 |
| Titel |
Putting global warming to use in accessing the 3D viscosity structure of the uppermost mantle beneath Iceland |
| VerfasserIn |
Peter Schmidt, Björn Lund, Thóra Árnadóttir, Harro Schmeling |
| Konferenz |
EGU General Assembly 2011
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 13 (2011) |
| Datensatznummer |
250053804
|
|
|
|
|
|
| Zusammenfassung |
| Located directly south of the arctic circle, approximately 11 % of Iceland is covered by
glaciers. However, a general regression trend of the Icelandic glaciers commenced around
1890. Over the period 1890-2004, the largest glacier in Iceland, Vatnajökull, is estimated to
have lost an ice-volume of 435 km3, causing the land to uplift due to Glacial Isostatic
Adjustment (GIA). At present, vertical velocities up to 20 mm/yr have been measured close
to the ice-edge.
Iceland is located on top of a mantle plume and cut through by the mid-Atlantic spreading
ridge. Due to the presence of the plume, the crust is found to be anomalously thick and
significant viscosity variations can be expected in the upper-most part of the underlying
mantle. Using 3D visco-elastic Finite Element models and including all major Icelandic
glaciers, we present the first GIA-study of Iceland investigating the 3D viscosity structure
in the mantle. In addition, we study the influence on the model results by lateral
variations of the elastic thickness of the lithosphere, Te, as well as the presence of the
spreading ridge. The models are evaluated against vertical velocities estimated from
observations of a nation-wide GPS network in 1993 and 2004 (ISNET), as well as
data from continuous GPS stations. Sensitivity tests indicate that the GPS data are
potentially capable of detecting viscosity contrasts at depths down to 250-500 km, for an
upper most mantle viscosity in the range of 1018 - 1019 Pa s as found in previous
studies.
Starting with simple models with an elastic layer underlain by a visco-elastic
half space, we find a best fit model with an Te of 35 km and a half-space viscosity
of 1019 Pa s. Although models with Te in the range 10-60 km combined with a
half-space viscosity in the range 6-15 x 1018 Pa s, detoriates only slightly in the fit to
the GPS data. We observe that the viscosity of the half-space is better constrained
than Te, in accordance with previous studies. We also note a slight preferred trend
of lower viscosity combined with higher Te. Incorporating lateral variations of
Te in the models, or a weak zone representing the spreading ridge, we observe a
significant effect on the vertical velocities predicted by the models. However, the
difference between the vertical uplift rates predicted from models with uniform Te and
models with laterally varying Te is most pronounced in regions below or close
to the glaciers, where we have few data. Therefore the fit to the dataset is hardly
affected.
Proceeding to simple plume models (vertical cylindrical conduit with a plate-like head),
we cannot find a model that fits the data well for a uniform plume viscosity, unless the plume
head extends to great depth and the viscosity contrast between the plume and the mantle is
small. For example, using a mantle viscosity 5 times higher than in the plume, the plume head
needs to extend to depths greater than 200 km. This can however be resolved if we allow for a
higher viscosity in the plume head, such that the viscosity in the plume head is higher than in
both the mantle and the plume conduit. The best fit model has an elastic thickness
of 30 km, a depth to the base of the plume head of 120 km and a conduit radius
of 100 km. The viscosity of the plume head is 2 x 1019 Pa s, of the conduit 3 x
1018 Pa s, and in the mantle 1019 Pa s. We note that this model can fit the data
almost as well as the best fit uniform mantle model. We systematically test a range of
models by varying one of the six model parameters: the viscosity in the head, the
viscosity in the conduit, the viscosity in the mantle, Te, depth to base of head, and
conduit radius. The predicted vertical velocities appear most sensitive to the viscosity
in the plume head and mantle, and least sensitive to Te and the viscosity conduit
radius.
Finally we present results from models where the viscosity structure is derived from
dynamic plume modeling. However, dynamic plume models generally demand higher
viscosities than those inferred from GIA studies of Iceland, which requires us to rescale the
former. We discuss the implications of this rescaling and show how the resulting GIA models
fit the GPS data. |
|
|
|
|
|
|