Western Drifting Stations
Historical drifting station synoptic observations from ice islands, sea ice floes or from ships drifting in pack ice were digitized at NOAA's National Climatic Data Center (NCDC). Data from the Fram were obtained from the German Weather Service. With the exception of some data from drifting ice station T-3 and the Fram that are in COADS (see "Description of COADS" under "Total and Low Cloud Cover" in the gridded fields section of the documentation) these data have not been readily available in the past. Table 18 gives the time period and parameters for each station. Use the Atlas HTML interface to see the length of the data record at any station for any of the parameters included in the Atlas, as well as the drift track of the station.
Data were keyed (digitized) at NCDC from the original or photocopies of the original records, or in the case of T-3, using some previously digitized sources. Each station's data were in a different format. Data were converted to standard units at NSIDC where necessary (with the aid of documentation provided by NCDC), sorted by time, and put in uniformat files (see "Uniformat - uniform format" under "Data Formats"). Generally, positions had to be converted to decimal degrees, longitude to degrees east, and time to UTC (for Maud data and parts of the T-3 data record). For ice stations Alpha, Charlie, and T-3, precipitation was converted from inches to millimeters.
While drifting stations are sometimes referred to as ice islands, an ice island is ice of land origin that has broken off an ice shelf. Most ice islands originate from the Ellesmere ice shelf. Of the drifting stations on this Atlas, only ARLIS II, T-3, and the Russian drifting stations NP-19, NP-22, NP-23, and NP-24 were on ice islands. Typically, stations were on multiyear ice floes.
This section of the Atlas is not a comprehensive catalogue of all western drifting stations. Only the major drifting stations, from T-3 (the first major, non-Russian, drifting station) through the AIDJEX program (the last western drifting stations prior to the SHEBA experiment of 1997 and 1998) are included. Data acquired during the drift of the Norwegian ships Fram and Maud are also included.
Table 18. Western drifting station, date range for data on the Atlas, and all parameters available in digital form. Uniformat parameters are included in uniformat files; all available parameters are included in data files in "raw" form (see text).
Station | Dates | Parameters |
AIDJEX (4 stations) | 4/1975- 4/1976 |
Press, Air Temp, Wind Dir/Spd, DewTemp |
ARLIS I | 9/1960 - 3/1961 | Press, Air Temp, Precip*, Sky Cover, Wind Dir/Spd, Weather, Visibility, Wet Bulb Temp, Cloud Layers, Min Temp |
ARLIS II | 6/1961 - 9/1962 | Press, Air Temp, Precip*, Sky Cover, Wind Dir/Spd, Weather, Visibility, Wet Bulb Temp, Cloud Layers, Min/Max Temp |
Ice Station Alpha | 5/1957 - 11/1958 | Press, Air Temp, Precip, Sky Cover, Wind Dir/Spd, Weather, Visibility, Dew Temp, Cloud Layers |
Ice Station Charlie | 5/1959 - 1/1960 | Press, Air Temp, Precip, Sky Cover, Wind Dir/Spd, Weather, Visibility, Dew Temp, Cloud Layers - Daily Summary: Min/Max Temp, 24 Hour Precip, 24 Hour Snowfall, Snow Depth, Peak Wind Dir/Spd |
T-3 (called Ice Station Bravo in IGY) | 1952-1971 (with gaps) | Press, Air Temp, Precip, Sky Cover, Wind Dir/Spd, Weather, Visibility, Dew Temp, Cloud Layers |
Maud | 9/1922 - 8/1924 | Press, Air Temp, Precip, Sky Cover, Wind D/Force, Weather, Cloud Layers |
Fram | 1893-1896 | Press, Air Temp, Wind Dir, Total Cloud, Low Cloud, Rel Humid.,Dew Temp, Wet Bulb Temp |
*Note: Precipitation data from ARLIS I and ARLIS II are not included in the uniformat files for these stations, because the data were determined to be unreliable by NCDC.
The literature resulting from drifting station studies is extensive. As a starting point, Cabaniss, [1962] summarizes much of the work done at Alpha, Bravo and Charlie. Crary [1958] summarizes characteristics of T-3. The proceedings of the Airlie Symposium on Arctic Drifting Stations [Salter, 1968] compiles research results from all the ice station programs that received funding from the U.S. Office of Naval Research (ONR). Jeffries, [1992] reviews the physical characteristics of ice islands and gives an overview of the history of their discovery and use as research stations.
For a historical perspective, Fletcher [1966] provides an account of early Soviet and U.S. ice station activities with an emphasis on the role of air support. Leary and LeSchack [1996] relates the development of U.S. interest in ice stations for military advantage, culminating in the story of a secret mission to an abandoned Soviet North Pole drifting station. Articles in National Geographic present interesting accounts of life on T-3 [Fletcher, 1953] and ARLIS II [Thomas, 1965].
Ice station data were keyed at NCDC under the direction of J. Elms, in a project initiated in 1992 for COADS. The University of Washington Polar Science Center and NSIDC assisted by helping to locate data, and NOAA's National Geophysical Data Center (NGDC) provided funding. The project is documented in Elms et al. [1993], with additional information in Elms [1999] and Elms [1992]. The objective of the project was to key data that were then missing from COADS.
Data were assembled from all sources that could be located. These included manuscripts, teletype, some digital data, and published data. NSIDC provided NCDC with copies of handwritten logs from AIDJEX (originals are at the University of Washington), and with messages printed on teletype paper for T-3 from 1963 through 1966. The University of Washington provided NCDC with copies of logs from the Maud (published in Sverdrup [1933]). Records from T-3 came from a variety of sources collected at NCDC, including synoptic teletype messages, "plain language" messages, and WBAN hourly surface observations.
Data from the Fram were obtained by NSIDC from the Deutscher Wetterdienst (the German Weather Service) with the assistance of V. Wagner.
Some of the ice stations reported parameters that are not included in the uniformat files (see Table 18 for the complete list of parameters available in digital form). These observations are included on this Atlas in files containing all the data as digitized by NCDC from the original records, and prior to any quality control at NSIDC. Files are in the directory FLOATING_PLATFORMS/WESTERN_DRIFT/ORIGINAL_DATA/. Supporting documentation was scanned and is included with the "raw" data files (00SUMMARY.htm). These additional data may be difficult to use, and we are including them on the Atlas only for completeness and because they are historical records that may not have been published elsewhere.
Records digitized at NCDC included station positions. When necessary, positions were inserted in records at NCDC by locating the needed positions in manuscripts. For some stations (ARLIS II, Ice Station Alpha, Fram, and parts of T-3) we used position data from the data set Arctic Ocean Drift Tracks from Ships, Buoys and Manned Research Stations, available from NSIDC (http://nsidc.org/NSIDC/CATALOG/ENTRIES/G01358.html) for the Atlas instead of positions in the NCDC records. These data are smoothed and interpolated from data sets of observations at the Polar Science Center, University of Washington [Colony and Thorndike, 1984] to give a position every two days. The position nearest in time to the observation time is used in the uniformat file. Sea ice drift speeds in the Arctic seldom exceed a few kilometers per day, so little error is introduced by this practice.
For the four stations named above, the Colony and Thorndike data set provided data that were not readily available elsewhere (for the Fram), or provided a smoother track where data had been recorded only to the nearest degree (for Alpha, ARLIS II, and parts of T-3). Still, at least part of the position record for ARLIS I, Alpha, Charlie, and T-3 is available only to the nearest degree. The positions have not been quality controlled beyond measures taken at NCDC. They are given on the Atlas as they were received from NCDC or as they appear in the Colony and Thorndike data set. The position flag for the western drifting station data is set to "1", even though some positions may have been interpolated (see "uniformat - uniform format" under "Data Formats").
Elms et al., [1993] note that data continuity problems plagued the effort to collect and digitize the arctic ice station records. Date and time were often miscoded, for example. Usually these problems were corrected in quality control procedures at NCDC.
After conversion to uniformat data, station meteorological data underwent further testing at NSIDC. Data points that exceeded minimum and maximum values suggested by a meteorologist were visually inspected and discarded if suspected to be erroneous. Parameters that do not vary smoothly, such as wind direction, wind speed, precipitation, and cloud, were checked for gross errors by plotting the data and looking for evidence of problems such as stuck gauges, or factor-of-ten jumps.
Air temperature was checked for out of range values (greater than 10 °C, less than -55 °C). Relative humidity was converted to vapor pressure, which is less sensitive to temperature, before quality control. For smoothly varying parameters, such as air temperature, pressure, and humidity, a difference method was used to identify bad points. The difference in adjacent measurements, (xn+1-xn), was found. This difference was compared with a "standard" difference, and if the tested point created a difference greater than the standard, it was visually inspected and a decision was made whether or not to discard the point. The standard differences were obtained by finding the maximum differences for each parameter in records from each of the 31 Russian North Pole drifting station data sets. The average of the 31 maximum differences was then taken to produce a "standard" maximum expected difference. The Russian North Pole drifting station data were used because the time series is very long, while the western drifting station data sets are generally too short to provide a reliable estimate of the maximum expected deviation between adjacent measurements.
The difference method was preferred over simply flagging points outside a few standard deviations from the mean. It assumes little about the data distribution, and it was found to work better on temperature data, where large temperature excursions in winter resulted in a standard deviation too large to properly screen data in summer. Figure 10 shows an example of how the difference method worked for pressure data from Ice Station Charlie.
All changes to data digitized at NCDC for the uniformat files were noted, and an inventory of changes is available from NSIDC.
Fig. 10. Data points in red are those flagged as questionable by the difference method of quality control in the record of sea level pressure from Ice Station Charlie.
Fram
The North Polar Expedition of 1893 through 1896 was conducted onboard the Fram. Norwegian researcher and explorer Fridtjof Nansen originated the idea of using a ship, beset in ice and drifting, as a research station after pieces of wreckage from the USS Jeannette, crushed in the ice of the Laptev Sea, washed up on the southwest coast of Greenland after three years. To Nansen, this event suggested the existence of a trans-Arctic current. The Fram, designed specifically for the expedition by naval architect Colin Archer, drifted from a starting point in ice north of the New Siberian Islands to Svalbard in what was later named the Transpolar Drift Stream.
The meteorological observation program for the expedition was the responsibility of H. Mohn. The instruments used, their characteristics, and the observations themselves are in "Meteorology," by H. Mohn, published as Volume VI of the Scientific Results [Nansen, 1900-1906]. Observations were taken every two hours, and recorded with the local time (converted to UTC at the German Weather Service). Following are brief summaries of measurement methods for the parameters for which we have digital data, obtained from Mohn's account.
Wind direction was reported on a 32-point compass scale (that is, points have 11.25 degrees of separation). Wind direction was noted to the nearest compass point in degrees True. Wind speed was measured with a Mohn hand-held anemometer and recorded in meters per second.
Atmospheric pressure was measured with three mercury barometers, one of which served as a standard. The observations were corrected to the height of the mercury at 0 °C and at sea level, and reduced to standard gravity.
Air temperature was measured with a variety of mercury, alcohol and toluol thermometers, as well as two thermographs. When the ship was frozen in the ice, the thermometer for air temperature and the hygrometer were placed on the ice in a Stevenson screen at a height of about 1.2 m.
Air humidity was measured using a psychrometer and a hair hygrometer. Absolute humidity and relative humidity were computed from the corrected readings of the dry and wet-bulb thermometer using psychrometric tables. When the wet-bulb temperature reading was below zero, its reading was corrected following Ekholm's recommendation (as cited by Nansen [1900-1906]).
Cloud amount was recorded on a scale of 0 to 10, and cloud form was noted.
Ship positions were interpolated to the time of each meteorological observation.
Maud
Following the success of the Fram expedition, Norwegian researcher and explorer Roald Amundsen organized an expedition of the Maud to perform a similar drift. The scientific program was lead by H. Sverdrup, and meteorological observations were published in Sverdrup [1933].
The expedition lasted from 1918 to 1925, but most of the observations from 1918 through 1919 were lost when two members of the expedition set forth along the coast, carrying the records. They died in the attempt to reach the Russian station at Dikson, and the records were destroyed by animals [Sverdrup, 1933].
The observations on this Atlas are from 1922 through 1924, when the ship was in the ice. Observations were taken every two hours, and recorded with the local time (converted to UTC at NSIDC).
Air pressure was reduced to 0 °C, standard gravity and sea level. Air temperature was measured by sling thermometer or Assman psychrometer at about 5 m above sea level. Relative humidity was calculated from readings of the Assman psychrometer.
Maud data include "shelter" and "masthead" temperatures. The shelter readings were used for the Atlas uniformat files. The shelter was about 4 to 5 m above the ice surface.
Wind speed observations from the Maud used the Beaufort scale. NSIDC converted the Beaufort force number to the middle of its corresponding range in knots, and then converted to meters per second. The anemometer was mounted about 5 m above the surface of the ice. Wind direction was recorded to the nearest 16 point compass direction (that is points have 22.5 degrees of separation).
Daily precipitation was recorded. Cloud cover was generally recorded six times a day prior to 1 June 1923, and every two hours after then, on a scale of 0 to 10.
The ship position was recorded at noon in tables separate from the meteorological data. NCDC assigned the position nearest in time to the meteorological observation time to each meteorological record.
T-3
The ice island that was later designated T-3 (T stands for Target) was sighted for the first time by North Americans when the crew of an U.S. Air Force B-29 flew over it on 27 April 1947 [Wood, 1956]. Lt. Col. J. O. Fletcher established the station on T-3 in March, 1952. A. P. Crary, Air Force Cambridge Research Center, soon joined the station to conduct geophysical investigations. Over more than 20 years of drift, T-3 served as a base for weather observations and research programs sponsored primarily by the U.S. Office of Naval Research and the U.S. Air Force. In an interesting footnote to the scientific work carried on at the station, complicated issues of jurisdiction were raised after a murder on T-3 in 1970 [Wilkes, 1973]. T-3 has also been called Fletcher's Ice Island and, during the International Geophysical Year (IGY, 1957-1958), the station on the ice island was called IGY-B or Ice Station Bravo.
T-3 was occupied continuously between March 1952 and May 1954 as a weather station operated by the Air Weather Service, and then reoccupied in spring and summer of 1955 to continue geophysical investigations [Cottell, 1960]. Beginning in 1957 the U.S. Weather Bureau took over responsibility for meteorological observations [Crary, 1966]. The station was continuously occupied from 1957 until 1971, with the exception of October 1961 to October 1962 [Wilkes, 1972].
Drifting station positions were obtained by theodolite using the stars in winter and the sun in summer. The Lamont-Doherty Geological Observatory recalculated positions using a computer for the period 1962-1970, and estimated the error in resulting coordinates to be less than 0.05 nautical miles. In 1967, a Doppler-based satellite positioning system was installed. The positions from this system were no more accurate than celestial navigation, but the system allowed more fixes for a smoother track [Hunkins and Hall, 1971].
A device dropped on the ice island in May 1976 allowed the island to be tracked by the NASA Nimbus satellite for four months, and again for a short time in July 1977 [Anonymous, 1977]. The ice island was visible in NOAA polar orbiter satellite images, and in Seasat synthetic aperture radar imagery from 1978 (Figure 11). After 1979 T-3 was not tracked, but in 1984 what was believed to be a remnant of T-3 was sighted at the southern tip of Greenland [Jeffries, 1992]. T-3 had made three circuits in the Beaufort Gyre before exiting the Arctic through the Fram Strait via the Transpolar Drift Stream.
Figure 11. T-3 as it appeared in a Seasat synthetic aperture radar image on 3 October 1978. The island was about 12 km long and 7 km wide. The high backscatter is thought to be due to the low corrugated ridges of rock debris [Fu and Holt, 1982]. Banks Island, in the Canadian archipelago, is at lower right. (Image provided by Benjamin Holt, Jet Propulsion Laboratory, Pasadena, CA).
Alpha, Bravo, and Charlie
Information on ice stations Alpha, Bravo, and Charlie was drawn from Cottell [1960], and Leary and LeSchack [1996]. Alpha, Bravo (the re-occupied T-3), and Charlie were established under Project Ice Skate through the agency of the U.S. National Committee for the International Geophysical Year (IGY). The U.S. Air Force provided logistic support, and the U.S. Weather Bureau directed the meteorological observations. These included upper air data, radiation, carbon dioxide, and ozone studies as well as synoptic observations.
The scientific program included heat budget studies, ice physics, magnetics, oceanography, and geology, and was carried out by a number of U.S. universities and agencies with participation by Canadian and Japanese agencies as well. After IGY, many of these studies continued under sponsorship of the U.S. Office of Naval Research (ONR) and the U.S. Geophysics Research Directorate. One significant bathymetric discovery from the program at Alpha was the existence of the "Alpha Rise," which parallels the Lomonosov Ridge.
Positions were determined from observations of sun, moon, and stars with precision theodolites or transits. The recorded positions were estimated to be accurate within one-half a nautical mile.
Ice Station Alpha was set up in April 1957, and the scientific program lasted from June 1957 to November 1958, when the station was abandoned due to the ice breaking up. ONR had agreed to run Alpha after IGY, and after Alpha's premature abandonment, ONR supported the establishment of Ice Station Charlie in April 1959. Observations at Charlie began in May 1959 and lasted until January 1960, when the station was abandoned due to the ice breaking up.
Ice Station Bravo was established by re-occupying T-3 in March 1957. Signals from the Sputnik satellites were monitored on Bravo for a time in 1957, when the position of the drifting station north of Ellesmere Island was advantageous for reception.
ARLIS I and ARLIS II
The Arctic Research Laboratory Ice Stations (ARLIS I and ARLIS II) were supported by the Arctic Research Laboratory (ARL) established near Barrow, Alaska, by the newly formed U.S. Office of Naval Research in 1947. In 1956, M. Brewer, the director of ARL, and M. Britton, director of arctic research at ONR, were successful in obtaining support for a program to establish drifting stations at relatively low cost. ARLIS I was established by icebreaker in September 1960, but abandoned in March 1961. ARLIS II was established on an ice island in May 1961 and operated until May 1965 [Thomas, 1965].
Of the four ARLIS stations, only data from ARLIS I and ARLIS II were located and keyed at NCDC. ARLIS II swept west in the Beaufort Gyre, and left the Arctic Basin through the Fram Strait via the Transpolar Drift Stream. ARLIS I had a much shorter track.
AIDJEX
The Arctic Ice Dynamics Joint Experiment (AIDJEX) grew out of an initial experiment design prepared by N. Untersteiner and K. Hunkins at the request of the U.S. Office of Naval Research and the U.S. Naval Oceanographic Office. The experiment was designed to collect, for the first time in the West, coordinated measurements of wind stress, ice strain, water motion, and water stress over at least one year, in order to have the right combination of data with which to understand atmosphere and ice interactions. Canadian and U.S. scientists collaborated in planning and carrying out the experiment, which included pilot studies in 1971 and 1972 as well as the main experiment in 1975 and 1976. J. Fletcher served as AIDJEX coordinator, 1970-1971, followed by N. Untersteiner for the remainder of the program (1971-1977). The National Science Foundation, the Office of Naval Research, and other U.S. and Canadian agencies funded AIDJEX.
All AIDJEX science and logistics activities are documented in the AIDJEX Bulletin series. Forty volumes were published between September 1970 and June 1978. Volumes 3, 16, and 17 were devoted to translations from Russian: evidence that AIDJEX scientists were intensely interested in Arctic research performed by Russian scientists. Pritchard [1980] has a wealth of information on AIDJEX science.
The main objective of AIDJEX was to improve understanding of ice dynamics, and meteorological observations were important for estimating wind stress. Observations of wind, temperature, pressure and radiation were taken at the main camp (Big Bear), and at three satellite camps (Caribou, Blue Fox and Snowbird) that were established 60 km from the main camp. Figure 12 is an aerial view of Big Bear. Data were acquired from April 1975 through April 1976.
AIDJEX was the first major project in the Arctic Basin to employ navigation by the highly accurate satellite navigation methods that had become available at that time. All previous station positions were taken by the traditional methods of celestial navigation. The positions of the AIDJEX camps were acquired approximately 10 times per day using the Transit navigational satellite or the NOAA Nimbus F Satellite.
Fig. 12. A view of Big Bear taken from a helicopter at an altitude of about 1000 m on 12 July 1975. The runway, covered with melt ponds as is the rest of the ice floe, runs across the bottom of the image.
Paulson and Bell [1975] describe routine observations taken at the manned camps. At each camp a 10 m tower was instrumented to measure wind speed and wind direction at 10 m, and air temperature at 1 m and 9 m (the 1 m observations are included in this Atlas). The tower was located away from large ridges, and 100 m from the hut housing the data recording instruments.
Wind was measured with a vane and cup anemometer calibrated before and after the experiment. Air temperature was measured with platinum resistance elements inside aspirated radiation shields.
Surface pressure was recorded with two microbarographs at each station, and also measured twice daily using mercury barometers.
Humidity was measured with an Assman ventilated psychrometer three times per day when conditions permitted (that is, when dry bulb temperatures were above than -10 °C). At very low temperatures the difference between dry and "wet" temperature becomes too small to be measured by ordinary thermometers.
Sky cover, cloud type, visibility, precipitation, ablation, and radiation were also measured, but these observations are not included on the Atlas because they are not readily available. See Paulson and Bell [1975] for more information.