 |
| Titel |
On the sensitivity of large scale sea-ice models to snow thermal conductivity |
| VerfasserIn |
O. Lecomte, T. Fichefet, M. Vancoppenolle, F. Massonnet |
| Konferenz |
EGU General Assembly 2012
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 14 (2012) |
| Datensatznummer |
250060888
|
|
|
|
|
|
| Zusammenfassung |
| In both hemispheres, the sea-ice snow cover is a key element in the local climate system and
particularly in the processes driving the sea-ice thickness evolution. Because of its high
reflectance and thermal insulating properties, the snow pack inhibits or delays the sea-ice
summer surface melt. In winter however, snow acts as a blanket that curtails the heat
loss from the sea ice to the atmosphere and therefore reduces the basal growth
rate.
Among the snow thermo-physical properties, snow thermal conductivity is known to be
one of the most important with regard to the sea-ice-related thermodynamical processes. In
the literature, both model and observational studies parameterize the snow thermal
conductivity as a function of density and several different relationships are used. For the
purpose of large scale modelling, one issue is then to have the snow density correctly
represented while, for computational cost reasons, a comprehensive snow scheme can
generally not be used in such models. Since it is known by observationalists that one of the
key atmospheric parameters that affect snow thermal conductivity and density is the wind
speed, one way to get around the problem is to try to have a realistic representation of the
snow density profiles on the sea-ice directly using observations or simple wind speed
depending parameterizations.
In this study, we analyze the importance of the snow density profile and thermal
conductivity in the thermodynamic-dynamic Louvain-la-Neuve Sea-Ice Model (LIM3),
which is part of the ocean modelling platform NEMO (Nucleus for European Modelling of
the Ocean, IPSL, Paris). In order to do this, a new snow thermodynamic scheme was
developed and implemented into LIM3. This scheme is multilayer with varying snow
thermo-physical properties. For memory and computational cost reasons, it includes only 3
layers but the vertical grid is refined in thermodynamic routines. Although snow
density is time- and space-dependent in the model, it is not a prognostic variable.
The shape of the density profile is either prescribed as a function of snow and ice
thicknesses and based on snow pit observations, or parameterized as a function of
seasonally averaged wind speeds. LIM3 exhibits a large sensitivity to the different
tested formulations in both the Arctic, due to considerable changes in the sea-ice
bottom growth and melting rates and therefore significant total sea-ice volume
changes between the various runs ; and the Antarctic, because of strong ice-ocean
feedbacks impacting on the sea-ice volume and extent of the whole Southern Ocean. |
|
|
|
|
|
|