<p>Fundamentally the amplitude, sign, and frequency of a magnetic signal are inherently linked to curvature. This thesis employs curvature analysis as a semi-automated tool for source signal extraction from a magnetic field surface represented by a grid. The first step is to compute the full, profile, and plan curvatures from the magnetic grid. These values are used in two approaches to curvature analysis: statistical and lineament. The descriptive statistics mean, standard deviation, kurtosis, and skew are computed for quantitative analysis. Mean is used in conjunction with kurtosis and skew to assess frequency content of the signal, magnetization and source dip. Standard deviation characterizes low, moderate, and extreme curvatures. A rapid technique to statistical analysis is applied using a graphical approach with histograms and scatterplots. Histograms display frequency distribution and scatterplots display the relationship between different curvatures. Curvature in the maximum dip direction is used to systematically identify surficial lineaments characterized as continuous troughs or ridges. These lineaments may represent geological sources or remanent acquisition artefacts. Lineaments representing faults and dykes are used in conjunction with <em>a priori</em> knowledge to determine mineralization vectors since many ore deposits are structurally controlled. Quality control of the aeromagnetic grid levelling application may be assessed using spatial correlation of flight lines and magnetic lineaments. In this work curvature analysis is applied to simple synthetic models and two Canadian aeromagnetic data sets. Curvature analysis was applied to magnetic data from the Wopmay Orogen to identify bedrock contacts, fault configurations, and dyke swarms. The data was also used to show lineaments displayed as rose diagrams may be used as an alternative to standard Fourier power spectrums for assessment of levelling. Magnetic survey data from Southern Ontario was used to show a statistical approach to identify regional dip, dominant magnetization, and interference in anomalies.</p> / Doctor of Philosophy (PhD)