UM-8 : Unit Hydrograph Derivation

Abstract

The unit hydrograph technique is a simple tool being used by most of the water resources development organisations for the estimation of flood flows. Although number of techniques are available for finding the unit hydrograph, but all of them have some advantages as well as limitations. The technique adopted for the derivation of unit hydrograph in this user manual is based on Nash's approach as it gives non-oscillating and physically realizable unit hydrograph.

Nash considered that the Instantaneous Unit Hydrograph (IUH) can be obtained by routing the unit impulse input through a cascade of N linear reservoirs of equal storage coefficient K. Thus, the two parameters, N and K, determine the shape of the unit hydrograph. These parameters are estimated in the user manual by (i) method of moment, and (ii) optimization procedure. The former method requires the first and second moments of input ( excess rainfall) as well as output ( direct surface runoff), which are used in solving the moment expressions to get parameters N and K. However, in the latter approach the set of parameters, N and K, are estimated minimizing the objective function i.e. the sum of squares of the differences between ordinates of observed and computed hydrographs using Quasi Newton optimization procedure. The computer programmes NASH.FOR and CONTI.FOR, based on the above two methods have been developed at the National Institute of Hydrology, Roorkee and have been tested on VAX-11/780 system for the test input. The programmes require the stations rainfall along with their Thies sen weights and observed discharge hydrographs for the isolated events as input. The data interval for the observed rainfall hyetograph and observed discharge hydrograph should be the same. The programmes give the following outputs; the base flow hydrograph, direct surface runoff hydrograph, excess rainfall hyetograph, the parameters N and K, IUH and UH (Unit Hydrograph) ordinates and computed discharge hydrograph. The programme also estimates the error functions such as: (i) model efficiency, (ii) average standard error, (iii) average absolute error, (iv) average percentage absolute error, (v) percentage absolute error in peak, and (vi) percentage absolute error in time to peak. The input and output specifications for the programmes have also been described. The programme scan be run on computers other than VAX-11/780 system having FORTRAN compiler, after making suitable modifications.