(r∂ zij’ ∂ w∂l īz,) 1. To separate a curve or a surface into its long wavelength (or low-frequency) parts, called the regional, and its short-wavelength (or high-frequency) parts, called the residual. Residualizing attempts to predict regional effects and find local anomalies by subtracting the regional effects. This separation is not unique. Residualizing methods (Figure R-10) include (a) graphical methods, in which a smooth regional is drawn on a profile or contours are smoothed and spaced more uniformly; (b) polynomial method, in which the regional is represented by a polynomial fit to the observed data; (c) spectral-domain filtering, in which certain wavenumbers are attenuated by filtering; (d) stripping method, in which the field of a model that represents certain parts of the geology is calculated and subtracted from the observed field; (e) upward continuation, which attenuates the effects of shallow sources. These methods can be thought of as 2D convolution operations (map convolution) and some of them produce halo effects about local anomalies. See Cowan and Cowan (1993).