Sir,

(Reference Cogley and Adams.Cogley and Adams 1998; hereafter CA) make an attempt to estimate the contribution of glacier melt to global sea-level change. They do this by means of a statistical analysis of mass-balance observations and make some firm statements, most notably that “glaciers in general [made] little or no contribution to sea-level change during 1961–90”. This is in conflict with the recent studies of (Reference Dyurgerov and Meier.Dyurgerov and Meier 1997; hereafter DM) and (Reference Zuo and Oerlemans.Zuo and Oerlemans 1997; hereafter ZO) who report values of about 8 mm for this period (Fig. 1).

Fig. 1.

Volume change of small glaciers and ice caps calculated by Reference Zuo and Oerlemans.Zuo and Oerlemans (1997) for the period 1865–1990. The cumulative change is shown as sea-level equivalent with reference to the year 1960. Results are given for three different initial conditions (ΔT), expressed as the global mean temperature for which the average 1865–95 glacier volume would be in equilibrium minus the actual global mean temperature for 1865–95. The solid line shows the estimate of Reference Dyurgerov and Meier.Dyurgerov and Meier (1997).

In the debate about anthropogenically induced climate change, sea-level rise is one of the hot issues. Glacier melt is generally assumed to make a significant contribution to sea-level rise and it is important to resolve the discrepancy between CA’s conclusion and the earlier studies mentioned.

CA base their conclusion on a statistical analysis of the mean specific balance of glaciers for which observations exist. To arrive at a global number, they use the glacier size distribution of the inventory from the World Glacier Monitoring Service, Zürich. Although this certainly has its limitations, it may provide a reasonable first estimate of the size distribution.

Probably more serious is the bias in the distribution of mass-balance observations over glacier size classes. Considering figure 7 in CA, one can indeed see that the larger glaciers have a less negative balance in the period considered. A problem here is that many of the larger glaciers for which mass-balance observations exist are in Norway. They are located in the same geographic region, and most have a positive or only slightly negative balance. I have made a detailed study of the balance of these glaciers, including inspection of local meteorological records, and it is clear that excessive precipitation during the last 10–15 years is the most important factor here. Although CA’s analysis includes some observations from larger glaciers in other parts of the world, the bias appears too large for the measured balance to be taken as representative. In my view the lack of regional differentiation and proper weighting of mass-balance observations in CA’s analysis leads to too large error bars.

Both DM and ZO, although following quite different approaches, recognise the need for regional differentiation. DM make a careful compilation and extrapolation for mass-balance observations from the period 1961–90. ZO use sensitivities derived from calibrated mass-balance models as the basis for their calculations. By using observed temperature records they show that it is important to include regional and seasonal resolution. The estimates of ice wastage made in these studies are totally independent, but yield remarkably similar results for the period 1961–90 (Fig. 1). In fact, Figure 1 lends some credibility to the calculation for the entire period 1865–1990.

I also want to comment on CA’s suggestion that larger glaciers are significantly less sensitive to temperature change. The temperature sensitivity of mass balance as a function of glacier size cannot be determined by comparing a hemispheric mean temperature signal with unevenly distributed mass-balance measurements. As noted above, many Norwegian glaciers now have a positive mass balance. Analysis shows that this is due to large amounts of precipitation, and that average temperature did not change very much. So including these mass-balance measurements in an estimate of temperature sensitivity of glaciers again creates a strong bias. I therefore cannot agree with CA’s conclusion that larger glaciers are less sensitive to thermal forcing.

There are other arguments against a strong dependence of mass-balance sensitivity on glacier size. In recent years a number of glacio-meteorological experiments have been carried out on small and large glaciers to obtain a better understanding of the relation between glacier mass balance and meteorological conditions. I mention experiments in central West Greenland (two summers; three meteorological stations along a 60 km transect across the entire melt zone; e.g. Reference Oerlemans and Vugts.Oerlemans and Vugts, 1993), on the Pasterze, Austria (one summer; five stations distributed on a glacier about 10 km long; e.g. Reference Greuell, van den Broeke, Knap., Reijmer., Smeets. and Struijk.Greuell and others, 1994, Reference Greuell, Knap. and Smeets.1997),’and on Vatnajokull, Iceland (8000 km²; 12 meteorological stations covering the entire ice cap; Reference OerlemansOerlemans and others, 1999). Together with the operation of automatic weather stations for longer periods of time, these have provided a wealth of data on the melt process without a clear bias towards small or large, dry or wet. The process of analyzing data and subsequent improvement of mass-balance models is still in full swing. A lot of work has already been done, however, and one can say that differences in climate sensitivity for glaciers of different size, inferred by CA, are not supported. The modelling and observational studies make clear that it is mainly the precipitation regime, not the size, that determines the sensitivity: wet glaciers are sensitive, dry glaciers not so sensitive.

Historic glacier fluctuations provide another argument against the idea that larger glaciers are much less sensitive to thermal forcing. One cannot say that larger glaciers showed little retreat compared to small ones. With proper correction for geometric factors, or better, explicit simulation with calibrated ice-flow models, it requires a certain mass-balance sensitivity to simulate the post-Little Ice Age retreat from observed meteorological records. Sensitivities thus found are in good agreement with the results from the meteorological fieldwork and mass-balance modelling referred to above. For further background the reader is referred to Reference OerlemansOerlemans and others (1998) and references therein.

Altogether, I do not agree with the main conclusion of CA that glaciers have contributed little to sea-level rise in the light of an alleged dependence of mass-balance sensitivity on glacier size. On the contrary, the good agreement between the independent estimates of DM and ZO makes it very likely that glaciers made a significant contribution.