Draft : 31 August 941 An Application of Statistics toMeteorology : Estimation ofMotionDavid

1 ABSTRACT Concern is with moving meteorological phenomena. Some existing techniques for the estimation of motion parameters are reviewed. Fourier-based and generalized-additive-model-based analyses are then carried out for the global geopotential 500 millibar (mb) height eld during the period 1-6 January 1986. Early in his professional career Lucien Le Cam was a statistician at Elec-tricit e de France. His Fourth Berkeley Symposium paper, "A stochastic description of precipitation", (Le Cam 1961), describes a conceptual stochas-tic model for (perhaps moving) rainnelds developed at that time. In particular , Professor Le Cam was concerned with the development of river stream ow following rainfall. The model involved a smoothing transformation of a point process with direction and velocity of movement included. There have since been many references building on his work including: In this paper the focus is on velocity estimation of moving disturbances. To begin consider a number, N; of plane waves, g n (n x + n y ? v n t), moving across a surface. The model of interest is Y (x; y; t) = N X n=1 g n (n x + n y ? v n t) + noise (1) with (x; y) location, t time and the n?th wave having velocity v n and direction cosines (n ; n). The velocity is the parameter of particular interest in this work. The functions g n (:) may be known up to a nite dimensional parameter , eg. g(u) = cos(mu+) or simply may be assumed smooth. The