Infilling Missing Flow Data 'Source Title' Modeling Hydrologic Processes, Proceedings Of 3rd International Hydrology Symposium
Bibliographic details
Summary
Infilling of missing sections of data is often necessary prior to the practical use of hydrological time series. Three methods of infilling are examined. The weighted average model, in which the parameters are inversely related to the distances between the stations with complete and incomplete records, provides a first approximation. The bias in the mean and variance is corrected in the second method, called the modified weighted average model. finally, in order to maintain the serial correlations in the infilled data and cross correlations with known data series, a multivariate model is proposed. A case study is presented in which the different methods are applied to a set of five concurrent 47 year monthly flow series from England and Wales. Sections from two of these series are hidden and after infilling the missing values comparison is made between the synthetic and historical data. The multivariate model is shown to provide the best results from hydrological and statistical considerations.