Open AccessA stable downward continuation by using the ISVD method *
Author(s) -
Fedi M.,
Florio G.
Publication year - 2002
Publication title -
geophysical journal international
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.302
H-Index - 168
eISSN - 1365-246X
pISSN - 0956-540X
DOI - 10.1046/j.1365-246x.2002.01767.x
Subject(s) - continuation , computation , computer science , algorithm , noise (video) , field (mathematics) , sampling (signal processing) , exploit , taylor series , series (stratigraphy) , space (punctuation) , analytic continuation , domain (mathematical analysis) , stability (learning theory) , sequence (biology) , resolution (logic) , mathematics , geology , mathematical analysis , artificial intelligence , machine learning , image (mathematics) , pure mathematics , programming language , operating system , paleontology , computer security , filter (signal processing) , biology , computer vision , genetics
Summary Downward continuation of potential fields represents a very interesting way to enhance the information content of a gravity or magnetic map. In fact, apart from the increase of resolution, shared with many recent methods involving the use of directional derivatives, the downward continued data have the advantage of maintaining the physical dimensions of the original ones. This means that the interpretative tools that may be used are the same as for the untransformed data, but the obtainable models can fully exploit the benefits of the increased resolution of the field. However, because of its inherent instability, this method has progressively lost popularity. In this paper a stable downward continuation algorithm is presented. It is based on the computation of stable vertical derivatives obtained by the ISVD method and Taylor series expansion of the field. The algorithm uses both frequency and space domain transformations. Tests on synthetic examples proved its utility especially in cases where the signal is corrupted by noise, when the continuation level is greater than the data sampling step or when the needed continuation level is close to the source top level. Its application to real gravity data of Southern Italy shows that a stable downward continuation of a potential field to a close‐to‐source level gives very valuable information. In this case, while our algorithm can give meaningful results even continuing the field to a level close to sources, the use of a standard method dramatically decreased the signal‐to‐noise ratio necessitating low‐pass filtering of the transformed map.