4.0 Potential field data

4.1 Loading observed potential field data

To load observed potential field data click on the relevant menu in the Menubar (e.g., Gravity) and select Load Gravity Anomaly. Next Navigate to the file you want to load and select it:

Gravity -> Load Gravity Anomaly

Provided the file is a simple space delimited ASCII text file and the X-coordinates match the model, the file will then be loaded into the model canvas

Note

These files should be ASCII text files with X values in the first column (in unit km) and the anomaly value (units: mGal, Eotvos, nT, m) in the second column. These files may have any suffix e.g., .txt or .xa (short for x-coordinate anomaly).

If you later wish to delete an observed data record from the modelling frame, navigate to, for example:

Gravity -> Gravity Data -> data name -> delete observed data

4.2 Calculating predicted potential field anomalies

The Menubar contains buttons G, V and M. These are switches for turning on/off the potential field calculations for the gravity, VGG and magnetic anomalies.

If the model becomes complex (i.e. contains many layers and nodes), turing off the calculation of the predicted anomalies can help speed up the GUI response time.

4.3 Potential field operations

4.3.1 Gaussian smoothing filter

A 1D Gaussian filter can be applied to any potential field data that has been loaded into a model by, for example, navigating to Gravity in the menubar and then selecting Gaussian Filter. This function uses the Scipy gaussian_filter1d operation, where sigma is the standard deviation of the Gaussian kernel.

Note

The Gaussian filter takes the potential field data as input and applies a mathematical operation called a Gaussian kernel to it. This kernel represents a bell-shaped curve, with the width of the curve determined by the user defined standard deviation parameter (sigma).

When the Gaussian filter is applied, it replaces each data point with a weighted average of itself and the neighboring points. The weights are determined by the values of the Gaussian kernel at those positions. This process effectively smooths out any sudden or sharp changes in the data, making it appear more gradual and continuous. This may be useful if data are particularly noisy or you are interested in modelling long wavelength features.

4.3.2 Horizontal derivative

The horizontal derivative of any potential field data that has been loaded into a model can be calculated by, for example, navigating to Gravity in the menubar and then selecting Take Horizontal Derivative. This uses a finite difference approach to estimate the horizontal gradient and reveals inflection points along the profile. These inflection points may delineate for example, layer edges or the location of faults. See, for example, Kearey et al., (2002) Pgs. 141-142 for more information.

Note

The horizontal derivative application requires an evenly spaced X coordinate increment, This is defined by the X increment parameter. If your data is not evenly spaced, it will first be interpolated evenly at the specified X increment.

The Filter Length parameter determines the length of the preconditioning median filter. By default, this value is set to 1, treating the input data in its raw state. Increasing this value applies smoothing to the data before calculating the gradient, which can be beneficial for noisy data. NB. This value must be an odd number.