 |
| Titel |
Clues of the deformation source at Campi Flegrei (Italy) from the use of the Akaike Information Criterion |
| VerfasserIn |
Antonella Amoruso, Luca Crescentini |
| Konferenz |
EGU General Assembly 2010
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 12 (2010) |
| Datensatznummer |
250040578
|
|
|
|
|
|
| Zusammenfassung |
| Volcanic risk in the explosive Campi Flegrei (CF) caldera is extremely high, because
of its location in a densely populated area about 15 km west of Naples inside the
Campanian Plain. The caldera has been generally subsiding (at about 1.5 cm/yr)
from 1538 (last eruption) till 1969. A substantial ground uplift, more than 1.5 m,
occurred in the period 1969-1972 and, after a small subsidence of about 30 cm
after 1972, a very large uplift occurred in the period 1982-1984 (about 1.8 m),
with subsequent partial recovery (about 60 cm in 2 yr). Superposed on the still
continuing subsidence are some short uplift phases (mini-uplifts with a few cm
amplitude); ground level still remains about 2.5 m above pre-1970 levels at the town of
Pozzuoli.
The interpretation of deformation data (levelings, EDM, GPS, SAR) is the subject of
long-lasting work and several papers. A number of different source models have been tested
and/or proposed, including points of isotropic expansion (Mogi sources), rectangular tensile
sources, pressurized cracks, expanding prolate spheroids, small triaxial ellipsoids, small
sources with a generic moment tensor. Also the number of sources (one, two, or even more) is
debated.
Up to now, there is no systematic comparison among the capability of the different
sources to account for deformation data. In case two or more sources are compared, the
model giving the lowest normalized chi square is usually preferred. But the minimum
normalized chi square is not a good criterion of choice. If nested models (the more
complicated model includes the simpler one as a particular case) have to be compared, the
F-test approach could be used. Unfortunately, the F-test is unreliable if residuals depart from
the normal distribution even slightly, and is not valid if the models are not nested. The Akaike
Information Criterion (AIC) is a measure of the goodness of fit of an estimated
statistical model. It is based on information theory and does not use hypothesis
testing, so there is no conclusion about statistical significance and rejection of a
model, but does not require the models to be nested, and thus all models can be
compared.
Here we show the results of a systematic comparison among different models of the
source(s) of the Campi Flegrei deformation, with particular emphasis on the 1982-1984
unrest, using AIC. |
|
|
|
|
|
|