Introduction

Accurate information on precipitation and accumulation is an essential prerequisite for understanding the hydro-logical cycle as well as glacier dynamics. These are also important quantities for estimating future changes of the ice sheet and the sea level, as the greenhouse-induced climatic change takes place. There have already been several attempts to chart the distribution of the annual accumulation of the Greenland ice sheet (Reference DiamondDiamond, 1958; Reference BaderBader, 1961; Reference BensonBenson, 1962; Reference MockMock, 1967). It is, however, worthwhile constructing a new map, because of the recent increase in information from ice cores on the ice sheet and the meteorological data in the coastal regions. For constructing the accumulation map, a special effort was made to obtain solid precipitation data for the coastal stations, which are necessary for calculating the winter accumulation for the lower regions of the ice sheet and glaciers. To assist in the use of the maps, digital information is provided in tables. The present results will be used for estimating the mass balance of the Greenland ice sheet, which will be reported in the near future.

Glaciological And Meteorological Data

The basic information on the pits and cores is presented in Table 1. Altogether, 251 pits and cores are used to calculate the distribution of the annual accumulation on the ice sheet. There are more accumulation data, especially from earlier expeditions, but they are considered to be either too short in terms of the time duration or contaminated by the melt, and thus have been excluded from the present analysis. The period for which the data are used encompasses 77 years from 1913 (de Reference Quervain and MercantonQuervain and Mercanton, 1920; Reference Koch and WegenerKoch and Wegener, 1930) to 1989 (personal communication from F. Nishio). The precipitation data at coastal sites were collected at 35 meteorological stations and are presented in Table 2. The meteorological data are from the following sources: the Danish Meteorological Institute (1954–62, 1969, 1969–82), ESSA (1968), NOAA (1987), and unpublished precipitation data in the archives of the Danish Meteorological Institute. The Canadian meteorological data were obtained from Reference Hare and ThomasHare and Thomas (1974) and Reference Ohmura and MüllerOhmura (1977). The data for Sondre Stromfjord were obtained from the Data Processing Division, ETAC, USAF. The base topographic map is based on the new map of the Greenland ice sheet by Reference OhmuraOhmura (1987).

Sources: Ambach Carrefour: Reference AmbachAmbach (1977); Benson: Reference BensonBenson (1962); GISP D2–6: Reference Dansgaard, Clausen, Dahl–Jcnsen, Gundestrup and C.UDansgaard and others (1985); GISP OHIO: Whillans (1987); GISP A-H/ 1985: Reference Clausen, Gundestrup, Johnsen, Bindschadler and ZwallyClausen and others (1988); all other GISP sites: Reference Radok, Barry, Jenssen, Keen, Kiladis and McInnesRadok and others (1982); Hamilton Northice: Reference HamiltonHamilton and others (1956); Hendrickson: Reference SchusterSchuster (1954); Koch-Wegener: Reference Koch and WegenerKoch and Wegener (1930); Langway: Reference LangwayLangway (1961); Lead dog: U.S. Army Transportation Board (1960); Merc-Quer: de Reference Quervain and MercantonQuervain and Mercanton (1920); Mock: Reference MockMock (1965); Mock-Alf: Reference Mock and AlfordMock and Alford (1964); Mock: Reference Mock and RagleMock and Ragle (1963) and Reference Ragle and DavisRagle and Davis (1962); Müller D. F., V and VT: Reference Müller, Stauffer and SchriberMüller and others (1977); Paterson: Reference PatersonPaterson (1955); Quervain: Reference Quervainde Quervain (1969); Nishio: personal communication.

Values in brackets are solid precipitation.

Because the glacier accumulation is used for estimating annual precipitation, it is in order to discuss the difference between the two quantities. The accumulation is the result of precipitation, drifting, and evaporation. Although accumulation and precipitation are different processes, the numerical values are similar for a number of glaciers (Reference Ohmura, Kasser and FunkOhmura and others, in press). While the evaporative and drifting loss is often considered to yield an underestimation of precipitation, the measured precipitation is more often smaller than accumulation. One of the reasons for a smaller value of meteorologically measured precipitation is, no doubt, the failure to capture snowflakes by the snow gauge. After investigating the annual precipitation and accumulation for 12 glaciers, for which relatively long-term observations of both quantities are available, Reference Ohmura, Kasser and FunkOhmura and others (in press) found that the meteorological precipitation is on average 17% smaller than the glaciologically determined accumulation. Therefore, uncertainty of the order of 20% must be considered inherent in the present results.

Comparison With Previous Maps

The present result for the distribution of accumulation is compared with previously published works which are often quoted in the literature (Fig. 1). They are Reference BaderBader (1961), Reference BensonBenson (1962), and Reference MockMock (1967). The oldest work on this topic by Reference DiamondDiamond (1958) is not used in the present comparison as its content is taken into account by Reference BaderBader (1961), and the mapping therein does not cover the entire ice sheet. In general, qualitative similarities are found between the present work and that of Reference BensonBenson (1961) for southern Greenland and with that of Reference MockMock (1967) for northern Greenland. Major improvements in the present work include depicting the belt of higher accumulation at 1500ma.s.l. on the northwest slope facing Nares Strait; more realistic accumulation data in the ice-cap area south of Inglefield Land; providing the accumulation for the ice cap in Steensby Land; presentation of a more accurate picture of the entire west slope of the ice sheet and the southern ice cap; and especially the correction of previous overestimates for the area of the ice sheet below 2000 ma.s.l. These improvements can mainly be traced to the use of data provided by Reference Quervain and Mercantonde Quervain and Mercanton (1920), Reference LangwayLangway (1961), Reference Quervainde Quervain (1969), Reference Müller, Stauffer and SchriberMüller and others (1977), Reference Ohmura and MüllerOhmura (1977), and Whillans (1987). For calculating more realistic accumulation for altitudes below 2000 m a.s.1., the separation of the annual precipitation into solid and liquid precipitation for the coastal meteorological data played an important role. Overall, the present work also provides the distribution of precipitation and accumulation in high areal resolution which makes it possible to interpret the precipitation distribution for the Greenland ice sheet from a climatológica! viewpoint, as is presented in the following section.

Results And Discussion

The distribution of annual precipitation is shown in Figure 2. The main features of the distribution are summarized as follows: a strong longitudinal gradient exists in southern Greenland, south of 65° Ν on the west and south of 70° Ν on the east slopes; within this region the east coast receives considerably more precipitation than the west coast; the largest precipitation is observed in the southernmost region of the east coast; an extensive area with extremely small precipitation is expected on the north-eastern slope of the ice sheet; there are some local peculiarities, such as the belt of higher precipitation on the middle west slope extending from 69° Ν at 2400 m a.s.l. to the area north of Melville Bay, where it descends to 1500 m a.s.l.; there are very dry patches around Sandre Stramfjord on the west coast and also around Narssarssuaq in southern Greenland. The mean annual precipitation for all Greenland is 340 mm w.e.

Fig. 1. Comparison of various accumulation maps for the Greenland ice sheet; a. Reference BaderBader (1961); b. Reference BensonBenson (1962); c. Reference MockMock (1967); d. present work.

Fig. 2. Annual total precipitation in mm for Greenland. Dots on glaciers are locations of cores and pits. Solid circles are meteorological stations.

The amount of precipitation is regulated primarily by atmospheric conditions, such as stability, water-vapour content, and circulation, often combined with topography. Climatologically important features of the atmos-pheric circulation, leading to the regional variation in precipitation, are more clearly depicted in the interplay between the topography and the monthly resultant wind field, rather than on daily synoptic maps. The resultant wind is a vector mean of instantaneous wind over a certain period. Resultant wind calculated thus becomes mathematically identical to the geostrophic wind computed on the time-mean pressure field. The resultant wind is a convenient concept to use to trace the transport of atmospheric constituents, such as water vapour and pollutants.

Monthly resultant wind is calculated for January and July for the level of 850 hPa (mbar) over Greenland. The 850 hPa level is chosen because it is very close to the mean altitude of the Greenland ice sheet of about 1500 m a.s.l. The resultant wind field is calculated with the geostrophic approximation based on the monthly 850 hPa charts by Reference ScherhagScherhag (1969) modified with additional radiosonde data provided by the National Climatic Center, NOAA, and by the data archives of the North Water Project at the Eidgenössische Technische Hochschule. The January and July resultant wind fields are expressed in terms of streamlines and are shown in Figures 3 and 4, respectively. The concentration of streamlines is expressed as being proportional to the wind speed.

The winter circulation is strongly dominated by two semi-permanent cyclones, the Baffin Bay low to the west and the larger Icelandic low to the southeast. The Greenland ice sheet is located under a weak saddle between the two depressions. This setting determines the main route of water-vapour flow. The southeast coast is directly hit by the onshore flow from the northern flank of the Icelandic low, with relatively high water-vapour content of 2.1 gm⁻³ from the Atlantic Ocean. This flow causes heavy precipitation on the southeast slope so long as the air mass is forced to ascend along the surface of the ice sheet. Once it starts to descend on the vast area north of Summit and on the west slope of the south cap, precipitation will be terminated. The illustration shows that the crest area of the ice sheet is under the influence of the Atlantic Ocean rather than the continental air mass from North America. The area of the west coast, north of 65° Ν also receives the onshore wind from the southwest. The air mass is originally continental, though modified slightly over Davis Strait and Baffin Bay, and is very dry, i.e. 0.7 gm⁻³ water-vapour concentration, and incapable of causing high precipitation. The winter precipitation on the west coast is caused primarily by migrating cyclones entering Baffin Bay from the Atlantic Ocean through Davis Strait.

Fig. 3. Monthly resultant wind stream lines at 850 hPa (mbar) for January.

Fig. 4. Monthly resultant wind stream lines at 850 hPa (mbar) for July.

The summer ciculation over Greenland is dominated by the pressure ridge extending from the northeast towards the centre of the ice sheet. Both Baffin Bay and Icelandic lows remain in their locations. The Polar basin to the north is covered by another low. On the southeast coast, the precipitation decreases somewhat compared with winter, owing to the shift of the streamlines which now run parallel to the slope. On the other hand, the onshore flow on the west coast is loaded with high watervapour content (4.5gm⁻³) and causes the summer peak of precipitation. The air mass (temperature 3°C, and dew point −1°C) reaches condensation level at an altitude of about 2200 m a.s.l. on the mid-west slope, causing major precipitation above this altitude. During the summer, the northwest slope of the ice sheet, facing Nares Strait, also receives up-slope advection from the west and receives some precipitation. These westerlies are the result of the appearance of the low over the Polar basin. The northeast slope of the ice sheet also remains during the summer in the precipitation shadow, both with respect to the southwesterlies and the westerlies, thus receiving the lowest precipitation on the ice sheet. Likewise, Narssars-suaq receives only one-quarter of the annual precipitation of Prince Christian Sund, 150 km to the southeast but on the other side of the ridge extending from the south ice cap. The region around Sondre Stromfjord is located to the north of a weak ridge on the ice sheet which leads to Sukkertopen Ice Cap to the west. The ridge blocks the southwesterlies year round.

The belt of higher preciptiation half-way on the west slope is a natural consequence of the condensation level, as explained in the preceding paragraph and the depletion of water vapour at higher altitude. This is also a common feature in the vertical distribution of precipitation in mountainous regions. This phenomenon is not limited to the area surrounded by 500 mm isolines on the west slope. A close examination of the illustration shows the existence of a maximum precipitation belt all along the northwest to northeast slopes down to Kap Tobin on the mid-east coast. A tendency of the higher precipitation belt to appear is also seen on the west side of the south ice cap. A similar high-precipitation zone docs not show up on the southeast slope. This is probably due to the lack of accumulation data between sea level and 2200 m a.s.l. Some pit observations by de Quervain (Reference Quervain and Mercantonde Quervain and Mercan-ton, 1920) above Angmasalik suggest the existence of higher precipitation below 2000 m a.s.l. Owing to partial melt in the snow profile, his data for this altitude are not taken into account. In addition, a steep surface gradient of the ice sheet on the southeast side makes the occurrence of such a phenomenon less conspicuous.

The streamlines in Figures 3 and 4 also suggest that the sites of the deep ice coring, Dye 3 and Summit, are under the influence of the Atlantic air mass during the entire year, while Camp Century is located more under the effect of the continental air mass from North America, modified by Baffin Bay.

Important topographic barriers are shown in Figure 5, together with geographical names used in the present work. These barriers are not necessarily major ridges in terms of altitude, but they play an important role in dividing ice-sheet surfaces, simply due to the way the relative direction of the barrier is directed with respect to the major stream lines of high water-vapour content.

The distribution of the annual accumulation is given in Figure 6. The overall pattern of the accumulation distribution resembles that for annual precipitation, the main difference being the liquid precipitation subtracted from the annual precipitation for the coastal stations. The greatest accumulation, exceeding 1500 mm w.e., is estimated to occur on the east-facing slope of the south ice cap between Kap Cort Adelaer and Prince Christian Sund.

Fig. 5. Important topographic barriers influencing the accumulation on the Greenland ice sheet: geographical names used in the present work are indicated.

The zone of relatively high accumulation sweeps from the cast slope to the west slope along the south slope of the south ice cap at around 2000 m a.s.l. Another zone of higher accumulation is located from the col between the south ice cap and the main ice cap on the west slope towards Thüle Peninsula. Within this zone, several locations with especially high accumulation are observed: 550 mm w.e. at 2200 m a.s.l. east of Jakobshavn, 650–700 mm w.e. at 1700 m a.s.l. on the slope facing Melville Bay, 200 km east of Thüle AFB. Very low accumulation of less than lOOmmw.e. is found on the northeast slope of the ice sheet and at the lower altitudes less than 800 m a.s.l. east of Sandre Strcmfjord. The ablation area of the outlet glaciers around Inglefield Bredning in northwest Greenland is also estimated to have accumulation of less than 100 m.

The mean annual accumulation on the Greenland ice sheet based on the results given in Figure 2, is 310 mm w.e. for the ice-sheet area of 1.676 × 10⁶ km². Within this definition of the ice sheet, the ice surfaces included are that of the main ice sheet and those of the ice caps which are connected to the main ice sheet through the accumulation areas. Valley glaciers, isolated ice caps, and the ice caps connected to the main ice sheet only through their ablation areas are excluded. The mean annual accumulation on all glacier surfaces in Greenland (1.75 × 10⁶ km²) is estimated at 317 mm w.e.

Fig. 6. Annual accumulation and solid precipitation in mm w.e. for Greenland. Dots on glaciers are locations of cores and pits. Solid circles are meteorological stations.

Acknowledgements

We thank Professor M. de Quervain for providing useful information concerning the loss of mass at several locations in East Greenland. Professor de Quervain took great pains to go through his father’s field book from the Swiss Greenland Expedition, 1912–13. Unpublished meteorological data were made available by the Danish Meteorological Institute. We are indebted to Dr A. Wiin-Nielsen and Mr G. Nielsen. The present work was made possible by financial support from Schweizerischer Nationalfonds zur Förderung der wissenschaftlichen Forschung research grant No. 21–27’449.89 and Eidgenössische Technische Hochschule Zürich research grant No. 41–1010.5 for the Greenland Project.