Surface nuclear magnetic resonance (NMR) is a geophysical technique that provides the ability to noninvasively image water content in the subsurface. To improve the ability of this method to produce images representative of the true subsurface structure, we require high spatial resolution. We derive a method to provide improved spatial resolution through the use of novel excitation strategies designed to enhance and exploit the information content within the quadrature component of the NMR signal. In a traditional surface NMR experiment, the frequency of the perturbing magnetic field () is chosen to equal the Larmor frequency of the hydrogen nuclei in the subsurface. In this case, it is assumed that the signal phase is determined entirely by the conductivity structure of the subsurface. Several studies have found that modeling the signal phase accurately and inverting a complex-valued NMR signal, can improve the spatial resolution of the surface NMR water content images. We propose alternative excitation schemes designed to generate a complex-valued signal, where the quadrature component can be controlled experimentally and was larger than that generated by the conductivity effects. This allowed a single excitation to provide two samplings of the subsurface properties, one stored in the real component and another in the quadrature component. To test if the alternative sampling strategies can provide improved spatial resolution in surface NMR, we evaluated a synthetic study contrasting the performance of three techniques. We contrasted two techniques designed to generate a complex-valued NMR signal during excitation, called off-resonance excitation and composite pulse excitation, to a traditional on-resonance excitation. We demonstrated that our proposed excitation schemes were able to better resolve boundaries between layers with contrasting properties, and we produced images with improved spatial resolution.