1:2 to 1:200-year floods. You confirm the delineated catchment before the run, and every method lands in one table and one chart.

Besides the Drainage Manual methods, Studio runs the NaFSAT 2.0 regional methods (Index Flood, PRM and QRT; Calitz and Smithers, 2020), which DFET does not include. It covers sites in South Africa only. The Design Flood Estimation tool page lists every method, its source and the area limits Studio applies.

Frequently asked questions

Is DFET free?

Yes. The workbook and its manual download from the DWS website without registration or payment.

Which version of DFET is current?

Version 1.4 is the version on the DWS DFET page as of October 2026. The manual on the same page was issued for the May 2023 Stellenbosch University flood hydrology course.

Does the DWS guide prescribe a method?

No. The guide describes itself as basic information on the methods DWS uses, not a handbook or manual, and not an attempt to prescribe one method over another. Choosing the method is left to the analyst’s judgement.

Where does the design rainfall for DFET come from?

From the TR102 or RLMA-SAWS station databases built into the workbook, or from the 1′ × 1′ RLMA&SI grid produced by the Design Rainfall Estimation in South Africa software. The design rainfall article covers where to download that software and how to use its output.

Do I need a flow gauge?

No. Deterministic and empirical methods work without one. Where a gauge with a usable record exists on the river, the DWS guide recommends a site-specific frequency analysis.

Design floods for culverts, bridges, dams and floodlines are the core of our hydrological studies. If a municipality, SANRAL or DWS will review the number, we can produce it with the method comparison documented.