A new technique for imputation of multivariate time series: application to an hourly wind dataset CARLOS LÓPEZ† and ELÍAS KAPLAN‡ Centro de Cálculo, Faculty of Engineering (11) CC 30, Montevideo, Uruguay †internet: carlos@fing.edu.uy ‡internet: elias@fing.edu.uy Abstract:The techniques employed in the treatment of an hourly surface wind database during the development and calibration phases of an objective wind field interpolator model are presented. The model itself has been applied to estimate the regional wind energy resource creating a layer in a GIS environment. The outlier detection phase is presented in a companion paper, and here the different techniques applied in order to imputate the missing values are described. The comparative results obtained with an hourly dataset of 15 years long are also presented. Two different problems have been simulated numerically: systematic missing values (i.e. at fixed hours) and non systematic ones. Five different criteria were applied: imputation with the historical mean value; linear time interpolation within single station records; optimum interpolation (kriging) and the two newly developed Penalty Of the Principal Scores and linear Time Interpolation of the Principal Scores which considers all station records in a multivariate fashion; they prove to be the most accurate for this particular wind dataset. There is also some evidence of oversampling in time. 1. Introduction 1.1 Presentation of the problem Since 1988 the Team was involved in evaluating the National Wind Energy Resource. Although it was not an explicit objective of the project, it was necessary to complete the time series to the longest available period. With these goals some algorithms have been implemented for imputation of missing values in the series. Some test have been carried out that confirm the kindness of their performance in controlled cases. Three different methods have been tested: a) Time interpolation of the principal scores (TIPS) (includes standard time series interpolation as special case) b) Penalty of the principal scores (POPS) and c) Optimum interpolation (Gandin, 1965). The first two were developed in López et al. (1994b). The third is a standard interpolation procedure in the inicialization of mathematical models in meteorology (Johnson, 1982) that allows to find estimates not only in the stations