Billings, S.Minty, B.Newsam, G.2006-06-192006-06-192003Geophysics, 2003; 68(4):1257-12660016-80331942-2156http://hdl.handle.net/2440/1322<jats:p>The first part of the paper presents a method for frequency domain deconvolution of airborne gamma‐ray surveys using a Wiener filter. A geometrical detector model is used to model gamma‐ray detection, with aircraft movement simply incorporated by a multiplicative term. The method requires estimation of the autocorrelation functions governing both signal and noise. The former is estimated through the radially averaged power spectrum of the survey data, whereas an error propagation analysis is used to estimate the latter, which is assumed white. Slight manual adjustments to the noise level are used to tune the reconstruction.</jats:p><jats:p>The technique is applied to a low‐altitude radiometric survey collected along closely spaced transects. Results are good for thorium, but are poor for both potassium and uranium. This can be attributed to the high noise levels in the potassium and uranium estimates, principally due to scattered gamma‐rays from high thorium concentration. Much better results are obtained when the method is applied to a survey with more typical radioelement concentrations. The reconstructions are improved significantly if an adaptive 2D Lee filter is applied prior to deconvolution.</jats:p><jats:p>The second part of the paper addresses how noise in the data and attenuation of signal due to the flying height limit the spatial resolution. The autocorrelation functions of signal and noise, along with the gamma‐ray model, can be used to determine how signal‐to‐noise ratio degrades with increasing height. The frequency where signal and noise are present in equal quantity can be used as an estimate of the spatial resolution. Predicted critical sampling rates range from 30 m at 20 m elevation to 60 m at 60 m elevation and 90 m at 120 m elevation.</jats:p>enJournal article002003201210.1190/1.15981180002208363000162-s2.0-014151446857789Newsam, G. [0000-0002-4492-3048]