A data collection platform was placed near the centre of a 950 m x 800 m near-rectangular iceberg in the Weddell Sea on 4 February 1979. Meteorological conditions and the dynamic behaviour of the iceberg were recorded for every three hours over a one-year period. The following parameters were measured: barometric pressure, wind speed and direction, air and snow temperatures, and iceberg heading, tilt, and strain. Platform elevation was 35 m a.s.l. corresponding to an iceberg thickness of about 210 m. The platform was deployed at lat. 70°5'S., long. 20°3'W, Figure 1 shows the iceberg drift and sea-ice conditions between 4 February and 4 May 1979.

Fig. 1.

Drift track of the iceberg for the period 4 February to 4 May 1979. The berg was surrounded by pack ice from about 6 March.

The tilt sensor was a Singer-Kearfott bubble-type two-axis electrolytic vertical sensing element mounted on a levelling platform. This was fixed on a 2.2 m long, 0.16 m diameter white-painted tube, which was tightly placed in a vertical drill hole in the firn, and frozen in by the addition of water (firn temperature was -5°C). The tilt sensor was level with the snow surface, and aligned so that tilt was measured along and across the main axis of the iceberg. The range of the tilt sensor was ±4.4 x 10-3 rad with a resolution of 3.7 x 10-5 rad (Bø and others 1979). The strain-meter consisted of a 1 m invar rod fixed at the passive end of the instrument to a mounting block and plate, and at the active end to the core of a linear variable differential transformer (LVDT), The LVDT body was mounted in a cage connected to a rezeroing system. The strainmeter was placed in a 1.1 m deep pit and fixed to vertical wooden posts placed 2 m deeper and frozen in. The range of the strainmeter was 1.2 x 10-3 and the resolution 10-7 (Bø and others 1979).

Tilt and strain were sampled 20 times at intervals of 6 sec. This measuring sequence was repeated every three hours. The data were stored and transmitted over the TIROS-N/NOAA satellite, which provided approximately 13 passes per day for transmission of data. Eight of these gave the location of the iceberg. Analysis has been carried out of the tilt and strain data from the instrumented iceberg for the period 4 Fébruary to 4 May 1979.

The causes of iceberg deterioration can be discussed under three broad headings: wave-induced melting at the water line; breaking; and melting at the top, bottom, and sides. A short summary of current understanding under the first two headings is presented. It is then argued, under the third heading, that the melt rate at the sides of a tabular Antarctic iceberg is likely to exceed that at the top and bottom. The behaviour of the entraining plume which forms at the side of an iceberg when it melts in water of uniform salinity is outlined. Another form of convection, occurring when the ambient water is stratified, is then described; in this case the melt water spreads out in a series of almost horizontal layers. Finally, field observations on iceberg melting are discussed.

The melt rates of Antarctic icebergs derived by Morgan and Budd (1978) were based primarily on the distribution of iceberg concentrations, data on iceberg size distribution being limited. Recently, more detailed data have been obtained, especially north of lat. 60°S which allow more precise calculations of the changes in size during drifting. On Australian National Antarctic Research Expedition (ANARE) voyages each summer since 1977, observations of icebergs have included photographs, number per unit area from ship’s radar, widths, heights, and shapes. These data have allowed separate calculations to be made for tabular icebergs which give a clearer indication of the contribution to average size changes of melting, breakage, and rollover. New data from satellite transponders placed on icebergs have provided more definite estimates of the average northward and then eastward drift rates of icebergs near 90°E. Combined with the new data on size more accurate melt rates have been derived. Detailed measurements with expendable bathy-thermographs (XBT) along the shipping routes have provided more definite data on water temperature which allow the melt rates to be studied as a function of temperature. Historical data on iceberg locations have been used to estimate the melt and decay rates of the icebergs at higher temperatures.

An expendable bathythermograph (XBT) survey around an iceberg grounded in 100 m of water shows that iceberg melt-driven convection significantly alters the surrounding water properties in two ways, by a general cooling of the water in the upper 40 m by as much as 3 deg and by the formation of isothermal layers with a 5 m vertical-length scale in the pycnocline. Both of these effects become more pronounced as the distance to the iceberg decreases. The overall cooling of the upper layers supports the upwelling concept of Neshyba (1977) and of Josberger and Martin (in preparation*), while the formation of a step-like structure supports the cell formation idea of Huppert and Turner (1978) and Huppert and Josberger (1980).

Synoptic surface surveys around the iceberg show the existence of melt plumes containing water of a lower temperature than the adjacent water, and these plumes are detectable at distances of 0.5 km.

Concentrated rhodamine placed at depths of 14 and 18 m adjacent to an ice wall sloping down-ward at a 30° angle flowed upward along the ice. The dye reached the surface immediately next to the ice in 240 s and 540 s, respectively, to give a characteristic upward velocity of approximately 0.07 m s−1. The dye then dispersed outward away from the iceberg until it was no longer visible.

Blocks of ice with the proportions of tabular icebergs have been observed melting in water of different temperatures and salinities. The sub-surface shape adopted by the blocks melting in water of the same salinity as sea-water was typically a ‘bath-tub’ one. The basal and mean-side melt rates were of a similar value. The melt rates obtained in the laboratory for icebergs in water of a low temperature match those inferred from population studies of Antarctic icebergs. The melt rate is proportional to the water temperature above the onset of freezing raised to the power 1.5 and melt rates at 18°C are likely to be greater than one metre per day.

Experimental tests were carried out on model icebergs. Natural drift tests with a 1:100 scale model on Brienzer See, Switzerland, show the effects of various environmental conditions. Froude similitude formed the basis for these tests. Other experiments, carried out with a 1:60 model in St. Malo harbour, France, show the need for information obtained at this scale, and include assessments of the drag coefficients for various model shapes with or without insulation skirts on the model, the effects of added mass and rotation of the model, and steering possibilities.

Construction of an iceberg processing plant at Saldanha Bay, Republic, of South Africa, is proposed. A reservoir would be constructed at Rieibaai for ice storage. Tidal forces would be harnessed to pump the warm water of Saldanha Lagoon over heat exchangers (using ammonia or propane gas as a heat exchange medium), thus providing power for electrical generators and for melting ice. A functional analysis of operations is presented, together with proposed costs. It is suggested that the fresh water and electricity produced by this system will cost 6¢/m3 and 3¢/kwH, respectively.