Generation of natural daily flow time-series in regulated rivers using a non-linear spatial interpolation technique

Daily streamflow time-series data are often in demand in water resource assessment, water quality and river ecology studies. Such studies normally require daily time-series representing the natural conditions in a catchment. The generation of these time-series by means of deterministic physically based daily models is a time consuming approach. This paper describes more pragmatic alternative techniques for generating daily flow sequences using the available observed daily flow records at the neighbouring gauging stations. The daily time-series representing natural flow conditions at any site of interest may be generated by means of a non-linear spatial interpolation technique, which is based on 1-day long-term flow duration curves for each calendar month of the year. This technique is used to transfer the daily flow information from a source site(s) with data to the destination site where these discharge data are either insufficient or completely missing. The transfer process is based on the assumption that daily flows occurring simultaneously at sites located close to each other correspond to similar probabilities on their respective flow duration curves. Several methods by which to establish the required 1-day long-term flow duration curves representing the natural flow regime at the destination site are presented and the conversion of these curves into a complete continuous time-series of daily discharges is illustrated with the examples from modified streams in South Africa. Copyright © 1999 John Wiley & Sons, Ltd.