 |
| Titel |
Sea level curves, geoid rate and uplift rate from composite rheology in glacial isostatic adjustment modeling |
| VerfasserIn |
W. Van der Wal, P. Wu, H. Wang, M. G. Sideris |
| Konferenz |
EGU General Assembly 2009
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 11 (2009) |
| Datensatznummer |
250024866
|
|
|
|
|
|
| Zusammenfassung |
| Laboratory experiments show that both diffusion creep and power-law creep can exist for
realistic mantle conditions. Therefore a composite rheology which includes both creep laws
might be a better approximation of the deformation process in the mantle. Here we study the
effect of such a rheology on glacial isostatic adjustment (GIA) observables. Composite
rheology has in the past been shown to provide a better fit to sea level data for a
wide range of parameters investigated with 2D finite element models. Here, we use
the Coupled Laplace Finite Element Method for an incompressible 3D spherical
self-gravitating Earth to study the effect of composite rheology on relative sea level (RSL)
curves, maximum present-day uplift rate and maximum present-day geoid rate with
the ICE-5G model. The long computation time of this model limits the number of
cases that can be investigated to a handful. The stress exponent is taken to be 3,
the pre-stress exponent (A) derived from a uni-axial stress experiment is varied
between 3.3 x 10-33/10-34/10-35/10-36 Pa-3s-1, and the Newtonian viscosity η is
varied between 1/3/9 x 1021 Pas. Because the ice models are developed under the
assumption of linear rheology any rheology with a non-linear component usually
under predicts the uplift rate and geoid rate. Therefore, to see if the ice model can
provide a better fit, we investigate simple modifications to the ICE-4G model, such as
i) scaling of the ice height by 1.5 and 2.0; and ii) delay in glaciation by 1 and 2
kyears.
The non-linear component in the composite rheology becomes important for high
effective stress, which is shown to occur at the edge of the ice sheet at the end of deglaciation.
As in previous studies, composite rheology is found to have a smaller misfit value with
observed RSL data than linear rheology. However, a purely non-linear rheology with
A = 3.3 x 10-35 Pa-3s-1 still has a slightly better fit than composite rheology,
although this is sensitive to outliers in the misfit computation and might also not hold
true when other values of the A and η are investigated. The best fitting composite
rheology has A = 3.3 x 10-35 Pa-3s-1 and η = 9 x 1021 Pas. For this model, transition
stress is 1 MPa, which is low enough so that most of the sea level curves follow
those of non-linear rheology. Exceptions are sites in the center of the ice sheet
(Laurentide) and the center and margin of the ice sheet (Fennoscandia), where there is a
significant contribution of linear rheology at the start of melting, and Antarctica,
where apparently stresses are not high enough for non-linear deformation to become
important. The maximum geoid rate of the best fitting composite rheology model is
1.0 mm/year (compared to 0.85 for the best fitting purely non-linear model and
the observed 1.4 mm/year) and the uplift rate is 7.5 mm/year (compared to 6.0
mm/year for the non-linear model and the observed 11 mm/year). Thus, the best
fitting composite rheology model increases the uplift and geoid rate compared to a
model with only power-law creep, at the expense of a small decrease in sea level
misfit.
The following conclusions are obtained from the simulations with a modified ICE-4G ice
history:
i) Increasing the ice thickness increases the present-day uplift rate, but only when the
non-linear deformation component in the model is small. The misfit value with sea level data
in Laurentide increases, and the fit with individual sea level curves becomes worse except for
two sites.
ii) A delay in glaciation increases the uplift rate for all values of A, and a 1 kyear delay
improves the misfit with Laurentide RSL sites, and even a 2 kyear delay improves sea
level fit for a number of stations. Therefore, delay in glaciation is a more promising
direction to adjust ice models for composite rheology than increasing the ice thickness. |
|
|
|
|
|
|