Future outlook

The apparent consensus above must not blind us to the need for future work. This point can be illustrated by Fig. 83.4 where results from Oerlemans (1993c) and Braithwaite et al. (2003) are plotted against gridded precipitation from the global climatology of New et al. (1999). Any correlation between mass-balance sensitivity could be used to calculate the global distribution of mass-


à \

• Oerlemans 1993a o Braithwaite and others 2003

4A •„ «

o o o o

°o X

Ô •

--------Model 1

u o


Gridded precipitation (m a-1)

Gridded precipitation (m a-1)

Figure 83.4 Mass-balance sensitivity versus gridded precipitation. Mass-balance sensitivity from Oerlemans (1993c) and Braithwaite et al. (2003) and precipitation data from the global gridded climatology of New et al. (1999). The two curves represent different logarithmic functions of precipitation.

balance sensitivity. However, two slightly different models achieve widely different results (Fig. 83.4). Model 1 involves the logarithm of precipitation (to base 10) following Oerlemans (1993a), whereas model 2 involves the logarithm of (precipitation + 1). Model 2 is more 'realistic' than model 1 because it passes through the origin, such that mass-balance sensitivity is zero when precipitation is zero, but model 1 gives a better fit to results for low precipitation. The two models diverge at high values of precipitation and we are left wondering whether the increase of massbalance sensitivity really flattens out for glaciers with high precipitation as indicated by model 1. This is an important question when assessing the contribution to sea-level rise from 'wet' areas such as the Gulf of Alaska.

It is significant that de Woul & Hock (in press) find large massbalance sensitivity for Icelandic glaciers, i.e. up to -2.0myr-1K-1 but data from many more glaciers than shown in Fig. 83.4 are needed to obtain a representative picture of mass balance sensitivity across the full precipitation range. Aside from the new data from Iceland, the approaches of Oerlemans (1993a) and Braithwaite et al. (2003) are unlikely to yield many more points to add to Fig. 83.4. There is, however, a new possibility based on the concept of an elevation on glaciers where a glaciological variable is similar to the area-weighted average of that variable for the whole glacier. For the special case that the glaciological variable varies linearly with altitude, the elevation is the area-weighted mean elevation of the glacier (Lliboutry, 1965). The mass-balance sensitivity was therefore calculated at the mean elevation for the 61 glaciers and compared with the corresponding sensitivity for the whole glacier. There is a strong correlation (Fig. 83.5) suggesting that mass-balance sensitivity need only be calculated at a

Whole glacier

Whole glacier

o Sub-polar • Temperate ° Tropical

At mean elevation o Sub-polar • Temperate ° Tropical

At mean elevation

0 0

Post a comment