Time series analysis in acid rain modeling: Evaluation of filling missing values by linear interpolation

Abstract This paper describes the results of a study to evaluate how well the linear interpolation of the two nearest bracketing values (LIBV) scheme for filling missing values might be expected to perform. Computer simulation techniques were used to generate simulated time series which were probabilistically equivalent to the types that are likely to be found in practice. Actual data on the frequency and duration of missing value episodes were analyzed to determine the probability models which described the behavior of these missing value episodes. The models were then used to inject missing value episodes into the simulated time series. The LIBV scheme was then used to fill these missing values. The values of the parameters of the time series model were then estimated. The estimated parameter values were then compared to the true values of the parameters which were used to generate the lime series. The results of these comparisons suggest that the LIBV scheme could yield good results if only one dominant seasonal component is present in the data and that it should be used with caution when more than one dominant seasonal component is likely to be present.