Answers are grounded in this paper and cite their sources.
[1] The paper by Raper and Braithwaite [2005], referred to here as RB05, addresses an important gap in our understanding of the relation of glacier wastage to sea-level rise. This gap is due to the incompleteness of a world glacier inventory. However, we believe that their synthesis methods, and therefore the results, are not entirely realistic. Meier and Bahr [1996, hereinafter referred to as MB96] attempted a similar synthesis; the two results are not directly comparable, but MB96 estimated a larger volume of glacier ice, considering the same glacier regions, than RB05. One problem is the attempt by RB05 to treat glaciers on a spatial grid, which is only appropriate for spatially-continuous variables such as temperature or precipitation. Calculations of ice-mass volume, however, should consider each mass as a whole entity. Bahr et al. [1997] showed that a power-law relation between area and volume follows directly from scaling theory. For typical values of the scaling parameters the power-law exponent is ≈1.36 for glaciers and ≈1.25 for ice caps. This relation has also been confirmed empirically by many authors. Some large glaciers exceed the size of a 1° × 1° grid cell, and cannot be divided into separate fractions for the calculation of volume using this nonlinear relation. For instance, the Malaspina Glacier in Alaska (≈5 × 103 km2) has a volume ≈2.9 × 103 km3 (water equivalent) whereas if it is fragmented into the five 1° × 1° grid cells and summed, the apparent volume drops 30% to ≈1.9 × 103 km3. We note that this equation has not been tested empirically for glaciers as large as the Malaspina or Bering, but it appears to be appropriate for other large, complex glacier systems including Columbia Glacier (>1 × 103 km2); see the highest point on Figure 6 in MB96. RB05 show no glaciers >212 km2 in area, which may surprise some Alaskans. [2] RB05 synthesized a global result by suggesting that the slope of the relation between fractional area that is glacier-covered and glacier size bin number (n in 2n size bins) is somehow inversely related to topographic roughness (their Figure 2), Only 7 points in the figure were used to form this relation: 2 in the upper left corner (high latitude, polythermal ice caps/icefields and other large glaciers), 3 in the lower right corner (“alpine” areas with small glaciers in temperate latitudes), and 2 in between (north and south Scandinavia, which appear to show rather different relations contrary to the illustration in MB96 for these same 2 regions). It is easy to imagine that other factors might be important in causing this supposed relation. To illustrate the inconsistency of this suggested relation, note that the regions that are the largest contributors to runoff such as around the Gulf of Alaska [Arendt et al., 2002; Meier and Dyurgerov, 2002], are maritime, highly glacierized, have high topographic relief, and much of the ice mass is at relatively low elevations. We suggest that a power-law/exponential distribution of glacier numbers versus size, which can be derived from percolation theory and is supported by data from many hundreds of real glaciers [Bahr and Meier, 2000], would have been a realistic alternative. This form of the distribution function was used in MB96. Bahr [1997] pointed out the danger in truncating the size distribution, if, for instance, it did not include very small and/or very large glaciers as in RB05. [3] RB05 discussed only those glaciers (termed GSIC) that lie outside of Greenland and Antarctica; GSICs total about 522 (RB05) or 540 [Dyurgerov, 2002, 2005] × 103 km2 in area. Their calculation is valuable, but it is incomplete. As they correctly note, an important question is how to account for the huge numbers of glaciers and ice caps that are peripheral to the large ice sheets and that are not included in GSIC. These should not be lumped in with the ice sheets for studies of climate and meltwater runoff, because these “small” ice masses are at low altitude and lower latitude, are warmer and wetter, have very different response times than the ice sheets, and are generally missed in the coarse grids used for Greenland and Antarctic models. Many of the small ice masses in the Greenland and Antarctic regions are currently thinning and sending runoff to the oceans; see, for instance Knudsen and Hasholt [2004]. This current wastage is consistent with the current increase of melting of coastal outlets of the ice sheets; see, for example, Krabill et al. [2004]. MB96 and Dyurgerov [2002] assumed a modest total area for these peripheral ice masses of 140 × 103 km2; Dyurgerov [2005] assumed a larger area (245 × 103 km2) and these numbers still may be conservative. Note that Shumskiy's [1969] estimate for the Antarctic small glaciers (169 × 103 km2, used by Dyurgerov [2005]) did not include East Antarctica or the Antarctic Peninsula. Thus, RB05 included only 73% or 53%, and perhaps even less, of the total area of small ice masses that may now be, or will be, contributing to sea-level rise. If one chooses not to include the ice peripheral to the two ice sheets in a particular calculation, the fact must be recognized that the calculation is a lower bound at best, and possibly a very low bound indeed. [4] Projections of future sea-level rise need glacier and ice cap volume, and the distribution of these volumes (or mean thicknesses) with ice-mass area [Church and Gregory, 2001; Van der Wal and Wild, 2001; Wigley and Raper, 2005]. Unfortunately, these authors have discussed and applied their models only to those glaciers and ice caps (GSICs) outside of the vicinity of Greenland and Antarctica, leaving out a large part of the contribution of ice melt to sea-level rise. RB05 are to be commended for noting this gap in our knowledge, but this gap needs to be more widely recognized and it remains to be seen how the gap can be filled.
No takes yet. Share an insight, caveat, or question.
Meier et al. (2005) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: