The present paper introduces the thin-plate spline as a potentially useful tool for spatial forecast verification. The approach can be used as both a neighborhood type (i.e., low-pass filter) method as well as a scale decomposition approach. In particular, a thin-plate spline can be represented in a symbolically equivalent manner as a wavelet decomposition. Briggs and Levine (1997) suggest the use of wavelets for forecast verification because of their ability to not only break a field down by physical scales, but also because the locations of the features at each scale are also preserved. For example, if one played a chord on a piano and applied Fourier analysis to decompose the chord, each individual note of the chord could be identified. However, if a second chord were played, again all of the notes can be identified, but no information is available as to which chords the notes belong. Wavelets, on the other hand, allow one to identify the individual notes, and to which chords they correspond (Ogden 1996; Meyer 1993).