Modelling Antarctic Climates and Ice Sheets

Aspects of global Middle Miocene to Pliocene climates and environments have been examined in palaeoclimate modelling studies. For example, the impact that oceanic gateways such as the Panama Gateway (Klocker et al., 2005), the Indonesian Throughflow (Cane and Molnar, 2001) and Tethyan circumglobal passage (Hotinski and Toggweiler, 2003) had on global oceanic circulation patterns and heat/moisture transports has received considerable attention. The role of Tibetan uplift, as well as other regional orographic and palaeogeographic variations, in changing atmospheric circulation patterns and the development of monsoon systems has also been a key area of research (Prell and Kutzbach, 1992; Kutzbach et al., 1993; Fluteau et al., 1999; Bice et al., 2000; Zhisheng et al., 2001; Liu et al., 2003; Kutzbach and Behling, 2004; Ramstein et al., 2005). Reconstructing and predicting the climatic implications of altered distributions of vegetation types (Dutton and Barron, 1996, 1997; Cosgrove et al., 2002; Haywood et al., 2002a; Haywood and Valdes, 2006), sea surface temperature gradients (SST) and latitudinal temperature gradients/heat transports have been studied in detail (Chandler et al., 1994; Sloan et al., 1996; Rind, 1998; Haywood et al., 2000; Barreiro et al., 2006). However, very few palaeoclimate modelling studies have specifically addressed the question of how climates and environments changed on Antarctica during this interval. The majority of global palaeoclimate modelling experiments have used specified Greenland and Antarctic ice volumes based solely on estimates of global sea level. Therefore, modelling studies conducted thus far could be considered as representing a suite of sensitivity experiments that have tested the response of climate models to different initial conditions rather than accurate climate ''retro-dictions''.

One example of a palaeoclimate modelling study that did focus on Antarctica is Haywood et al. (2002b). In this study, they attempted to assess the sensitivity of a climate model to different Antarctic Ice Sheet configurations and to address, indirectly at least, the controversy existing over the behaviour of the EAIS during the Pliocene epoch and whether or not it was possible for Nothofagus remains founds in Sirius Group deposits in the Transantarctic Mountains to be Pliocene in age. Three sensitivity experiments were performed for the Pliocene using the UK Meteorological Office's General Circulation Model (GCM). Each experiment used a different configuration of Antarctic ice conforming to a range of sea-level estimates supported by geological evidence (specifically 35, 25 and 12-15 m higher than modern sea level). Climate outputs from these experiments were used to drive a biome model capable of predicting the equilibrium vegetation state for a given climate forcing. For the experiments using an Antarctic Ice Sheet configuration equivalent to a +12 to 15 and +25 m sea level, the biome model predicted the occurrence of an exclusively tundra type of vegetation on deglaciated regions of Antarctica (Fig. 10.7). In the +35 m experiment, climate ameliorated sufficiently to allow the model to predict deciduous taiga montane forests in areas of Antarctica where fossil Nothofagus material has been recovered from Sirius Group sediments (Fig. 10.7). This work suggested that it is feasible for Nothofagus to have existed in Antarctica during the Pliocene with ice-sheet extents calibrated to global sea-level records.

This research had certain limitations relating to the experimental design. The configuration of Antarctic ice was specified within the GCM and the resulting climatologies were only used to drive an offline biome model. A more sophisticated approach would use a coupled climate-ice and vegetation model capable of predicting, rather than being prescribed with, the equilibrium condition for Antarctic ice cover and vegetation with climate. As a step towards this, Francis et al. (2007) dynamically coupled the Top-down Representation of Interactive Flora and Foliage Including Dynamics (TRIFFID) dynamic global vegetation model to the HadAM3 GCM set up for the mid-Pliocene. The results supported the earlier study by Haywood et al. (2002b) since the TRIFFID model predicted tundra vegetation on deglaciated regions of Antarctica.

A number of studies have utilized ice-sheet models to predict the response of the EAIS to increasing surface temperatures the results from which, in a few cases, have been related to the debate over how dynamic or stable the Antarctic Ice Sheet was likely to have been during the Middle Miocene to Pliocene (e.g. Barker et al., 1999b). Huybrechts (1993) used a three-dimensional ice-sheet model to determine ice-sheet geometries under various kinds of climatic conditions. A surface temperature warming of between 17 and 20°C was needed to generate the ice-free corridor over the Pensacola and Wilkes subglacial basins hypothesized by Harwood (1983) and Harwood and Webb (1986). For temperature rises of less than 5°C, the model predicted that the EAIS increased in size due to an increase in snowfall. Even with the most favourable model set up for ice loss, a surface temperature rise of 15°C was still required to make the Pensacola and Wilkes subglacial basins ice-free which Huybrechts (1993) concluded was unlikely for the interval in question.

However, there are a number of limitations inherent in such an approach. First, by assessing the response of the modern EAIS to a certain climatic forcing, an assumption is made that the EAIS during the Middle Miocene and Pliocene was the same as it is today. Other important differences may

Pliocene control experiment

Pliocene control experiment

Pliocene dyanamic experiment

Pliocene dyanamic experiment

Biome key:


Cushion forb lichen moss tundra

Prostrate shrub tundra

Dwarf shrub tundra

Shrub tundra

Steppe tundra


Deciduous taiga/montane forest Evegreen taiga/montane forest Open conifer woodland Temperate sclerophyll woodland Temperate xerophytic shrubland Cool conifer forest Temperate conifer forest Water

Pliocene stable experiment

Pliocene stable experiment

Figure 10.7: Predictions of Antarctic vegetation distributions for each of the Pliocene experiments using climatological means derived from the UK Meteorological Office General Circulation Model within a biome model

Figure 10.7: Predictions of Antarctic vegetation distributions for each of the Pliocene experiments using climatological means derived from the UK Meteorological Office General Circulation Model within a biome model

also have existed. For example, the Transantarctic Mountains may have been substantially lower in the past (McKelvey et al., 1991). Kerr and Huybrechts (1999) tested this possibility in an ice-sheet-modelling study and concluded that the surface elevation of individual mountain blocks has only a very local effect on EAIS dynamics. Second, by applying a uniform climate forcing within the model, the possibility for important regional-scale variations in temperature and precipitation, resulting from the differences in Middle Miocene and Pliocene boundary conditions, is ignored. Third, the ice-sheet model used did not include a representation of ice shelves or ice streams. Ice-sheet theory states that the stability of an ice sheet grounded below sea level depends on its surrounding ice shelves that are believed to buffer the ice sheet from oceanic changes (Lingle, 1984). Any weakening of these ice shelves may lead to a runaway process in which the entire marine ice sheet potentially collapses. Support for this hypothesis comes from recent observations of the terrestrial APIS. After the collapse of the major part of the Larsen A ice shelf in 1995, the flow of glaciers, which had fed the former ice shelf, accelerated significantly (De Angelis and Skvarca, 2003). The ice shelves provide a mechanism for linking ocean temperature change with the retreat of ice shelves and grounded ice-sheet instability that may have been important during the Middle Miocene to Pliocene interval.

Current research is using an alternative approach to explore the nature and behaviour of the Antarctic Ice Sheet during the Middle Miocene to Pliocene interval. Instead of attempting to reconstruct the actual state of the ice sheet at a specific time, which is extremely difficult given the incomplete geological record and uncertainties inherent in the techniques used for palaeoenvironmental reconstruction, modelling studies are now focusing on developing ensemble predictions which encompass the range of plausible behaviour of the ice sheet for particular periods. For the Pliocene, a suite of palaeoclimate modelling experiments which cover the range of possible boundary conditions, and which have been validated against global Pliocene palaeoenvironmental proxies, are being used to provide three-dimensional thermomechanical ice-sheet models with the required climatological forcings to predict the extent of the both the Greenland Ice Sheet and EAIS. Ranges of different initial conditions are used within ice-sheet models so that icesheet hysteresis can be explored in detail. The dependence of these results on the GCM boundary conditions and the initial state of the ice sheets leads to a range of ice-sheet reconstructions, but it seems likely that the mean state for the Pliocene Greenland Ice Sheet and EAIS were, to some extent, smaller than they are today (Fig. 10.8). For Antarctica, the model results clearly show that the Wilkes and Aurora Basins are the areas of the EAIS that are most susceptible to increased temperatures. Nevertheless, even in the most

Figure 10.8: Examples of (A) smallest and (B) largest modelled EAIS in equilibrium with reconstructed Pliocene climate.

extreme scenario, the predicted EAIS still covers all but the northernmost reaches of the Wilkes subglacial basin, which is suggested to be a source of the diatoms found within the Sirius Group sediments (Hill et al., 2007). Comparison of the model-predicted ice sheet and sea-level mean state to observed Pliocene highstands (e.g. Dowsett and Cronin, 1990) provides a potential match, which may be improved in future modelling exercises that incorporate variability in orbital forcing.

0 0

Post a comment