Chapter 6

Quantitative Interpretation

Quantitative interpretation means quantification of qualitative models. “Qualitative”

is not “purely qualitative”. All models are approximate estimates. There is no clear
distinction between “qualitative” and “quantitative”. “Quantitative” is never “abso-
lutely accurate” or error-free, and model parameters must be complemented by their
error bounds. What is considered sufficiently accurate depends on the study aims,
data situation, numerical realisation etc.
The present chapter concentrates on how the forward problem is best solved by
calculating gravity effects δ g of given or assumed mass anomalies. The different
strategies for organizing the calculations are emphasized, based on qualitative mod-
els of Chap. 5. The ultimate aim of gravity interpretation requires detailed routines;
however, they are usually available or should be written for the given problems.
This chapter tries to show up the possibilities on the theoretical basis laid down in
Chap. 2. Trial and error, optimization and inversion build on this chapter.

6.1 Introduction: From Qualitative

to Quantitative Interpretation

Complex geological bodies must be constructed on the basis of simple body types.
Specific parametrizations offer themselves which are most appropriate to the de-
scription based on the coordinates of points and the associated densities: quadru-
plets {xk , yk , zk , ρk } or quintuplets {xk , yk , zk , ρk , tkl } where the index k signifies
a countable model parameter and l an instant in time. Any kind of coordinates are
in use. Frequently only two geometrical coordinates are specified, the third is as-
sumed infinite; such models are called “two-dimensional” or 2D (see Sect. 2.9.7).
Generally the basic model bodies are uniform in density.
Since the gravity effects of geological bodies are always small relative to standard
earth gravity g o in a study area, the associated vertical deflections are generally
negligible and in a fixed (x, y, z) coordinate system with z parallel to g o . The gravity
effect is thus always assumed to be the z component of the gravitational vector effect

W. Jacoby, P.L. Smilde, Gravity Interpretation, 233


c Springer-Verlag Berlin Heidelberg 2009
234 6 Quantitative Interpretation

although the disturbing vector will generally have also x, y components and thus
deflect g from g o and the fixed z direction.

6.1.1 Principal Considerations: Qualitative

and Quantitative Interpretation

In quantitative gravity interpretation essentially the locations, depths, shapes and

density contrasts of geological bodies are to be defined as “accurately” as possi-
ble. It may be desirable to first isolate those parts of the observed gravity anoma-
lies which are caused by the target bodies, and this is part of interpretation. It
requires independent additional a priori information, which, of course, may be
ambiguous in itself. The whole task, thus, will be to approach “the truth” by
combining all available information and adjust it mutually within the error lim-
its of each. The compromise between all pieces of conflicting evidence is largely
a matter of judgement and experience. The traditional manual approach by trial
and error is necessarily subjective and uncertain. Its quantification is Bayesian
inversion described in Chap. 7 which, thus, is the top aim of quantitative
interpretation.
Due to the ambiguity problem gravity can never define or prove a “true” model
but can definitively prove an assumed model wrong – in line with Karl Popper’s
view that science advances by falsification of preliminary hypotheses. One can
learn how wrong a model is. Gravity interpretation is never final: any new data
warrant better quantifying the models. Another valuable aspect of gravity interpre-
tation is suitability to aid interpolation between gaps in other types of information,
for example, between boreholes or seismic surveys. Standard methods as seismic
reflection may also leave “blind spots” where gravity modelling can be applied prof-
itably to fill such gaps. Thus, gravity is and remains a unique and economic tool for
exploration.
Non-ambiguous quantitative information, contained in gravity anomalies, is the
total amount of the anomalous mass and its horizontal centre of gravity (see
Sect. 2.7.6). The accuracy of the information depends on the accuracy of the
data and, to a large extent, on the definition of the zero level of an investi-
gated anomaly. In a somewhat similar fashion, maximum depths may also be es-
timated. Downward continuation of gravity anomalies through homogeneous layers
can also give some direct insights by computing the idealized mass anomaly
in the form of the equivalent stratum, but it is only a guide to quantitative
interpretation.
A model is as accurate as the given error bounds permit, i.e. errors from the data,
from neglected model aspects, from additional information. The best to be achieved
is good estimates of most likely model parameters and of their error bounds. They
are essentially of statistical nature, but that is too simple because of the ambiguity
problem, and the possibility to construct totally wrong models that fit within arbi-
trarily narrow error bounds.