Skip to end of metadata
Go to start of metadata

Surface Interpolators

Some methods are more sensitive to local, especially unique high amplitude, variation than others.With respect to the interpolation problem, it is useful to consider if local variation should be analyzed or not. If sampling is sparse, an over emphasis of local variation leads to the creation of pits and spikes at sample locations. Conversely, highly generalized approaches may not sufficiently accentuate localized low amplitude, but potentially useful, variation. Magnetometry is one case in point. Thermally produced magnetic anomalies of an archaeological sort are generally not very strong or high amplitude, yet there is variation. Inverse distance weighted can be a useful approach to representing this ind of variability. Yet, this same technique may not be optimal for other interpolation tasks like topography.

Interpolating Magnetometry Data

  1. Here are a couple of links that describe magnetometry at a very general level. Jstor is another place to look for resources related to magnetometry. Basically, magnetic field anomalies have both a positive and negative charge. These are like the poles of a magnet. When prospecting for magnetic anomalies, one is looking for places where positive and negative charged areas join. The location of the anomaly is where the two charged zones meet.
    1. Wikipedia has a decent page on Magnetic Surveys for archaeology
    2. The Boston University project at the historic Nathan and Polly Johnson House has a good description written by Holt.
  2. Import the block9iv_low.txt file and convert it to a shapefile.
  3. Make sure the Spatial Analyst extension is activated.
  4. Under Spatial Analyst, select Interpolate to Raster and observe the three options available.
    1. Inverse Distance Weighted
    2. Spline
    3. Kriging
  5. Read about each of these three interpolators available via the Spatial Analyst extension and make a prediction as to which is most likely to be suitable for representing Magnetometry data.
  6. Next try out each of the interpolators. In each case, be sure to set the Z field to READING_1. These are the readings of the magnetometer in units of nanoteslas.
  7. When comparing interpolators
    1. Use the name number of points in the search radius.
    2. Ensure that the Output Cell size is the same.
    3. The default display for the output raster is often categories. However, the variability of the data are probably better represented as a standard deviation stretch. Thus, when comparing be sure that the various grids produced are displayed with a standard deviation stretch and that the same palette is applied.
  • No labels