Spatio-temporal covariance functions are main parts of geostatistical analysis of environmental data. Many classes of covariance functions along with a large number of parameters have been proposed to estimate correlation structure of the spatial-temporal data. This paper surveys using and comparing of two important heuristic algorithms Bees Algorithm (BA) and Genetic Algorithm (GA) for finding optimal values of spatio-temporal covariance parameters based on full likelihood, Composite and Weighted Composite likelihood function. Next, the accuracy and computational time of these methods are compared based on a simulation study. Finally, the combined method is used to explore the spatio-temporal correlation structure of the monthly level of groundwater data in Ilam province, Iran.
