Rainfall

Design rainfall and IDF curves

Design rainfall is the depth of rain expected for a given storm duration and return period at a site. This tool returns the depth-duration-frequency table from NOAA Atlas 14 in the United States, the Bureau of Meteorology IFD in Australia and the K5/1060 grid in South Africa, or from your own table anywhere else, and fits IDF curves to it.

Data sources and coverage

Design rainfall is estimated from long rain gauge records, regionalised and mapped by a national study. The tool reads the source that covers the site, so a depth for Washington, DC and a depth for Durban come from different studies:

SourceCoverageDurationsFrequenciesUncertainty
NOAA Atlas 14United States and territories5 min to 60 days1 to 1000 years ARIUpper and lower confidence limits
BoM IFD (2016 design rainfalls, ARR 2019)Australia1 min to 7 days12 EY to 1 in 2000 AEPBest estimate only
K5/1060 grid (Smithers and Schulze 2003)South Africa, 1-arcminute grid of 421 027 points23 durations, 5 min to 7 days1:2 to 1:1000 yearsLower, mean and upper bounds
User-definedAnywhereAt least 3 durations per return periodAny you enterAs supplied

For the US and Australia the tool fetches the published table for the clicked point. For South Africa it computes the table at the nearest grid point. A user-defined table can be entered as depth or intensity, in millimetres or inches.

How the South African K5/1060 method works

The South African grid uses a regional L-moment method. The country is divided into 78 growth-curve clusters, and the design depth for return period TT and duration DD is:

PT,D=GCT⋅L1,DP_{T,D} = GC_T \cdot L_{1,D}
SymbolMeaningUnits
PT,DP_{T,D}Design rainfall depthmm
GCTGC_TDimensionless growth-curve factor for the cluster and return period–
L1,DL_{1,D}At-site mean (first L-moment) of the annual maximum series for duration DDmm

L1L_1 is estimated by regression at anchor durations (5 min, 15 min, 2 h and 24 h for sub-daily storms; 1, 3 and 7 days for multi-day storms) and interpolated between them in log-log space, with the slope held non-negative so depth never falls as duration rises. The lower and upper bounds combine the uncertainty in L1L_1 with the cluster’s lower and upper growth curves.

How the IDF curves are fitted

Any of the four sources can be turned into IDF curves. The default is the Sherman (also called Chen) equation, fitted separately for each return period:

i=a(d+b)ci = \frac{a}{(d + b)^c}
SymbolMeaningUnits
iiRainfall intensitymm/h
ddStorm durationmin
aa, bb, ccFitted coefficients for one return periodaa in mm h−1 minc\text{mm h}^{-1}\,\text{min}^{c}, bb in min, cc dimensionless

The fit is made in log space by least squares, with a grid search on bb, and the tool warns when R² falls below 0.95. For the NOAA Atlas 14, BoM IFD and K5/1060 sources only durations up to 24 hours are fitted; the multi-day depths stay in the depth table, and the fitted intensity table stops at 24 hours. A user-defined table is fitted over every duration you enter. Three other forms are available: Bernard, i=a/dci = a/d^c; Talbot, i=a/(d+b)i = a/(d+b); and Koutsoyiannis, i=λTκ/(d+θ)ηi = \lambda T^{\kappa}/(d+\theta)^{\eta}, which fits one equation across all return periods together.

Worked example: NOAA Atlas 14, Washington, DC

At 38.8895° N, 77.0353° W (the National Mall), NOAA Atlas 14 Volume 2 gives these 1:10-year partial-duration depths. Intensity is depth divided by duration in hours:

DurationDepth (in)Depth (mm)Intensity (mm/h)Sherman fit (mm/h)
5 min0.56514.4172.2173.2
15 min1.1429.0115.8116.8
30 min1.6642.284.381.2
1 h2.1654.954.952.5
2 h2.5364.332.132.5
6 h3.3083.814.014.4
24 h4.81122.25.15.0

Fitted to the 1:2- to 1:100-year tables at the Atlas 14 durations from 5 minutes to 24 hours, the 1:10-year Sherman coefficients are aa = 1448.0, bb = 10.22 min and cc = 0.780, with R² = 0.9994 in log space. The 1:100-year curve is aa = 1666.1, bb = 10.90, cc = 0.723.

For a catchment with a 25-minute time of concentration, the fitted 1:10-year intensity is 1448.0/(25+10.22)0.7801448.0 / (25 + 10.22)^{0.780} = 90.0 mm/h, a depth of 37.5 mm. Interpolating the published depths in log-log space between 15 min (29.0 mm) and 30 min (42.2 mm) gives 38.2 mm, or 91.7 mm/h. The fitted curve runs 1.8% below the published table at 25 minutes and 4.2% below it at an hour; over all six return periods and every duration up to 24 hours it stays within 6.1% of the table.

The tool fits only durations up to 24 hours. A three-parameter curve cannot follow a table from 5 minutes to 60 days: fitted to every Atlas 14 duration, the same 1:10-year curve reads 5% low at 25 minutes and 8.7% low at an hour, and the 1:2-year curve 10.9% low at 30 minutes. Even within 24 hours the fit is a smoothing: for a single design duration, the published depth is the value to use, and the fitted equation is for the durations in between.

NOAA Atlas 14 outside the US: what to use instead

NOAA Atlas 14 is a statistical analysis of US gauge records. Its values stop at the US border, and a value borrowed from a “similar” US climate is not defensible in a design report. Outside the US:

  • Australia: the Bureau of Meteorology 2016 IFD design rainfalls, built into the tool.
  • South Africa: the K5/1060 grid, built into the tool, with the nearest Design Rainfall Estimation stations listed for comparison.
  • Other countries: use the national standard and enter its depths or intensities as a user-defined table. Examples are the Flood Estimation Handbook rainfall model in the UK, the Engineering Climate Datasets IDF files from Environment and Climate Change Canada, and HIRDS from NIWA in New Zealand.
  • No national product: extract and fit the annual maximum daily series from a long gauge record with the daily rainfall statistics tool or flood frequency analysis. That gives daily depths at one point; treat any sub-daily depths derived from them with caution.

AEP, EY and ARI conversion

Australian practice since ARR 2016 uses annual exceedance probability (AEP) and exceedances per year (EY); US and South African practice mostly uses return period. The exact conversion is ARI=−1/ln⁡(1−AEP)\text{ARI} = -1/\ln(1 - \text{AEP}):

EYAEPAEP (1 in x)ARI (years)
163.2%1.581.00
0.6950%21.44
0.539.35%2.542.00
0.2220%54.48
0.218.13%5.525.00
0.1110%109.49
0.055%2019.5
0.022%5049.5
0.011%10099.5

Below 10% AEP the two are close enough to treat as reciprocal. Above it they are not: a 50% AEP depth is the 1.44-year ARI depth, not the 2-year ARI depth.

In Studio a return period of 1:TT means 1 in TT AEP, the third column, as the K5/1060 annual-maximum estimates are defined. The BoM columns are labelled the same way: 50% AEP is 1:2, 0.5 EY is 1:2.54 and 0.2 EY is 1:5.52, while 2 EY and more frequent columns keep their EY labels. NOAA Atlas 14 is read as NOAA’s partial-duration series, so its 1- to 1000-year columns are ARIs, as NOAA publishes them.

Inputs and checks

  • Location. Click the map or enter coordinates. For South Africa, check the reported MAP and altitude of the grid point, because depths can change over a few kilometres in mountainous terrain. For a catchment, the South African source can also average the grid over a catchment polygon.
  • Return periods and durations. Pick the durations around your time of concentration from the time of concentration calculator, and the return periods your design standard requires.
  • Bound. Use the mean or median for ordinary design. Consider the upper bound for spillways and high-consequence structures, where the cost of underestimating is large.

Limitations

  • Design rainfall is a point value. Catchments larger than a few square kilometres need an areal reduction factor.
  • The estimates assume stationary climate statistics. A climate change uplift can be applied, but the factor has to come from your local guidance.
  • South African values are for the nearest 1-arcminute grid point, and sub-daily values carry more uncertainty than daily ones because fewer recording gauges stood behind them.
  • BoM IFD rare-event depths for short durations are derived from daily data by ratios, and BoM publishes no confidence bounds.
  • A fitted IDF equation smooths the published table. Check it against the table at the durations you design for, as in the example above.
  • Depths rarer than 1:200 years rest on records much shorter than the return period. Treat them with caution; spillway design floods need probable maximum precipitation or a dedicated study.

What the full Studio tool adds

The Studio tool shows the full depth table and the IDF and depth-duration charts, with lower, mean and upper values where the source provides them. You can switch between Sherman, Bernard, Talbot and Koutsoyiannis fits, apply a climate change uplift as one factor or by duration or return period, and estimate the return period of an observed storm from its depths. Results save to a project, export to CSV, Word and Excel, and feed the rational method calculator and the design storm generator.

FAQ

Common questions

What is the difference between design rainfall and an IDF curve?
Design rainfall is the depth, in millimetres or inches, expected for one duration and one return period. An IDF curve plots the same information as intensity (depth divided by duration) against duration, one curve per return period, usually as a fitted equation such as i = a/(d + b)^c. The table is the published data; the curve is a smoothed form of it that is convenient for any duration.
Can I use NOAA Atlas 14 outside the United States?
No. NOAA Atlas 14 is fitted to US rain gauge records and covers only US states and territories; there is no valid Atlas 14 value for a point in another country. Use the national equivalent instead: the Bureau of Meteorology IFD in Australia or the K5/1060 grid in South Africa, both built into this tool. Elsewhere, take the published national IDF values and enter them as a user-defined table.
How do I read rainfall intensity for the rational method?
Set the storm duration equal to the time of concentration, read the depth for that duration and return period, and divide by the duration in hours. Where the time of concentration falls between tabulated durations, interpolate in log-log space or use the fitted IDF equation, and check the result against the published depths on either side.
What is the difference between AEP and return period?
Annual exceedance probability (AEP) is the chance that a depth is exceeded in any one year. The average recurrence interval (ARI), or return period, is the average time between exceedances. They are related by ARI = −1/ln(1 − AEP). For rare events they are nearly reciprocal, so 1% AEP is about 100 years, but 50% AEP is 1.44 years, not 2.
Which IDF equation should I fit?
The Sherman (Chen) form i = a/(d + b)^c is the default and fits most data well because the b term flattens the curve at short durations. Bernard (i = a/d^c) and Talbot (i = a/(d + b)) have fewer parameters. Koutsoyiannis fits one equation across all return periods together. Use the form your local design standard prescribes, and accept a fit only if R² is at least 0.95.
Full documentation: Design Rainfall
Every input, option and equation in the Studio tool, with its limitations.
Run it from your AI agent
This engine is open to Claude, ChatGPT, Cursor and other MCP clients as compute_design_rainfall. The agent passes the inputs, the engine does the calculation, and the record is saved to your project for you to review.

References

  1. Smithers, J.C. and Schulze, R.E. (2003) Design Rainfall and Flood Estimation in South Africa. WRC Report No. K5/1060. Water Research Commission, Pretoria.
  2. Bonnin, G.M., Martin, D., Lin, B., Parzybok, T., Yekta, M. and Riley, D. (2004) Precipitation-Frequency Atlas of the United States, NOAA Atlas 14, Volume 2, Version 3.0. NOAA National Weather Service, Silver Spring, Maryland.
  3. Ball, J., Babister, M., Nathan, R., Weeks, W., Weinmann, E., Retallick, M. and Testoni, I. (eds) (2019) Australian Rainfall and Runoff: A Guide to Flood Estimation. Commonwealth of Australia.
  4. Green, J., Xuereb, K., Johnson, F., Moore, G. and The, C. (2015) The revised intensity-frequency-duration (IFD) design rainfall estimates for Australia: an overview. 36th Hydrology and Water Resources Symposium, Hobart.
  5. Hosking, J.R.M. and Wallis, J.R. (1997) Regional Frequency Analysis: An Approach Based on L-Moments. Cambridge University Press.
  6. McCuen, R.H., Johnson, P.A. and Ragan, R.M. (2002) Highway Hydrology, Hydraulic Design Series No. 2 (HDS-2), 2nd ed. FHWA-NHI-02-001. Federal Highway Administration, Washington DC.
  7. Koutsoyiannis, D., Kozonis, D. and Manetas, A. (1998) A mathematical framework for studying rainfall intensity-duration-frequency relationships. Journal of Hydrology, 206, 118–135.

Last reviewed by the Delta Hydro engineering team.

Keep the whole study in one place

Delta Hydro Studio puts this tool alongside catchment, rainfall, runoff and hydraulics tools, with projects, team sharing and Word and Excel reports. Run them by hand, or let your AI agent run them for you to review.