In this paper we consider numerical algorithms for stochastic simulation of time series and space fields, which take into account the specific character of real hydrometeorological processes and fields.We propose and investigate a special algorithm for simulating inhomogeneous threedimensional fields of a discrete argument in which the field correlation structure is height-dependent. We consider two special stochastic models of binary stationary time sequences related to precipitation series. We give model estimates for the distributions of arid and rainy periods, which are needed to investigate extreme precipitation regimes.