MGSTOOLS WORKSHOP July 9-10, 2005 MGSTOOLS Marine Gradiometer Software Workshop Dr. Richard Hansen...

MGSTOOLS WORKSHOPJuly 9-10, 2005

MGSTOOLSMarine Gradiometer Software

Workshop

Dr. Richard Hansen

PRJ, Inc.

Presenter

MGSTOOLS WORKSHOPJuly 9-10, 2005

MGSTOOLS

• A tool for enhancing ability to discriminate between external signals and geology

• For use with longitudinal magnetic gradiometer systems

• Includes computation of power spectrum, sensor lag and ship’s heading effects

MGSTOOLS WORKSHOPJuly 9-10, 2005

MGSTOOLS Features (con’t)

• Provides ability to reconstruct total field data minimizing temporal variations and ship noise

• Provides tool for generation of synthetic data sets for testing parameters

• In conjunction with MagMap2000, provides graphical user interface for plotting raw and reconstucted data

MGSTOOLS WORKSHOPJuly 9-10, 2005

Geological Field Example

MGSTOOLS WORKSHOPJuly 9-10, 2005

External Field Example• Solar wind, field line compression

MGSTOOLS WORKSHOPJuly 9-10, 2005

External Field Example• Long term diurnal

MGSTOOLS WORKSHOPJuly 9-10, 2005

External Field Example

• Micropulsations

MGSTOOLS WORKSHOPJuly 9-10, 2005

External Field Example

• Magnetic Storm

MGSTOOLS WORKSHOPJuly 9-10, 2005

Superposition of geology and external fields

• Components cannot be distinguished with only one sensor

• On land, stationary base station is used (no geology)

MGSTOOLS WORKSHOPJuly 9-10, 2005

Gradiometer Configuration

• Physical Arrangement

MGSTOOLS WORKSHOPJuly 9-10, 2005

Gradiometer Configuration

• Show records at equal times, with geological source

MGSTOOLS WORKSHOPJuly 9-10, 2005

Gradiometer Configuration

• Because sensors are in different positions, geology is shifted

MGSTOOLS WORKSHOPJuly 9-10, 2005

Gradiometer Configuration

• External field anomaly has no appreciable shift

MGSTOOLS WORKSHOPJuly 9-10, 2005

Gradiometer Configuration

• Difference: external field effects are gone, but geology is a directional gradient

MGSTOOLS WORKSHOPJuly 9-10, 2005

Gradiometer Configuration

• Need to reconstruct geology field from difference

MGSTOOLS WORKSHOPJuly 9-10, 2005

Reconstruction - General

• Look again at combination of geology and external field: equal spatial position

• For short-wavelength external field, shift is clear

• For long-wavelength external field, shift cannot be seen

• Reconstruction cannot distinguish at long wavelengths (compared to sensor separation)

MGSTOOLS WORKSHOPJuly 9-10, 2005

Reconstruction - General

• Look again at combination of geology and external field: equal spatial position

MGSTOOLS WORKSHOPJuly 9-10, 2005

Reconstruction - General

• For short-wavelength external field, shift is clear

MGSTOOLS WORKSHOPJuly 9-10, 2005

Reconstruction - General

• For long-wavelength external field, shift cannot be seen

MGSTOOLS WORKSHOPJuly 9-10, 2005

Reconstruction - General

• Reconstruction cannot distinguish at long wavelengths (compared to sensor separation)

MGSTOOLS WORKSHOPJuly 9-10, 2005

Naïve Reconstruction – why it fails

• Trapezoid rule – show transfer function

• Short-wavelength errors (sensor noise, wave noise, …) unimportant because operator is small at short wavelengths.

• Long-wavelength errors are amplified: subheadings follow?

• Heading error (how do we show this? – We could show striping or swing curves)

• Leeway

• Vertical Offset

MGSTOOLS WORKSHOPJuly 9-10, 2005

Naïve Reconstruction

• Trapezoid rule transfer function

MGSTOOLS WORKSHOPJuly 9-10, 2005

Naïve Reconstruction

• Example Short-wavelength errors

MGSTOOLS WORKSHOPJuly 9-10, 2005

Naïve Reconstruction

• Example Long-wavelength errors

MGSTOOLS WORKSHOPJuly 9-10, 2005

Naïve Reconstruction

• Example Heading errors

MGSTOOLS WORKSHOPJuly 9-10, 2005

Naïve Reconstruction

• Example Leeway

MGSTOOLS WORKSHOPJuly 9-10, 2005

Naïve Reconstruction

• Example Vertical Offset

MGSTOOLS WORKSHOPJuly 9-10, 2005

MGS Tools Algorithmgeneral strategy

• Roll off transfer function at long wavelengths

• Minimizes effects of long-wavelength errors

• Lose ability to discriminate external field at long wavelengths (e.g., diurnal)

• Need to adopt a rational strategy for designing roll-off

• Fit spectrum using a set of exponentials

• Identify components dominated by geology, external field, short-wavelength noise

• Use this to design operator

MGSTOOLS WORKSHOPJuly 9-10, 2005

MGS Tools Algorithm

• Roll off transfer function at long wavelengths

MGSTOOLS WORKSHOPJuly 9-10, 2005

MGS Tools Algorithm

• Minimizes effects of long-wavelength errors

MGSTOOLS WORKSHOPJuly 9-10, 2005

MGS Tools Algorithm

• Lose ability to discriminate external field at long wavelengths (e.g., diurnal)

MGSTOOLS WORKSHOPJuly 9-10, 2005

MGS Tools Algorithm

• Need to adopt a rational strategy for designing roll-off

MGSTOOLS WORKSHOPJuly 9-10, 2005

MGS Tools Algorithm

• Fit spectrum using a set of exponentials

MGSTOOLS WORKSHOPJuly 9-10, 2005

MGS Tools Algorithm

• Identify components dominated by geology, external field, short-wavelength noise

MGSTOOLS WORKSHOPJuly 9-10, 2005

MGS Tools Algorithm

• Use data to design operator