Extension of runoff series using empirical orthogonal functions

Open Access
Publisher Website
Google Scholar
Abstract
Methods for extrapolating runoff series are important in order to gain maximum information from hydrological observations. The main disadvantages of traditional methods are that they are subjective and often time consuming. In this paper, empirical orthogonal functions (EOF) are adapted for the extension of time series. By a linear transformation of a set of runoff series from a region, a new set of series, EOFs, describing the runoff variations in time, and a set of weight coefficients describing the spatial variations, are created. Using this information it is possible to extend a runoff series within the region. A study has been carried out using Norwegian runoff series of daily values. The method was compared both to regression and to use of a conceptual model. In general the EOF method gives results as good as the traditional methods, and the amount of work involved is minimized while the objectivity is maximized.