MAXIMUM-LIKELIHOOD INFERENCE OF SPATIAL COVARIANCE PARAMETERS OF SOILPROPERTIES

The statistical description of the spatial variability of soil propert ies is a preliminary step to the application of other quantitative tec hniques, for example, optimal spatial interpolation by kriging. The sp atial correlation between the experimental data is characterized by th e covariance function that is usually inferred from the experimental m easurements. An appealing method for performing this inference is maxi mum likelihood estimation. This paper describes the results of the app lication of maximum likelihood estimation to the inference of the spat ial covariance of infiltration measurements for a small catchment, Bec ause there are a large number of experimental data, the approximate ma ximum likelihood method has been used to overcome the computational bu rden imposed by the calculation of the complete likelihood. As will be shown, this method, used in conjunction with classical graphical vari ogram analysis (method of moments), provides more objective inference of covariance parameters.