The synthesis of daily river flow rates from daily rainfall records

Abstract

The work commences with an explanation of the need for synthesised riverflow values for reservoir yield calculations and a review of the factors which influence the rainfall-runoff relationship. This is followed by a critical resume of the modern literature on mathematical models of hydrologic systems which involve the use of a digital computer for runoff synthesis. The problem of synthesis from limited data is discussed and a line of investigation is proposed. The investigation is carried out using data from a catchment lying on the borders of Co. Durham and Northumberland. Details of this area and the recording instruments on the catchment are given. The actual daily riverflow values are first processed and then plotted in the form of a cumulative deficiency diagram. The storage conditions revealed by this detailed curve are then compared with published values obtained by traditional methods. The distribution of daily riverflow is investigated for various calendar groupings and conclusions drawn. The seasonal correlation of rainfall-runoff values, grouped according to antecedent precipitation index values is then performed. Since the store capacity of the computer available for the analysis was only 8k the data had to be read" in and out in groups and as a result the processing time was approximately 8 hours. An alternative method was devised, by which a large number of pieces of data could be held in a store of limited capacity. This system, which performed the analysis in less than a twentieth of the time, is then described. The effect of varying the recession factor through a range of values from 0.85 to 0.95, in steps of 0.01, is investigated. As a result of this investigation a further analysis is attempted. In this instance runoff on dry days is correlated with A.P.I, whilst runoff on rainfall days is correlated with daily rainfall in moving, fifty day, A.P.I, groupings. Daily runoff values are then synthesised, from the equations derived from the last analysis, and diagrams of cumulative excess of run off values are drawn. Finally, details are given for an alternative method of analysis using potential transpiration value in place of A.P.I. values for all but the winter months. The programmes written for this thesis are given in appendix I together with samples of the input data required to run them, and the output data derived from them. Copies of the fifty-three graphs drawn by the plotter are shown in appendix II.