Insights from glacierclimate modelling

Studies to date all indicate the dominant sensitivity of glacial systems to changes in temperature; ablation rates are much more variable than accumulation rates and are the primary control on glacier mass balance and long-term glacier distributions (Oerlemans & Reichert, 2000; Marshall, 2002). On the Greenland Ice Sheet, for instance, a 1°c warming translates to roughly a 50%

increase in present-day melt area extent (Abdalati & Steffen, 2001). Temperature changes have a generally greater impact than precipitation changes because of compounding influences; a warming, for instance, will expand the ablation area, lengthen the melt season, increase the extent of melt at a given site, and increase the proportion of precipitation that falls as rain rather than snow.

The seasonality of temperature changes is important for these influences. Winter warming, for instance, has little effect on annual ablation and does not have an impact on rain versus snow events in most glacierized regions, as it is too cold in winter. In this situation, warming has the potential to increase winter moisture supply, hence net accumulation, although this depends on the synoptic-scale controls of precipitation in a region.

There are local exceptions to the general rule of temperature influences winning out over precipitation variability. Maritime ice masses experience more net accumulation and greater interan-nual variance in precipitation/accumulation rates. Precipitation changes in wet or dry years can be tens of per cent in coastal settings, whereas interannual temperature variability can be buffered by the ocean. This gives precipitation variability a larger role in mass balance fluctuations. In some regions, such as Norway, the strong positive phase of the North Atlantic Oscillation in the 1980s and 1990s has increased regional precipitation and resulted in growth or stability of many glaciers, bucking the global trend in this period (Dyurgerov & Meier, 1997). Even in this situation, however, recent warming has overtaken the precipitation increases in a number of regions (e.g. Iceland) and glacial systems have moved into a negative mass balance regime. Recent warming in many other maritime regions (e.g. Alaska, coastal North America, Patagonia, the high Arctic) has also resulted in near-unanimous ice retreat (Dyurgerov & Meier, 1997; Arendt et al., 2002; Meier et al., 2003).

The most important broad exception to the general temperature control of glaciation is in Antarctica, where it is too cold for significant melt over most of the ice sheet. In this setting, large-scale changes in snow accumulation have an impact on mass balance over a very large area, whereas a modest warming or cooling has very little effect on ice loss. Warming-induced moisture increases in interior Antarctica may increase mass balance in the decades ahead. The Antarctic Peninsula is an exception (see the review by Vaughan, this volume, Chapter 42), as annual temperatures are significantly warmer on the Peninsula relative to the interior plateau of the continent.

For past or future climate-change scenarios, AGCMs can be used to simulate changes in precipitation patterns associated with changing temperatures and surface boundary conditions. In contrast, stand-alone glaciological models that explore temperature perturbation scenarios generally adopt present-day spatial precipitation patterns, using either modelled or observational climatology. A common treatment is to assume that climate warming/cooling will be accompanied by increases/decreases in moisture availability

P(1,9, t) = P (l, 9, 0) exp[[P (T (l, 9, t) - T (l, 9, 0))] (14)

where bP parameterizes the temperature-humidity relationship in accord with the Clasius-Clapeyron relationship. Time t =

0 refers to present-day climatology (precipitation rates and temperature).

The validity of this assumption of a globally consistent positive correlation between increasing temperature and precipitation is not certain, as changes in moisture supply are more complex than a simple thermodynamic control. The variable feed-backs associated with shifts in storm tracks and orographic forcing under different synoptic and topographic conditions are important in long-term glacier-climate forecasts.

Globally, temperature and precipitation are positively correlated (Fig. 32.2a), but the relationship is weak or even negative in some regions. Figure 32.2b plots the zonally averaged correlation between surface temperature and precipitation calculated from the global gridded observational climatology of Legates & Willmott (1990a,b), at 0.5° resolution (V2.01 of global precipitation data and V2.02 of surface temperature data). The average global correlation is 0.52 for annual mean temperature and precipitation, but the relationship weakens at regional scales (see e.g., Leung & Qian, 2003, fig. 9) and at subannual time-scales. The average monthly correlation of the Legates & Willmott data is 0.42. There is a slightly stronger relationship in high latitudes and at cold temperatures, well-illustrated in a continental-scale subset of this data shown in Fig. 32.2c. This figure plots mean climato-logical (1971-2000) temperature versus precipitation for all available climate stations in Canada (N = 1077; Environment Canada, 2003). The linear correlation value of this long-term annual data is 0.48, indicating a positive relationship but also illustrating that much of the variance in precipitation rates is due to factors other than temperature.

I have downplayed the importance of precipitation, as typical climate change scenarios for the decades and centuries that lie ahead predict global temperature changes of a few °C, which will be much more important for glacier mass balance than the predicted precipitation changes, of the order of 10%. However, long-term mass balance forecasts need to consider changes in both precipitation and temperature, as the climatic controls of glacier mass balance differ in all regions and the patterns and rates of anticipated climate change will differ significantly between regions (Church et al., 2001).

Was this article helpful?

0 0

Post a comment