The primary use of the Natural Resources Conservation Service (NRCS) curve number (CN) is to compute total storm runoff based on total rainfall. The method was originally created to determine the mean daily depth of runoff during flood producing events on small agricultural watersheds. CN values were determined using daily rainfall and runoff data. Practically, it did not rain for 24 hours during many, perhaps most, of the events, but since the data were recorded as daily rainfall, 24 hours became the implicit duration for values input to the curve number runoff model. NRCS references do not specifically state the CN applies only to the 24-hour storm. Even so, it may be inferred from what is published that the standard CN applies to the 24 hour duration storm.

Many methods and computer models used for the analysis and design of stormwater management systems incorporate the NRCS CN method. Because some designs and performance evaluations are based on rainfalls with durations less than 24 hours, there is the need for a method to modify CN values for shorter duration events. It goes against basic hydrologic principles if the same CN is used for storms of all durations. Not yet formally published, the NRCS recently developed a procedure to modify CN values for rainfall durations less than 24 hours. With encouragement from the NRCS, introducing that method to the engineering community is the goal for this paper.

The impact of adjusted CN values was demonstrated by calculations comparing runoff depths computed with standard and duration modified CN values for rainfalls of 1, 2, 3, 6, 12, and 24 hour duration. The standard CN significantly under-predicted runoff depths compared to the duration modified CN values. The differences increased with shorter duration storms.

The impact of adjusted CN values also was demonstrated during a forensic assessment of the performance of a stormwater detention pond in a residential subdivision. The pond was designed compliant with regulations to limit the post-development peak discharge rate at or below the pre-development peak runoff rate for 2- and 10-year frequency 24-hour design storm events. Even though the pond design met regulatory standards for 24-hour design storms, downstream flooding and sediment problems frequently occurred during short duration events. As part of the forensic study, runoff hydrographs were simulated for pre-development, construction phase, and post-development land use conditions for rainfalls of 1, 2, 3, 6, 12, and 24 hour duration. The simulation results for post development conditions showed successful pond performance for the 24-hour rainfall. However, the peak outflow rates for storms with durations less than 24-hours were greater than the 24-hour pre-development peak runoff rate.

The simulation results emphasize pond design calculations and decisions should include pond performance for events with duration less than 24 hours and should use duration modified CN values. It is recommended controlling regulations specify design events such as the 2- and 10-year 24-hour rainfalls, but include a mandatory check of other events, such as the 1, 2, 3, 6 and 12 hour events. Prudent and ethical practice suggests pond design be upgraded for the critical rainfall event.