Regional Gravity Field Modeling by Radially Optimized Point Masses

Case Studies with Synthetic Data

authored by: Miao Lin, Heiner Denker, Jürgen Müller
Abstract: A two-step point mass method with free depths is presented for regional gravity field modeling based on the remove-compute-restore technique. Three numerical test cases were studied using synthetic data with different noise levels. The point masses are searched one by one in the first step with a simultaneous determination of the depth and magnitude by the Quasi-Newton algorithm L-BFGS-B. In the second step, the magnitudes of all searched point masses are readjusted with known positions by solving a linear system in the least-squares sense. Tikhonov regularization with an identity regularization matrix is employed if ill-posedness exists. One empirical and two heuristic methods for choosing proper regularization parameters are compared. In addition, the solutions computed from standard and regularized least-squares collocation are presented as references.
Organisation(s): Institute of Geodesy
Type: Conference contribution
Pages: 233-239
No. of pages: 7
Publication date: 2016
Publication status: Published
Peer reviewed: Yes
ASJC Scopus subject areas: Computers in Earth Sciences, Geophysics
Electronic version(s): https://doi.org/10.1007/1345_2015_92 (Access: Closed)
: Details in the research portal "Research@Leibniz University"

BibTeX

@inproceedings{6de70eaff5ba48989c6ebab48a8a98a1,
title = "Regional Gravity Field Modeling by Radially Optimized Point Masses: Case Studies with Synthetic Data",
abstract = "A two-step point mass method with free depths is presented for regional gravity field modeling based on the remove-compute-restore technique. Three numerical test cases were studied using synthetic data with different noise levels. The point masses are searched one by one in the first step with a simultaneous determination of the depth and magnitude by the Quasi-Newton algorithm L-BFGS-B. In the second step, the magnitudes of all searched point masses are readjusted with known positions by solving a linear system in the least-squares sense. Tikhonov regularization with an identity regularization matrix is employed if ill-posedness exists. One empirical and two heuristic methods for choosing proper regularization parameters are compared. In addition, the solutions computed from standard and regularized least-squares collocation are presented as references.",
keywords = "Free depths, Least-squares collocation, Point mass method, Regional gravity field modeling, Tikhonov regularization",
author = "Miao Lin and Heiner Denker and J{\"u}rgen M{\"u}ller",
note = "Funding information: Three anonymous reviewers are acknowledged for their valuable comments which improved the original manuscript. The first author is financially supported by China Scholarship Council (CSC) for his PhD study in Germany.; 150th Anniversary with a Scientific Assembly, IAG 2013 ; Conference date: 02-09-2013 Through 06-09-2013",
year = "2016",
doi = "10.1007/1345_2015_92",
language = "English",
isbn = "9783319246031",
series = "International Association of Geodesy Symposia",
publisher = "Springer Verlag",
pages = "233--239",
editor = "Chris Rizos and Pascal Willis",
booktitle = "IAG 150 Years - Proceedings of the 2013 IAG Scientific Assembly",
address = "Germany",
}