Estimation of the accuracy and precision associated with a gravity survey is not an easy task. We must estimate not only the accuracy and precision associated with each observation of the Earth's gravity field, but also, because most modern surveys are acquired from moving platforms (or individual land stations are leveled as interconnected networks), we must estimate the spatial resolution of the survey. In one of the first attempts to understand this problem, Alan Herring of EDCON prepared the following graph, summarizing the estimated range of accuracies and spatial resolution of gravity acquisition systems. Individual estimates taken from specific surveys agree fairly well with this envelope, but vary depending on actual survey conditions. This graph plots survey accuracy as a function of shortest observable wavelength for various types of gravity acquisition systems.
Perhaps the best way to estimate survey accuracy and spatial resolution is to record “repeat lines,” that is, to make multiple observations of the gravity field over the same traverse. Fourier analysis of the repeat lines provides a good estimate of the spatial reproducibility of the survey and allows characterization of the noise statistics.
More commonly, gravity surveys are acquired as a series of intersecting profiles. Harmonization of the line intersections required to “level” the survey and the resultant “mis-tie” histograms often are quoted as indicative of the resolution of the survey, but they provide little information about the spatial resolution