Why do I get different peak flows when using different time increments?
Should this be a USER input? -- I don't see it in other programs.
At three minutes, your peak value falls between two time increments. At one minute you capture that value. This is much more prevalent when your Tc is short and you use the SCS Curve.
Mind you, SCS Curves are defined in 6 minute increments.
Problem with variation appears to be something more than interpolation between time increments and short Tc.
Compared hydrographs -- increasing Tc=21 minutes and time intervals of 1 and 3 minutes, holding Area, CN, storm type constant.
Flow (1min Time Incr)
Flow (3 min Time Incr)
The peak flows were both at 12.10 hours but the flows with with the 3 minute increment were consistently lower by about 10%.
Alternate scenarios varying Tc from 6 to 21 minutes for 1 and 3 minute Time Intervals—Peak Flows still had considerable variation as Tc increased.
Peak Flow (1min Time Incr)
Peak Flow (3 min Time Incr)
Important: A unit hydrograph is not the final runoff hydrograph. It only reflects the watershed characteristics and geologic factors.
With a known unit hydrograph of a watershed, you can apply any design storm and compute the final runoff hydrograph. Many hydrologists use the SCS 24-hour storms, but any storm of any length can be used with the unit hydrograph method. The Bulletin 71, Huff, and Hydraflow Hydrographs Extension synthetic distributions are other examples.
Hydraflow Hydrographs Extension computes SCS runoff hydrographs by convoluting rainfall hyetograph through a unit hydrograph. This method is also used in SCS TR-20. Convolution is known as linear superpositioning, and means that each ordinate of the rainfall hyetograph is multiplied by each ordinate of the unit hydrograph, thus creating a series of smaller hydrographs. These hydrographs are then summed to form the final runoff hydrograph. For example, if the rainfall hyetograph (design storm) contained 1,440 ordinates and the unit hydrograph contained 30, then a total of 43,200 ordinates (1,440 x 30) would have to be computed before reaching the final hydrograph. This method is highly accurate, but not very practical without the use of a computer.
Related: Many projects contain small watersheds with short times of concentration (Tc). Since the time interval (unit duration) of the unit hydrograph should be 0.133(Tc), a great number of ordinates will be required to perform the convolution when modeling long design storms such as the SCS 24-hour storms. A drainage basin with Tc of 30 minutes requires a time interval of four minutes, 0.1333(30). The resulting hydrograph, when using a 24-hour storm, needs at least 360 ordinates (1,440 minutes / 4). Hydraflow Hydrographs Extension can use up to 2,880 points for each hydrograph, allowing you to model very small watersheds (with Tc as low as two minutes) while using a 24-hour storm with accuracy that is difficult to obtain by shortcut methods. Hydraflow Hydrographs Extension creates its design storms compatible with the selected time interval of the final hydrograph. This enables you to achieve the highest degree of accuracy and eliminates any interpolation between fixed ordinates or the need to convert the final hydrograph to match the time interval.
Matt, how does this work in SSA? When looking at peak flows of basins only using TR-20, the results in SSA are substantially different when I've tested it. I have meant to dive into this for a while, just haven't had the time to. I know the intervals for the rainfall hyetograph are fixed at 6 minutes for 24 hour type II storms, and the routing interval is one second, but guessing this is for application of St. Venant equations - is the TR-20 method of constructing hydrographs different in SSA?
Log into access your profile, ask and answer questions, share ideas and more. Haven't signed up yet? Register