Optimal Robust Methods for Prediction and Design in Spatial Studies Running title: Spatial Prediction and Design

We develop and test robust methods for design construction, for estimation and for prediction in spatial studies. The designs are robust against misspecified variance/covariance structures, and against misspecified regression responses. Robustness against contaminated error distributions is provided by the use of generalized M-estimators in the estimation and prediction procedures. The loss function is based on the mean squared error of the predicted values. This is maximized, analytically, over the various neighbourhoods quantifying the departures from the fitted model. This maximum is then minimized numerically by simulated annealing, or sequentially in order to obtain the optimal designs.