CELMOD5 - a semi-distributed cell model for conversion of rainfall into runoff in semi-arid watersheds Academic Article uri icon


  • This paper introduces the general outline of CELMOD5, a parametric, semi-distributed, quasi-linear model, for conversion of rainfall into surface runoff. The model considers the watershed as a series of interconnected cell units, each representing a specific portion of the area of the watershed. In contrast to grid models using a large number of rectangular elements or cells, the number of cells in CELMOD is relatively small and their boundaries are chosen according to the watershed topography. For each cell, the program computes the surface runoff hydrograph at the cell outlet, related to a specified record of total rainfall data at a number of rain gages. If measured surface runoff data are available for some locations in the watersheds, the program can compare these data with computed values of surface runoff at the corresponding points of the cell model. Detailed descriptions are provided for the main model procedures — computation of rainfall excess, conversion of rainfall excess into surface outflow, routing the channel inflow and subtraction of channel losses. Special attention is given in this model to the specific conditions of arid or semi-arid watersheds. This paper is also concerned with a technique for calibrating and testing a forecasting model of storm hydrographs with emphasis on two objective functions — runoff volume and peak discharge. A method for evaluation procedure is presented based on the following five steps: trial and error calibration; sensitivity analysis; bilinear interpolation optimization; testing the model on different storm events; testing the model on a different watershed. Results are presented for all the larger storm events with reliable data during 12 years in two subwatersheds of the Walnut Gulch Experimental Watershed in southeastern Arizona. The evaluation procedure is demonstrated for one particular rainfall-runoff event.

publication date

  • January 1, 1994