Optimal Spatial Weighting Methods for Estimation of Missing Rain Gage Records

Deterministic and stochastic weighting methods are the most frequently used methods for estimating missing rainfall values at a gage based on values recorded at all other available recording gages. Distance-based weighting methods suffer from one major conceptual limitation based on the fact that Euclidian distance is not always a definitive measure of the correlation among spatial point measurements. Several spatial weighting methods are proposed, developed and investigated for estimation of missing precipitation records. These methods use mathematical programming formulations and evolutionary algorithms. Historical daily precipitation data obtained from 15 rain gauging stations from a temperate and semi-arid climatic region are used to test and derive conclusions about the efficacy these of methods. Results suggest that the weights derived based on mathematical programming formulations and surrogate parameters for correlations are superior to those obtained from scalar distance-based weights used in spatial interpolation methods for estimation of missing rainfall data at points of interest.