Geoscience Reference
In-Depth Information
fill this gap (see, e.g., Schramm et al. 2007 ; Bense et al. 2009 ; Frampton et al. 2011 ),
thereby responding to the need for an increased realism of numerical permafrost models
highlighted by Riseborough et al. ( 2008 ) and Woo et al. ( 2008 ).
As the development of mesoscale permafrost-related hydrology models has only gained
momentum within the last years, it is not surprising that the situation is comparable or even
worse for climate models. Until recently, the representation of frozen ground physics, as
well as of above-mentioned characteristics (Sect. 3.1 ) like reduced permeability of frozen
soil and impeded infiltration of spring melt water, was, if at all, represented in a rather
simplified way. An intercomparison study of different land surface schemes especially with
respect to cold regions' climate and hydrology revealed large differences between the
models, even in case the implementation of frozen ground physics was constructed in a
similar way (Luo et al. 2003 ). Due to missing processes and related deficiencies of their
land surface schemes, climate models often show substantial biases in hydrological vari-
ables over high northern latitudes (Luo et al. 2003 ; Swenson et al. 2012 ). Therefore, large
efforts are ongoing to extend ESMs in this respect, in order to improve simulated soil
moisture profiles and associated ice contents, river discharge, surface and subsurface
runoff. The ESM improvement over permafrost areas is, e.g., one of the research objectives
of the European Union Project PAGE21 ( http://www.page21.org ).
Given the substantial range in the level of complexity and advancement of permafrost-
related processes implemented in the ESMs, the large variety of results from the CMIP5
models is not surprising (see Fig. 8 ; Koven et al. 2012 ). The most comprehensive ESM land
surface schemes include freezing and melting of soil water, the dependency of soil thermal
properties on water and ice content, multilayer snow schemes with snow on the top of the soil
instead of blending upper soil layers and snow and the representation of soil organic matter
(e.g., the Community Land Model (CLM) of Lawrence et al. 2011 ). In contrast, many models
incorporate only few of these processes. Soil hydrology is assessed using multilayer schemes
that compute vertical flows using Richards law (Richards 1931 ) or some of its derivations
(e.g., Oleson et al. 2004 ), thereby replacing more and more the formerly used bucket schemes.
The reduced permeability can thus be considered via coupling of soil thermodynamics and
hydrology, i.e. hydraulic properties are functions of liquid soil water content only, instead of
the total water content. This is refined in some models through the implementation of a
freezing point depression. Reduction in infiltration at the surface is partially assessed, ranging
from very simple approaches like total blocking soils in case of freezing to the consideration
of subgrid scale variability, based on power law relationships between infiltration and the
degree of soil freezing. Examples for global ESM land surface schemes that are in various
states of ongoing development are CLASS (Verseghy 1991 ), ORCHIDEE (Gouttevin et al.
2012 ), JSBACH (Ekici et al. 2013 ) and JULES (Best et al. 2011 ). The latter three also
participate in the Page21 ( http://www.page21.eu ) model improvement activities. One of the
planned Page21 improvements is the development of a global scheme for the formation and
drainage of thermokarst lakes that has not been implemented in any ESM up to now.
It is important to note that also for models that represent the same processes the results
may diverge markedly. This can be attributed to differences in parameterization schemes
and the choice of parameters, e.g., soil column depth and thickness of its layers, as well as
to the choice of input data, e.g., soil porosity or heat capacity of the soil's dry material. In
addition, initial conditions play a role due to the long spin-up times of model soils. These
may comprise several years for liquid and frozen soil water content, several years
(de Ridder 2008 ) up to two decades for soil temperature and several centuries to millennia
for soil carbon storages (Wutzler and Reichstein 2007 ; Hashimoto et al. 2011 ). In this
respect, Christensen ( 1999 ) pointed out the importance of an adequate initialization of soil
Search WWH ::




Custom Search