Four ice cores from the Ross Sea drainage, Antarctica, show patterns of δ18O variations on a time scale of decades to centuries over the last 1400 years without change in the long-term average δ18O. Century scale δ18O fluctuations in the two cores drilled in the Ross Ice Shelf at Station J-9 (82°23′S, 168°38′W, elevation 60 m) are highly correlated (P < 2 × 10−4). The long isotope record (>30 000 a) of the 1978 J-9 core thus represents local conditions over at least 102 m and on time scales of 100 years and longer.

Regional correlations between the J-9 δ18O records and those from Ridge BC (82°54′S, 136°40′W, elevation 509 m) and the Dominion Range (85°15′S, 166°10′E, elevation 2700 m) are barely significant (P ≈ 0.05 for J-9 '76 and Dominion Range, 580 to 1400 years ago) or absent. The failure to find clear regional isotope trends related to climate fluctuations may reflect the finding that between 1957 and 1982 the area was in the transition zone between areas with opposite temperature trends, and showed little or no temperature change. The fact that the records nevertheless show significant δ18O fluctuations highlights the need to base regional climate reconstructions on a regional suite of ice-core records.

Currently-available evidence of 20th-century warming from glaciers and permafrost is briefly reviewed. The signals are clear and strong: warming of polar firn and permafrost, and mass losses of glaciers at lower latitudes, were most striking towards the middle of the century. The easily observable length-reduction of mountain glaciers confirms the global character of the evolution. A probably intermittent reversal of the trend was observed in places after about 1950.

Mass-balance investigations on glaciers in Svalbard at high latitudes (78°N) show that the ice masses have been steadily decreasing during the period 1950–88. Detailed annual observations have been carried out on Brøggerbreen since 1966 and Lovénbreen since 1967. The mean specific net balances are −0.46 and −0.37 m year−1 water equivalent respectively. Only one year had positive net balance in this period. The cumulative mass lost in the period is then more than 10% of the volume in 1967. Zero net balance would be obtained if the summer temperature was lowered about 1°C or if the winter precipitation increased about 50%. There is a strong correlation between the net mass balance and the height of the equilibrium-line altitude (ELA). Because of the high amount of superimposed ice (10–30% of winter balance) stake readings are necessary to find the ELA. There is no sign of climatic warming through increased melting. The trend analysis of the data from the last 20 years shows stable conditions with a slight increase of the winter balance. The net balance is then slightly increasing and thus less negative than 20 years ago.

Geostatistical methods are applied in the analysis of radio-echo data from Scharffenbergbotnen, Dronning Maud Land, East Antarctica, in order to allow the following investigations: detailed and reliable cartography of subglacial bed topography and ice thickness, comparison of recent ice flow patterns and ice flow during earlier glacial maxima, and mass balance studies in relation to climatic changes.

Baldwin Peninsula, northwest Alaska, is a middle Pleistocene push-moraine complex composed of marine, fluvial, and glaciogenic sediments. The peninsula was formed by three ice lobes emanating from the De Long and Baird mountains and the Selawik Lowlands in the southwest Brooks Range during the Anaktuvuk River glaciation. This glaciation was nearly an order of magnitude more areally extensive than late Pleistocene glaciations in the same region and occurred c. 500 to 600 ka B.P. based on paleomagnetism, and amino-stratigraphic and morphostratigraphic correlations with other numerically-dated northwest Alaskan deposits.

Extensive deposits of massive and laminated clayey silt with striated, faceted stones indicate that local sea level was high as glacial ice reached its maximum extent. Glacio-isostasy does not seem to have been important in maintaining a high relative sea level; therefore we infer that eustatic sea level remained high during ice advance. For this to occur, high latitude glaciation must have preceded the build-up of ice in lower latitudes. The source of moisture for such massive glaciation may have come from submerged Bering and Chukchi shelves, enhanced flow of North Pacific/southern Bering Sea winter storms, reduced intensity of the winter Arctic High, or a combination thereof.

A complete three-dimensional thermo-mechanical ice-shect model for the entire Antarctic ice sheet, including an ice shelf, grounding line-dynamics and isostatic bed adjustment, is employed to simulate the response of the ice sheet during the last glacial-interglacial cycle with respect to changing environmental conditions. To do this, the Vostok temperature signal is used to force changes in surface temperature and accumulation rate and sea level prescribed by a piecewise linear sawtooth function. Model calculations started at 160 ka B.P. In line with glacial geological evidence, the most pronounced fluctuations are found in the West Antarctic ice sheet and appear to be essentially controlled by changes in eustatic sea level. Grounding occurs more readily in the Weddell Sea than in the Ross Sea and, due to the long time scales involved, the ice sheet does not reach its full glacial extent until 16 ka B.p. The concomitant disintegration of the West Antarctic ice sheet is triggered by a rise in sea level and takes around 6000 years to complete. The ice sheet then halts close to the present state and no collapse takes place. This Holocene deglaciation appears to have added 6–8 million km3 of ice to the world oceans, corresponding with an Antarctic contribution to world-wide sea level of 12–15 m.

Decadal oscillations of the ice cover in the Barents Sea are examined for the period since 1950. They are highly correlated with atmospheric circulation when that circulation has an anomalous low pressure over the Barents Sea and Eurasian Basin, while the ice cover is weakly correlated with local air temperature. A feedback mechanism between Barents Sea ice and the atmospheric circulation is suggested; increased cyclonic wind-stress curl reduces cold Arctic flow to the Barents Sea and reduces the sea ice. The reduced ice cover encourages heat flux from the Barents Sea to the atmosphere, tending to reinforce the low pressure. This positive feedback amplifies the oscillations of the air–ice–ocean system driven by external forcing with relatively weak decadal variability. A two-level ocean model, which is driven by prescribed buoyancy flux and wind stresses, confirms that Arctic outflow to the Barents Sea decreases during a cyclonic wind stress.

The purpose of this study is to investigate the interannual variability of winter snow cover in the northern hemisphere, using the satellite-derived monthly snow cover data (NOAA/NESDIS) from 1967 to 1987. An empirical orthogonal function (EOF) analysis is made to see the typical pattern of snow cover variations during winter. EOF1, which represents about 40% of the total variance, shows concurrent snow cover patterns between Eurasia and North America. EOF2, which represents over 20% of the total variance, shows a negatively correlated pattern between the eastern and western parts of the continent in Eurasia and North America.

We examined time series of mean snow cover for key regions of the northern hemisphere to investigate variations depicted in the EOF patterns in more detail. We selected two key regions which represent continental-scale snow variation. One is the eastern part of Eurasia and the other is the western part of North America. The time series of the two key regions show an apparent one-year lag relationship of heavy snow cover years; winters with extensive snow cover over Eurasia tend to be followed by extensive snow cover over North America during the succeeding winters.

A computer-based climate monitoring project is described. Data sets include monthly and annual mean surface temperatures and pressures for occupied stations in Antarctica, the Southern Ocean and South Pacific Ocean; and monthly Antarctic sea-ice extent at each 10° of longitude.

Simple statistical analyses of the data sets reveal a mean warming of ~0.15°C (10 a)−1 since the mid 1950s for Antarctic coastal stations and of ~0.04°C (10 a)−1 since the mid 1940s for the ocean stations. The sea-ice record from 1973 to 1988 reveals that the average northern ice limit has decreased at ~0.23°lat. (10 a)−1. Despite apparently compatible long-term trends of temperature and sea-ice extent, annual fluctuations of temperature and ice extent are highly variable and are not well correlated.

A linear relation between total gas content in ice and the elevation of ice formation (i.e. pore close-off) was obtained from seven shallow ice cores in Mizuho Plateau, Antarctica. The derived relation was applied to the vertical profile of total gas content in a 700 m long ice core at Mizuho Station. A general trend of gradual increase in total gas content was observed from 600 to 200 m in depth, from which toward the upper layer it showed a steep increase. After eliminating the effect of down-slope flow of ice around Mizuho Station, it was estimated that the thickness of the ice sheet decreased by about 350 m at maximum during the last 2000 years. This tendency also appears in the δ18O profile of the same ice core.

With a total of 723 km2 of glaciers (1970) the Cordillera Blanca includes the largest glacier-covered area in the tropics. The climate is characterized by relatively large daily and small seasonal temperature variations as well as by a distinct succession between a dry (May–September) and a wet season (October–April). Since the early 1970s an ablation stake network has been installed on the tongues of the glaciers Uruashraju and Yanamarey. The determination of the equilibrium-line altitude at each end of a wet season was possible, showing a fair correlation with temperature, but not with the precipitation records of the nearby climatological station Querococha. Mean ablation rates at the lowest parts of the glacier tongues are markedly higher during the wet season than during the dry season. Reasons are presumably to be found in the seasonal variation of cloudiness and air moisture rates. Terminus variations of four glaciers in the Cordillera Blanca have been monitored since the early seventies, earlier positions are reconstructed back to 1948 by vertical air photographs. For the glaciers Uruashraju and Yanamarey the terminus positions of 1939 are known from an early map. The general retreat of glaciers in the Cordillera Blanca during the last five decades correlates with the global attitude of glaciers and especially with the attitude of glaciers in other tropical areas. Decreased recession rates with minor advances (1974–79 and 1985–86) are accompanied by lower annual temperatures and preceded and accompanied by years with relatively high annual precipitation sums.

Sea ice is an important factor in controlling the ocean–atmosphere energy exchange in polar regions, and has an important impact on climate. This is because sea ice insulates the relatively warm ocean from the cold winter atmosphere, and has a higher surface albedo than the ocean. In this study the effect of sea-ice transport on the energy exchange between the atmosphere and ocean, and thus on climate, is examined using a coupled energy-balance climate–thermodynamic sea-ice model.

Transport of sea ice produces a thinning of the zonally annually-averaged sea ice in the poleward-most zones, and an extension of the ice edge equatorward. Thinning of the poleward-most ice produces an increase in the annual sensible heat flux from the ocean to the atmosphere, which produces an increase in the mean annual atmospheric temperatures. Sensible heat flux is decreased in the zones that contain the ice edge. However, warming in the poleward-most zones produces an increased meridional energy convergence at the ice edge that offsets the decrease in the sensible heat flux to the atmosphere, resulting in a net warming.

The sensitivity of thermodynamically-varying sea-ice and surface air temperature to variations in solar radiation on the 104 to 105 time scales is examined in this study. Model simulation results show the mean annual sea-ice thickness is very sensitive to the magnitude of midsummer solar radiation. During periods of high midsummer solar radiation between 115 ka B.P. and the present the sea ice is thinner, producing larger summer time leads and longer periods of open ocean. This has an effect on the mean annual sea-ice thickness, but not on the mean annual air temperature. However, the changes in sea ice are accompanied by similar variations in the summer surface air temperature, which are the result of the variations in the solar radiation and meridional energy transport.

The long series of mass-balance data obtained by the Norsk Polarinstitutt on Brøggerbreen for the period 1967–88 has been correlated to climatological parameters from the meteorological station in Ny-Ålesund. The best multiple correlation coefficient was obtained between mass balance and positive summer and autumn temperatures combined with winter precipitations, for which R = 0.90. The regression equation gives a very good agreement between observed and estimated mass balance for Brøggerbreen. A test for 8 years which includes summer long-wave radiation, gives a coefficient of R = 0.98. Based on the good correlation between temperatures at several stations, we reconstructed summer temperatures in Longyearbyen and Brøggerbreen mass balances for 1912–88. The end of the cold period before 1918 is connected with historical observations of the maximum advance of cirque glaciers. After an increase between 1912 and 1920, summer and autumn temperatures decreased slowly. From 1920 this decay has been of −0.7°C. Mass balance has been negative since 1918, and the total mass lost at Brøggerbreen is 34.35 m of water equivalent. For the period with available data in Ny-Ålesund (1969–88), a very slight cooling during the ablation period and an increase in winter precipitation has maintained the reduction of the net balance deficit.

To improve the interpretation of surface cryospheric albedo from satellite sensor data, diurnal measurements of the spectral bi-directional reflectance of a commonly-found fresh-water ice type were made, from which hemispherical reflectance can be derived. The purpose of this study is to document its clear-sky, bi-directional reflectance characteristics in the visible (650–670 nm) and near-infrared (810–840 nm) region, assess the diurnal nature of the reflectance, and quantify the surface anisotropy. Bi-directional reflectances of the re-frozen slush ice measured show a spectral dependence and change significantly with solar zenith angle. Considerable variation occurs at each view angle and among view angles throughout the day. Although diurnal reflectance patterns were similar in both bands, magnitudes varied greatly, being highest in the visible and lowest in the near-infrared region. With the exception of peak saturated (specular) values in the forward scatter direction, bi-directional reflectance was generally highest in the morning when the surface and the illumination were most diffuse in character. The exitance (πN) computed from nadir radiance (N) is compared to the measured hemispheric exitance (M). The πN/M ratios, an index of anisotropy, reveal an anisotropy that increases with increasing solar zenith angle and is more pronounced in the near-infrared region.

Seasonal mean changes in the surface elevation of the ablation zone of West Greenland to 72°N between spring 1985 and summer 1986 are measured using radar altimeter data from the 18-month Geosat Geodetic Mission. Semi-variograms are used to estimate the noise in the data as a function of position on the ice sheet. Mean elevation changes are computed by averaging the elevation differences measured at points where orbits ascending in latitude are later crossed by orbits descending in latitude (or the reverse), with each cross-over difference weighted in proportion to the inverse square of the noise level in the neighborhood of the cross-over point. The mean surface elevation of the ablation zone, relative to spring 1985, ranged from 1.5 ± 0.6 m lower during summer 1985 to 1.7 ± 0.4 m higher during spring 1986.

The winter precipitation at the equilibrium line altitude of a wide selection of modern mountain glaciers is well explained (r2 = 0.81, n = 25) by multiple regression using summer temperature and the logarithm of continentality (distance from a moisture source) as independent variables. Continentality serves as a proxy for inland decreases in energy input through longwave radiation and condensation and increasing energy losses through evaporation and sublimation. Winter precipitation in presently inaccessible areas and former precipitation across broad regions can thus be estimated accurately from two simple variables.

For the first time, a cold ice cover of a summit in the central Swiss Alps has been sampled from the surface to the bed for determining its isotopic composition in δD and δ18O. Results of the analyses show a striking decrease of δ-values with depth. The δ-value differences are greater than those explicable by a direct temperature effect, but a substantial increase in melt water percolation through the firn since the formation of the deepest ice layer may explain the situation of this high-altitude ice.

Ice-sheet advance and retreat chronologies reflect climatic change in a manner that is difficult to decipher. Especially difficult is the placement of records into a chronologic sequence. Multiple age estimates obtained from three stratigraphic positions at a site in Ohio show that organics within deposits of the Miami sublobe, along the southern margin of the Laurentide ice sheet, may be up to 3000 years older than the age of the maximum Late Wisconsin extension of that sublobe. In addition, recent studies on organic accumulations above glacial drift provide bracketing ages for ice recession. When the existing radiometric ages for the Miami sublobe are interpreted with these new radiometric constraints, several fluctuations suggested by prior workers are unsupported. A simpler chronology for the Miami sublobe suggests that in late Wisconsin time the southern margin of the Laurentide ice sheet advanced through Ohio about 22 ka to its maximum extent at 19.7 and remained near there until 15 ka. This is in agreement with newly-refined stratigraphic histories of other Laurentide lobes.

The effect of volcanic emission of acidic aerosols on climate is well documented. The presence of acid droplets in the stratosphere can reduce transmissivity and hence decrease surface temperatures. Since the amount and chemical composition of erupted material has important effects on regional climate, knowledge of past volcanic events is of extreme importance. Detailed glaciochemical records provide the only milieu wherein the geochemistry of paleovolcanic events can be fully documented. We present a detailed sulfate and chloride record from an ice core drilled at site 20 D, 40 km SW of Dye 3 in southern Greenland. The record spans the time period 1869–1984 with chemical analyses of approximately eight samples per year. Time series decomposition and locally weighted scatter plot smoothing techniques were used to extract long term trends from the data so that individual volcanic eruptions could be documented. A number of events identified here have been unnoticed previously and a high percentage of the major chemical signatures documenting these events is associated with large decreases in temperature in the latitudinal zone 60–90 °N. Many authors have pointed out that the amount of volcanic acids such as HCl and H2SO4 injected into the atmosphere has a very important influence on global climate, yet this volcanic input has been difficult to quantify prior to ∼1960. Our data help to alleviate this problem. These individual events can be compared to available frost tree ring data from North America, further establishing a volcanism-climatic linkage.