TR-44 : Application of Kinematic cascade model 'Kingen' for flow computation in a hilly catchment

Abstract

Estimation of Runoff in mountainous areas is required for the design, development and management of water resources projects. Runoff in mountainous areas result from precipitation in the form of rainfall or snowfall. The response of watershed to the precipitation may be conceptualised into three components; viz. infiltration into the soil; overland flow and channel flow; collectively called as surface runoff.

Surface runoff from a small catchment is predominantly overland flow with channel flow being comparatively less significant. The overland flow from mountainous areas is generally recognised as non-linear process. Among the two approaches of modelling non-linear processes viz. system approach and hydro-dynamic approach; it is usually the later one which has mostly been used for modelling overland flows.

Based on hydrodynamic approach, a kinematic cascade model namely KINGEN has been developed at Colorado State University. The model incorporates infiltration model to compute excess rainfall. The model requires simplification of the watershed into a cascade of planes and channel elements. The input data to the model include details of various elements; such as length, width, slope, etc. roughness parameters, in the form of Chezy’s C or Manning's n and soil related data for infiltration calculations. Computed excess rainfall is routed over the planes and in the channels upto the outlet of the catchment to yield surface runoff hydrograph. The rainfall is assumed to be uniformly distributed over the entire catchment for the purpose.

For the present study, the catchment of Malaprabha upto Khanapur in Western Ghats region has been selected. The watershed geometry has been represented by a cascade of fourteen elements. Infiltration losses were computed using Smith's theory of soil water flow. The kinematic cascade model KINGEN based on kinematic wave equation has been used to simulate overland flow and channel flow.

Using data of some storm spells in the catchment, the model has been calibrated. The calibrated model was verified with data for the other storm events. The direct-surface runoff hydrograph simulated by the model was compared with the observed runoff after separating base-flow. Peak flow and time to peak were simulated fairly accurately. The volume of the simulated hydrograph was found to be in agreement with the observed flow volume.