Spatial–Temporal Analysis of Temperature Using Smoothing Spline ANOVA

A new method, smoothing spline ANOVA, for combining station records of surface air temperature to get the estimates of regional averages as well as gridpoint values is proposed. This method is closely related to the optimal interpolation (also optimal averaging) method. It may be viewed as a generalization of these methods from spatial interpolation methods to a method interpolating in both spatial and temporal directions. The connection of this method to the commonly used anomaly approach is discussed in the context of correcting biases resulting from incomplete sampling. A main strength of this new method is its ability to borrow information across both space and time just like optimal interpolation does across space. This increases not only the accuracy of estimates but also the ability to correct various biases resulting from incomplete sampling. Some of these biases are ignored by the anomaly approach.