FM256 Data Quality
Processing and Analysis with Geoplot
To realise the full potential of magnetometer data, sophisticated data processing and analysis is essential. Geoplot software provides this support in the FM256 system. It provides simple download and dual gradiometer data merge facilities, combined with excellent graphics, editing, processing, analysis, interpretation, presentation, import and export facilities.
Often only minimal processing will be required before analysis and presentation. However, in common with all magnetometers, there will be noise present in the data which reflects a combination of soil noise, instrument noise, and field (operator) noise. This can limit feature detection when operating near measurement limits. Geoplot's comprehensive range of functions provides the means to reduce significantly these noise components, allowing very weak features to be detected. Data analysis can further extend the confidence with which features are identified.
Geoplot's toolkit of specialised processing functions include the following : Absolute, Add, Clip, Compress, Cut and Combine, Deslope, Despike, Destagger, Edge Match, High Pass, Low Pass and Median Filters, Interpolate, Multiply, Periodic Filter, Power, Randomise, Search and Replace, Spectrum, Standard Deviation / Variance Map, Zero Mean Grid, Zero Mean Traverse. Correct processing of the data shown on the right using Geoplot results in an overall noise level of 0.1nT - this processing was achieved even in the presence of very strong and extensive ferrous responses. In another example, a 50cm thick Pueblo midden was detected (and confirmed by excavation) 1.9m below sand and clay layers and registered as a 0.05nT anomaly after appropriate processing.
Geoplot provides a number of analytical tools including statistics, spectrum, standard deviation maps. Particularly useful for gradiometer data is statistical detection, illustrated by the data set below. A collection of weak hearths are visible in the left hand plot. A combination of high sampling density and appropriate data processing results in a background noise level of 0.13nT. This includes soil, instrument and field method noise. Setting a detection threshold of 2.5 standard deviations (high statistical confidence), reveals a large number of hearths (right plot) subsequently confirmed by excavation.