depends on the precip depth and pattern, arial reduction, type of storm, watershed size, slope and routing. Also depends on upstream diversions, storage etc. Sometimes duration of flow and volume is more critical than peak flow rate.
There is no simple answer, except that generally for a smaller area, a shorter duration storm is more critical and for a larger area, a longer duration is worse.
I was going to post after I read the original post, but you sais it quicker and more clearly than I could. I guess ir depends on how much land the "general" owns or is downstream of.
The duration of storm necessary to produce the peak flow for any location in a drainage system is a period known as the time of concentration. This is the time taken for water to flow from the outermost point in the system to the subject location. Storms of a shorter duration (and higher intensity) may cause failure of part of the drainage system upstream of the subject location, but not at that location. Longer storms will not produce a flow in excess of the maximum peak flow of this critical duration storm unless there is a rainfall over a period of time equal to the critical duration within the longer storm that is more intense than the critical duration storm itself.
A peak rainfall event does not necessarily equate to the same flood magnitude. For example, if the storm occurs when the drainage area is dry, the flood magnitude may be considerably less than if the drainage area is saturated from previous storms.
Are we talking about a constant-intensity event (Rational method) or a variable intensity rainfall distribution (SCS method)?
bimr's reply is appropriate for the Rational method.
For the SCS/NRCS method, the peak RAINFALL intensity should be the same for the 2-hr and 24-hr events. This is because the 24-hr rainfall contains all shorter storms, including the 2-hr storm. And since the 24-hr storm will have a greater rainfall depth, the runoff volume and peak will be greater. This happens because the initial abstraction is satisfied earlier in the storm, resulting in more runoff when the peak hits.
Also, the Time of Concentration is a construct, having llittle to do with reality. Flow down a basin from a 3" rain will be faster than from a 1" rain.
Every basin is different, and another thing not considered is the orientation of the basin compared with the probable direction of travel of the storm. Thunderstorms traveling up-basin will not produce the flow of one traveling down-basin.
Psmart - are you saying that a 1 in 100 year 24 hour event is made up of incremental increases each equal to a 1 in 100 year return event. i.e that the 2 hour rainfall depth within a 1 in 100 year 24 hour event is also a 1 in 100 year event; similar for 3 hour 4 hour etc. ?
Bris - My comment refers specificially to the SCS/NRCS synthetic rainfall distributions. The middle hour of the 24-hour storm contains the 1-hour event, with the corresponding depth. The middle 2 hours contains the 2-hour event, etc.
RWF7437 - thanks for the link - I am not in the US and was not familiar with the SCS distribution.
Psmart - thanks for your response - The uniform hyetograph of the rational method and the SCS extreme peak hyetograph represent the two extremes of possible, but improbable storm profiles. The actual being somewhere in-between. I now understand and agree comment that using the SCS curve the 24 hour event will always give a higher runoff than a 2 hour event regardless of area. (But because the SCS hyetograph represents a much rarer event than 1 in 100 - not true in reality ?)
Bris, The exact SCS rainfall distrubtion will never occur, partly because no two storms will ever be exactly the same. However, when you use the the 100-year rainfall depth, you're modeling a 1% probability event.
