Simultaneous stochastic simulation of daily precipitation, temperature and solar radiation at multiple sites in complex terrain

Stochastic generation of daily precipitation, maximum temperature, minimum temperature and solar radiation, simultaneously at a collection of stations in a way that preserves realistic spatial correlations, is described and illustrated. The procedure is a generalization of the familiar Richardson `WGEN' approach in that the same basic model structure and local parameter sets are used. The precipitation model is extended to multiple sites by driving the local models with spatially correlated random numbers. The nonprecipitation variables are generated using a high-dimensional autoregression that is a direct extension of the conventional formulation. The necessary spatial correlations for both components are specified as functions of horizontal and vertical distances between stations so that the procedure can be applied to arbitrarily located station networks, for example at a set of regular gridpoints. Results are illustrated for a network of 62 stations in an area of varied terrain in the western United States.