Abstract:
The two-parameter weibull distribution is presented as an alternate for developing synthetic hydrographs after a review of the existing applications like the popular gamma distribution method. The suitability of weibull distribution as Synthetic Unit Hydrographs (SUH) is explored for two cases using field data (i) for limited data conditions, and (ii) for no data condition, i.e. ungauged catchment. To estimate the parameters of the weibull distribution, an analytical procedure is proposed instead of existing numerical solutions. Taking the data of fifty-six catchments from India, equations for peak flow and time-to-
peak are developed using two new methods: (i) non-linear optimization, and (ii) artificial neural network. Both these methods use a set of non-dimensional groups formed out of geomorphological variables as inputs. The workability of the proposed method is checked using a test catchment. The weibull distribution parameters are related to statistical properties of the hydrograph considering an equivalent triangular hydrograph, which eventually helps in analyzing the parameter sensitivity. The results show that the weibull distribution is flexible and more adaptable to derive SUH in ungauged catchments.
