On accurate geoid modeling: derivation of dirichlet problems that govern geoidal undulations and geoid modeling by means of the finite difference method and a hybrid method

The geoid is the reference surface used to measure heights (orthometric). These are used to study any mass variability in the Earth system. As the Earth is represented by an oblate spheroid (Ellipsoid), the geoid is determined by geoidal undulations (N) which are the separation between these surface...

Full description

Bibliographic Details
Main Author: Eduardo Del Rio
Format: Article
Language:Portuguese
Published: Universidade Federal do Paraná
Series:Boletim de Ciências Geodésicas
Subjects:
Online Access:http://www.scielo.br/scielo.php?script=sci_arttext&pid=S1982-21702014000200007&lng=en&tlng=en
Description
Summary:The geoid is the reference surface used to measure heights (orthometric). These are used to study any mass variability in the Earth system. As the Earth is represented by an oblate spheroid (Ellipsoid), the geoid is determined by geoidal undulations (N) which are the separation between these surfaces. N is determined from gravity data by Stokes's Integral. However, this approach takes a Spherical rather than an Ellipsoidal Earth. Here it is derived a Partial Differential Equation (PDE) that governs N over the Earth by means of a Dirichlet problem and show a method to solve it which precludes the need for a Spherical Earth. Moreover, Stokes's Integral solves a boundary value problem defined over the whole Earth. It was found that the Dirichlet problem derived here is defined only over the region where a geoid model is to be computed, which is advantageous for local geoid modeling. Moreover, the method eliminates several of the sources of uncertainty in Stokes's Integral. However, estimates indicate that the errors due to discretization are very large in this new method which calls for its modification. So, here it is also proposed an optimal combination of techniques by means of a Hybrid method and shown that it alleviates the uncertainty in Finite Difference Method. Moreover, a rigorous error analysis indicates that the Hybrid method proposed here may well outperform Stokes's Integral.
ISSN:1982-2170