Sir,
Koerner (1997) presents an interpretation of ice-core data collected from the circumpolar Arctic. As a member of a Russian research team, I was involved in the coring and study of some of the ice cores referenced in his paper. The results of our research were published in the Russian scientific literature. This letter presents relevant data not included in Koerner (1997), and corrects a few inaccuracies. This additional information should contribute to a clearer understanding of the complexity of the ice-formation processes in glaciers subjected to intensive melting.
Data
Lomonosovfonna. Three locations were studied along the ice divide on Lomonosovfonna. In the southern part of the plateau at 1025 m a.s.l., Zinger and Mikhalev (1967) determined the average accumulation rate to be 988 mm w.e. a-1 for the period 1957-64. The 201 m ice core was recovered in 1976 in the northwest part of the plateau at about 1080 m a.s.l.; results were published in Gordienko and others (1981). Because the thickness of the ice is unknown, the time-scale based on stable isotopes remains questionable. The firn-ice transition at the 1976 drilling site was in the depth range 27-32 m (Zagorodnov, 1985). In addition, a 135 m ice core was recovered in 1982 from a site (about 1000 m a.s.l.) approximately 10 km south of the 201 m core. This borehole was plumb from surface to bottom (Zagorodnov and others, 19841. Melt features comprised 34% of the glacier sequence, and the 1975-82 balance was 620 mm w.e. a-1; for the period of 1962-82 the average balance was 658 mm w.e. a-1 (Zagorodnov and Samoylov, 1985). Note that Koerner (1997, table 1) gives the balance for the 1976 drilling site as 820 mm w.e. a-1.
Table 1. Mass balance at Austfonna summit
Austfonna. The elevation of the drilling site on the Austfonna summit was close to 750 m a.s.l. (Dowdeswell, 1986; Dowdeswell and others, 1986; Arkhipov and others, 1987; Dowdeswell and Drewry, 1989). The 566.7 m ice core was recovered from a plumb borehole. Mass-balance data for the drilling site are presented in Table 1. Assuming that the tritium peak at 14 m was from 1963 (Vaykmyae and Punning, 1989), the average (27 year) mass balance was 526 mm w.e. a-1; without tritium data the average (5 year) balance was 598 mm w.e. a-1. Both these values are lower than 794 mm w.e. a-1 given by Koerner (1997, table 1).
The melt scale for the Austfonna record appears to be incorrect and the curve given by Koerner (1997, corrected fig. 4) appears shifted. Using data from Tarussov (1992), and assuming that the profile was digitized from Zagorodnov and Arkhipov (1990) or from Kotlyakov and others (1990), with a net accumulation of 794 mm w.e. a-1 (Koerner, 1997, table 1), the Austfonna melt record presented by Koerner (1997) cannot be reproduced; melt values exceed the net accumulation. A similar problem exists in the Akademii Nauk melt record. The original percentage melt record was obtained from two ice cores, one drilled in 1985 (0-561 m) and the other in 1987 (250-761 m). These melt records are internally consistent. Using the melt percentage profile (Kotlyakov and others, 1990) and assuming annual net accumulation of 315 mm w.e. a-1, again the values presented in Koerner (1997, corrected fig. 4) cannot be reproduced. In order to obtain the best approximation of original data, the melt values have to be multiplied by a factor of 1.22 and the curve has to be shifted down so that the last peak on the right equals 19 g cm-2 a-1. In spite of these errors, the general trends and major elements of the profiles are consistent with the original data; therefore, it appears that inaccurate graphs were produced due to digitization of graphs that were originally published as rather small figures.
Discussion
In some cases, representation of melt data in g cm-2 a-1 units could obscure details of the nature of the ice-formation process. Specifically, Austfonna, Vestfonna (Nordaustlandet) and Akademii Nauk ice cores have high (75-100%) concentrations of melt features (infiltration and infiltration–congelation ice in Shumskii’s (1964) notation) in the upper 50-60 m. Below 350 m, the Austfonna ice core consists of 100% infiltration ice, which means that in the past the ice was formed under extensive melting conditions with a thin or no firn layer at the glacier surface. Similar conditions were described by Koerner and Paterson (1974) on Meighen Ice Cap, where the ice accumulation by melting and refreezing of snow was interrupted several times in the past, and there were several periods of negative mass balance associated with a warm climate. Zagorodnov (1985) reported that runoff from the infiltration–congelation zone (firn thickness < 12 m) is possible when the concentration of infiltration ice exceeds 75%. Therefore, at high concentrations of infiltration ice, runoff is possible and accumulation can be less than precipitation. The presentation of melt as a fraction of the depth interval (which may or may not be equal to annual accumulation), as proposed by Koerner (1977), creates a universal parameter which does not depend on accumulation, and probably does not depend on the vertical deformation of ice layers with depth. It therefore can represent the climate history of glaciers subjected to melting.
The possibility of an ice cap building up to the thickness and temperature state of Austfonna during the late Holocene was studied using a numerical model (Larina and Zagorodnov, 1989b). Using a variable balance (0-500 mm w.e. a-1), allowing runoff in earlier stages of glacier growth, and assuming surface temperatures ranging from 0° to -12°C (determined using an infiltration and infiltration–congelation ice-concentration–summer-temperature transfer function (Zagorodnov 1985)), a 560 m thick ice cap could form in 3000 years. Its growth begins once it attains a thickness of 50 m. The model suggests that the growth of the Austfonna ice cap was completed about 1000 years ago. Therefore, only the top section of the ice core (above 350-400 m) represents regional climatic changes, and the bottom section (400-566.7 m) records changes in surface elevation and local climate. The high dust content below 350 m in the Austfonna core (Zagorodnov and Arkhipov, 1990) supports the idea of snow/ice-free land on Nordaustlandet which could reflect warmer climate and/or lower precipitation. High concentrations (>75%) of melt features below 400 m also suggest the likelihood of negative mass balance creating gaps in the stratigraphic record. Domination of melt ice in the bottom (100% in the depth interval 170—200 m) of Vestfonna ice cap on Nordaustlandet (Zagorodnov, 1985) also supports our suggestion that there has been recent glacier regrowth subsequent to the Holocene thermal optimum.
Borehole temperatures indicate that the Little Ice Age was the coldest period recorded in Eurasian Arctic ice caps. At the very end of that period (110-130 years ago), Austfonna and Akademii Nauk near-surface temperatures (15 m depth) were 6-10°C colder than at present (Larina and Zagorodnov, 1989a, b). Similar changes were recently found in the Windy Dome (Franz Jozef Land) borehole (unpublished data, this author). In general, the new interpretation (Koerner, 1997) of Eurasian Arctic ice-core data is in agreement with our vision of the climate conditions and state of glaciation during the Holocene (Kotlyakov and others, 1990).
Acknowledgements
The author is grateful to E. Mosley-Thompson for the clear formulation of this paper.
