Rainfall Frequency Not Equal to Runoff Frequency

It is quite common, in California , to hear radio/TV commentators say "watershed conditions are saturated from recent rains, so additional rainfall will cause substantial runoff". Early in the winter rainy season in California , rainfall may cause very limited runoff. If a given rainfall can cause different amounts of runoff then rainfall frequency and runoff frequency cannot always be the same: 100-yr recurrence interval rainfall doesn't always cause a 100-yr peak flow.

What does cause the 100-yr peak flows? Would use of a "runoff coefficient" from an published table make runoff frequency equal to rainfall frequency? Where do the runoff coefficient values in an "published tables" come from?

The questions above are certainly not new. A Stanford Ph. D thesis by Curtis Larson {A Two Phase Approach to the Prediction of Peak Rates of Runoff from Small Watersheds, Tech. Report No. 53, Dept. of Civil Eng. , Stanford University } explored the hydrologic processes that runoff coefficients represent. More recently Laura Marino and Allen Bradley developed a very useful method for calculating the relationship between rainfall and runoff frequency {Precipitation Frequency-Runoff Frequency Relationship in Hydrologic Design, AGU Fall Annual Meeting, San Francisco , 1986}.

Are rainfall frequencies and runoff frequencies related? What is the influence of “watershed state”, or soil moistures as a storm begins?

There is a small stream near Stanford University called Redwood Creek (USGS 11162800, 1.82 sq. mi.) that participants in our Simulation Workshops calibrate. The watershed is in the foothills of the Coast Range , and has rolling, moderate topography. Much of the watershed has low density suburban development, and there is a golf course in the lower basin near the stream gage.

The following "design storm" sequence of hourly rainfall occurs on Redwood Creek.


Rainfall (Inches/Hour)














This storm contains a 1/100 year one hour rainfall in the 5th hour, and the total rainfall over six hours is the 1/100 year six hour rainfall. The hourly pattern in the storm is typical of "cold front" rainfall in Redwood Creek, although the rainfall intensities are very high.

If the above storm had occurred on the following historic dates, the following flood peaks would have been observed in the 1982 water year.


Peak Flow (cfs)

September 15, 1981


January 15, 1982


March 15, 1982



 A ‘100-yr’ rain storm creates peak flows that differ by an order of magnitude!


1. The last rainfall in Redwood Creek prior to September 15, 1981 was on May 20th, 1981.

2. The recorded peak of record on Redwood Creek was 644 cfs on Jan 31, 1963. The period of record at the gage is 1959 to 1997.

3. Water year 1982 was relatively wet in Redwood Creek. The total precipitation on the watershed that year was 41.3 inches. The mean annual precipitation is 25.9 inches. The maximum recorded peak flow during 1982 was 379 cfs on January 4th.

Back to Index

Hydrocomp, Inc.