Leveling Corrections For Ultra-Low Altitude Magnetic Surveys

Abstract

Six million hectares at about 1500 sites in the U.S. are contaminated with unexploded ordnance (UXO) with the cost of cleanup estimated at $500 billion. The development and implementation of ultra-low (<10 m) altitude airborne methods to detect UXO has the potential to cut this cost by half. Unfortunately, magnetic data acquired at these low altitudes are severely contaminated with flight-height related noise which arises because of the strong dependence of the anomaly strength, T, on the detection distance, r, Toc1 / r^3. Survey altitudes commonly fluctuate by more than 25% even in ideal conditions. Data acquired above the mean flight height are attenuated with respect to the data acquired at the mean height, while data acquired below the mean height are magnified. This causes large corrugations (easily 100% of the background magnitude) in the data, inducing error in the calculation of target depth and position, and creation of false positives. The noise may be removed by migration of the survey measurements to a common altitude using upward and downward continuation. Three published algorithms (Taylor-series, Chessboard, and equivalent source) were tested on four synthetic data sets and on data acquired at a UXO test site on the Badlands Bombing Range in South Dakota. Corrections were applied to both the profile and the gridded data. The results indicate that the grid-based equivalent source method, combined with a preferential low-pass filter to stabilize downward continuation, is the most accurate and versatile of the methods. Application to field data showed an eightfold increase in the signal-to-noise ratio, which was characteristic of the synthetic data as well. These results are promising for the future of ultra-low magnetic surveys to detect UXO.