This study is aimed to estimate missing rainfall data by dividing the analysis into three different percentages namely 5%, 10% and 20% in order to represent various cases of missing data. In practice, spatial interpolation methods are chosen at the first place to estimate missing data. These methods include normal ratio (NR), arithmetic average (AA), coefficient of correlation (CC) and inverse distance (ID) weighting methods. The methods consider the distance between the target and the neighbouring stations as well as the correlations between them. Alternative method for solving missing data is an imputation method. Imputation is a process of replacing missing data with substituted values. A once-common method of imputation is single-imputation method, which allows parameter estimation. However, the single imputation method ignored the estimation of variability which leads to the underestimation of standard errors and confidence intervals. To overcome underestimation problem, multiple imputations method is used, where each missing value is estimated with a distribution of imputations that reflect the uncertainty about the missing data. In this study, comparison of spatial interpolation methods and multiple imputations method are presented to estimate missing rainfall data. The performance of the estimation methods used are assessed using the similarity index (S-index), mean absolute error (MAE) and coefficient of correlation (R).

In meteorological and hydrological researches, missing rainfall data has always been one of the most challenging problems which need to be faced by the researchers. The problems of missing rainfall data are due to the wrong technique used when measuring the rainfall, relocation of the rain station and malfunctioned of instrument. Finding the suitable method to solve missing data problem is very critical before going to the next level of data analysis. Most researchers used the spatial interpolation method to estimate the missing rainfall data at a particular target station which is based on the available rainfall data at their neighboring stations. The spatial interpolation method is one of the traditional weighting factors which also consider the correlation between the stations. This study uses the modified of spatial interpolation weighting methods to estimate the missing rainfall data in Pahang and only assume that the particular target station has the missing value. A new modified method of normal ratio and inverse distance weighting with correlation is proposed by abbreviated by NRIDC. The performance of the modified spatial interpolation weighting methods used are assessed using the similarity index (S-index), mean absolute error (MAE) and coefficient of correlation (R) for different percentage of missing values (5%-30%).

No abstract

No abstract

No abstract