Report written by
Jonathan D. W. Kahl
Department of Geosciences
University of Wisconsin-Milwaukee, U.S.A.
kahl@uwm.edu
|
NOTE: This is an early version of an article that appears in the October, 1999 issue of the Bulletin of the American Meteorological Society. To view the published article, click here. |
An historical archive of over 25,000 radiosonde observations from
the formerSoviet "North Pole" series of drifting ice
stations has been compiled and made available to interested researchers.
This archive is the only long-term set of meteorological sounding
data over the Arctic Ocean. The digital archive is a result of
the multi-year, collaborative efforts of a group of U.S. and Russian
scientists and keypunch operators working under the auspices of
Working Group VIII, an area of study within the U.S.Russian Federation
Agreement for Protection of the Environment and Natural Resources.
The archive contains soundings from 21 drifting stations over
the period 1954-1990. The sounding archive is being distributed
by the National Snow and Ice Data Center in Boulder, Colorado.
Return to top
Project History and Data Set Description
1. INTRODUCTION
The Soviet Union launched their first Arctic Ocean drifting research station, North Pole 1 (NP1), in 1937. The pioneering scientists aboard this tenuous sea-ice platform lived in fur-lined tents, produced electricity with bicycle-powered generators (Figure 1), and received a heroes' welcome when they returned to Moscow in 1938. In 1991, the very year that the former Soviet Republics split apart into independent states, unstable ice in the Beaufort Sea forced the abandonment of the last drifting station, NP31. This 54-year-long record of scientific measurements aboard 31 drifting ice platforms in one of the world's harshest environments represents a monumental, lasting achievement by hundreds of dedicated Russian scientists (Belt, 1997).
Radiosonde observations were one of many regular geophysical measurements made at the drifting stations (Figure 2). The majority of these >25,000 historical soundings were archived as bound volumes of handwritten tables at the Arctic and Antarctic Research Institute (AARI) in Leningrad (now St. Petersburg). While these upper-air meteorological measurements were well-known to Russian scientists, they had been given little attention by western Arctic and climate researchers.
In 1989 we initiated an international project to obtain and digitize this exotic set of meteorological soundings measured over the Arctic Ocean. This joint project was conducted under the auspices of Working Group VIII (WG VIII), which is one area of study within the U.S.Russian Federation Agreement for Protection of the Environment and Natural Resources (Tatusko, 1990). The goals of WG VIII are to foster cooperative projects between the U.S. and Russia that will improve our common understanding of the Earth's climate and its sensitivity to natural and anthropogenic climate changes.
The original plan consisted of three parallel efforts: digitizing the tables through manual key-entry at AARI and the Central Aerological Observatory (CAO) in Moscow; delivery of the data to the University of Wisconsin-Milwaukee (UWM); and processing the records into a consistent format at UWM. The second step was not as straightforward as it might seem, as international "snail-mail" was unreliable and access to electronic means of data transfer was then extremely limited in Russia. Instead, a steady stream of floppy disks (over 70 in all) and several magnetic tapes were delivered by a number of "scientific couriers", i.e. scientists traveling to or from Russia as part of WG VIII. Our project was expected to last three years: two years for the manual key entry and an additional year for data processing.
The dissolution of the Soviet Union in 1991 dealt the project an enormous blow. Disastrous economic conditions often postponed or prevented the delivery of paychecks to research laboratories, and many scientific personnel at AARI and CAO were forced to seek outside employment. Between maintaining multiple jobs and enduring long grocery lines, little time was left for tedious key-entry work. Keypunch activities nevertheless persisted, albeit at a drastically-reduced rate. The keypunch operators' ongoing efforts in the face of extreme hardships was in many ways a continuation of the efforts of the polar scientists who performed the challenging series of upper-air meteorological observations aboard the North Pole drifting stations.
During the past eight years a number of related developments provided moral support despite the slow progress of our archiving project. Widespread concern about global warming, and about Arctic climate change in particular, sparked an interest in obtaining and analyzing climate observations over the Arctic Ocean. Several long-term Arctic meteorological data sets were constructed (Kahl et al., 1992; Serreze et al., 1995), including a set containing surface meteorological measurements from the North Pole drifting stations (National Snow and Ice Data Center et al., 1996). These data sets, as well as subsets of the drifting station radiosonde records, have begun to be used in climate research (Serreze et al., 1992; Kahl et al., 1993a,b; 1996; Skony et al., 1994; Zaitseva et al., 1996) and model reanalysis projects (Kalnay et al., 1996). In addition, the Gore-Chernomyrdin commission made great strides in releasing previously-secret data from both U.S. and Russian drifting stations (Belt, 1997).
In mid-1995, with approximately 90% of the manual data entry completed, the decision was made to release the archive to the scientific community. Our 1996 efforts focused on processing and quality-controlling the currently-digitized data. These tasks are now complete and the data set may be freely accessed via the Internet.
2. DATA SOURCES
The Arctic Ocean Drifting Station Sounding Archive was assembled from several different sources. All of these sources are ultimately derived from the original set of handwritten tables kept at AARI in St. Petersburg, Russia. The vast majority of the keypunch activities were done at CAO and AARI. Small portions of the data set were keypunched at UWM and at the National Climatic Data Center in Asheville, North Carolina.
Different portions of the data were received at UWM in varying formats, ranging from ~20,000 soundings in individual floppy disk files to ~700 soundings published as data tables in Russia. The individual data sources are summarized in Table 1. The measurement periods for each drifting station in the archive, as well as number of soundings and file sizes are shown in Table 2. Our sounding archive begins with NP3 in 1954. Radiosonde measurements had been made for seven months in 1950 on the earlier drifting station NP2 (Gaigerov, 1967) but we were unable to obtain these data. Several stations from the late 1960s to the late 1980s are similarly absent. In all, the archive contains over 25,000 soundings from 21 drifting stations spanning the period 1954-1990.
3. MERGING OF DATA SETS
A major part of the data processing effort involved merging the original data sets (Table 1) into a single archive. The primary focus of the merge procedure was blending the lower portions (original data set "3km") together with the upper portions (original data set "Floppy").
Although in theory there should be a unique match between the lower and upper sounding sets, in practice this was not the case for most soundings. Differences in time units often prevented an unambiguous match, as local time, Moscow time and GMT units were interspersed among and within the sparsely-documented original data sets. Inconsistencies in longitude units provided additional problems. In some cases the longitude field ranged from 0-180 with a hemisphere indicator of "E" or "W". The hemisphere indicator was often missing or later determined to be incorrect. In other cases longitudes ranged from 0-360 or from -180 to +180. We employed continuity checks and computed drift speeds to flag and correct many position errors. Discovery of longitude errors was especially difficult when the drifting stations were close to the date line.
A secondary focus in the merge procedure was eliminating redundancy. Many soundings were present in more than one data source, for example those from NP22 and NP26 in sets "Floppy", "Aerostan", and "Condigital" (Table 1). Like the vertical blending procedure described above, this process was not straightforward because of uncertainty in station positions and balloon release times. In many cases, soundings represented in more than one original data source had differing vertical resolution and/or altitude ranges. Considerable time and effort was expended into identifying such cases and blending the sounding data appropriately.
Due to the time, longitude and redundancy problems described above, some soundings were unable to be merged, particularly those from original sets "Floppy" and "3km" (Table 1). As a result, a minority of the soundings extend only to 3 km in altitude, or begin slightly above 3 km. In order to facilitate identification of such soundings, the minimum and maximum altitude of the sounding is included in the header record of each sounding (Kahl and Bacon, 1997). An example of this problem is shown in Figure 3, which contains minimum and maximum sounding altitudes for station NP5. Most of the 731 soundings for this particular station begin at the ice surface, although there are a few that begin at 3 km altitude. Except for a brief period in early 1956, most NP5 soundings extend no higher than 16 km altitude. The data for several other stations similarly contain periods when soundings terminate below the usual, lower-stratospheric vertical extent of radiosonde observations. We hope that, in the future, these high-altitude data can be added to the digital sounding archive.
4. DATA CHARACTERISTICS AND QUALITY CONTROL
a. Station Positions
After the soundings in the original data sources were merged, station positions were plotted along with positions contained in two independent data sets: monthly positions provided by AARI, and positions contained in the drifting stations surface meteorological data set (National Snow and Ice Data Center et al., 1996). There was general agreement found between positions given by the three data sources. For all stations except NP3 and NP26, station positions from National Snow and Ice Data Center et al. (1996) were used because they were interpolated to six-hour intervals and had already been subjected to quality control procedures (I. Rigor, Polar Science Center, University of Washington, personal communication, 1995). For station NP3 we used a detailed table, provided by AARI, of positions corresponding to each sounding release (this resource wasn't available for the other stations). For station NP26 the correct positions were determined to be a "reflection" of the National Snow and Ice Data Center et al. (1996) positions about the international date line.
Position plots for stations NP4, NP6, NP16 and NP22 are shown in Figures 4, 5, 6 and 7, respectively. These four stations alone contain 10,634 soundings, over 40% of the total number of soundings in the archive (Table 2). Position plots for each individual station, as well as the data set documentation (Kahl and Bacon, 1997) are available on the World Wide Web at http://www.uwm.edu/~kahl/Arctic/np.html. Soundings are generally available once or twice daily.
b. Meteorological Data.
Following our experience in assembling the Historical Arctic Rawinsonde Archive (Serreze et al., 1992; Kahl et al., 1993b) we decided to retain all data values present in the individual soundings, including quality indicators. This approach gives users the option of making their own determination of data quality. Nevertheless, we did apply a rudimentary quality control procedure which has been described fully by Serreze et al. (1992) and Kahl et al. (1993b). This procedure consists of identifying "gross" errors such as negative wind speeds, and also "probable" errors such as extreme values. In general, the percentage of non-missing data values that passed the error checks was > 99%. Non-missing pressure, geopotential height and temperature data were available in virtually all of the 933,164 measurement levels contained in the 25,756 soundings in the archive (all stations combined). Missing data percentages for other meteorological variables are shown in Figure 8. Cloud information is available only for stations NP22, NP26, NP28 and NP31. Wind data are available for all stations, although the percentage of available data ranges from 20% for NP3 to 98% for NP11. The percentage of available relative humidity observations (though uncertain in high latitudes, e.g. Elliott and Gaffen, 1991) ranges from 51% to 100%.
5. SUMMARY AND DATA ACCESS
An historical archive of over 25,000 radiosonde observations from the former-Soviet "North Pole" series of drifting ice stations has been compiled and made available to interested researchers. This archive was constructed through an international, collaborative project involving numerous U.S. and Russian scientists and keypunch operators. The project was conducted under the auspices of Working Group VIII, an area of study within the U.S.Russian Federation agreement for Protection of the Environment and Natural Resources.
These upper-air meteorological soundings were measured at 21 drifting
research stations in the Arctic Ocean during 1954-1990. This
unique data set is the only long-term record of upper-air meteorological
measurements made over the Arctic Ocean. The data set is being
distributed by the National Snow and Ice Data Center (NSIDC) in
Boulder, Colorado. For further information contact NSIDC via
electronic mail at nsidc@kryos.colorado.edu, or access the data set directly at
www-nsidc.colorado.edu/NSIDC/CATALOG/ENTRIES/nsi-0060.html.
Return to top
Arctic and Antarctic Research Institute, 1959: Observational data of the drifting scientific research stations "North Pole-4" and "North Pole-5" for the years 1956/57. Leningrad, 951 pp.
Arctic and Antarctic Research Institute, 1960: : Observational data of the drifting scientific research stations "North Pole-6" and "North Pole-7" for the years 1957/58. Leningrad.
Arctic and Antarctic Research Institute, 1962: Trudy 249 of the Institute of Arctic and Antarctic Research: : Observational data of the drifting scientific research stations "North Pole-6" and "North Pole-7" for the years 1958/59. Leningrad, 628 pp.
Belt, D., 1997: An Arctic Breakthrough. National Geographic, 191, 36-57.
Elliot, W.P. and D.J. Gaffen, 1991: On the utility of rawinsonde humidity archived for climate studies. Bull. Amer. Meteor. Soc., 72, 1507-1520.
Gaigerov, S.S., 1967: Aerology of the Polar Regions. Israel Program for Scientific Translations, Jerusalem, 280 pp. (Translated from Russian: Aerologiya Polyarnykh Raionov, Gidrometeorologicheskoe Isdatel'stvo [Otdelenie], Moscow, 1964.)
Kahl, J.D.W. and D.M. Bacon, 1997: Arctic Ocean Drifting Station Sounding Archive Data Base Documentation. Technical report, Department of Geosciences, University of Wisconsin-Milwaukee. (Available on-line at http://www.uwm.edu/~kahl/Arctic/np.html or in printed form from J. Kahl).
Kahl, J.D., M.C. Serreze, S. Shiotani, S.M. Skony and R.C. Schnell, 1992: In-situ meteorological sounding archives for Arctic studies. Bull. Amer. Meteorol. Soc., 73, 1824-1830.
Kahl, J.D., D.J. Charlevoix, N.A. Zaitseva, R.C. Schnell and M.C. Serreze, 1993a: Absence of evidence for greenhouse warming over the Arctic Ocean in the past 40 years. Nature, 361, 335-337.
Kahl, J.D., M.C. Serreze, R.S. Stone, S. Shiotani, M. Kisley and R.C. Schnell, 1993b: Tropospheric temperature trends in the Arctic: 1958-1986. J. Geophys. Res., 98, 12825-12838.
Kahl, J.D.W., D.A. Martinez and N.A. Zaitseva, 1996: Long-term variability in the low-level inversion layer over the Arctic Ocean. Int. J. Climatol., 16, 1297-1313.
Kalnay, E. and coauthors, 1996: The NCEP/NCAR 40-year reanalysis project, Bull. Am. Meteorol. Soc., 77, 437-471.
National Snow and Ice Data Center; Applied Physics Laboratory; Polar Science Center; Arctic and Antarctic Research Institute, 1996: Arctic Ocean Snow and Meteorological Observations from Drifting Stations, 1937, 1950-1991. CD-ROM available from nsidc@kryos.colorado.edu. Boulder, Colorado: NSIDC, University of Colorado at Boulder.
Serreze, M.C., J.A. Maslanik, M.C. Rehder, R.C. Schnell, J.D. Kahl and E.L. Andreas, 1992: Theoretical heights of buoyant convection above open leads in the winter Arctic pack ice cover. J. Geophys. Res., 97, 9411-9422.
Serreze, M.C., R.G. Barry and J.E. Walsh, 1995: Atmospheric water vapor characteristics at 70 degrees N. J. Climate, 8, 719-731.
Skony, S.M., J.D.W. Kahl and N.A. Zaitseva, 1994: Differences between radiosonde and dropsonde temperature profiles over the Arctic Ocean. J. Atmos. Ocean. Technol., 11, 1400-1408.
Tatusko, R.L., 1990: Cooperation in Climate Research: An Evaluation of the Activities Conducted Under the US-USSR Agreement for Environmental Protection Since 1974. U.S. Dept. of Commerce, National Oceanic and Atmospheric Administration, Washington, DC, 121 pp.
Zaitseva, N.A., S.M. Skony, and J.D. Kahl, 1996: Temperature inversions
in the Western Arctic from radiosonde data. Meteorology and
Hydrology, no. 6, 11-24.
Return to top
Acknowledgements. Portions of this work were sponsored
by the National Oceanic and Atmospheric Administration's (NOAA)
Climate and Global Change Program, the National Science Foundation,
and the Electric Power Research Institute. We thank Gus Shumbera
(NOAA, National Climatic Data Center) for securing additional
support at a particularly critical time. We gratefully acknowledge
the assistance provided by Renee Tatusko, U.S. Executive Secretary
of Working Group VIII. Ms. Tatusko coordinated the travel of
WG VIII scientists and thus directly arranged the delivery of
much of the drifting station data. Ignatius Rigor (Polar Science
Center, University of Washington) contributed the position data
for most of the stations, and Dennis Joseph and Roy Jenne (National
Center for Atmospheric Research) provided many helpful insights
concerning data quality. This project benefited greatly from
the efforts of a large number of other individuals during the
past eight years, and we apologize if we neglect to specifically
acknowledge certain people. We are particularly grateful for
the assistance provided by Roger Barry, Bob Etkins, Chuck Hakkarinen,
Bill Murray, Joe Fletcher, Carol Waldvogel, Brian Holt, Donna
Charlevoix, Dawn Redhail, Chris Kuk, Jennifer Galbraith, Claire
Hanson, Tony Hansen and Dave Rasmussen.
Return to top
Table 1. Summary of individual data sources.
Click here for Table 1.
[Hit your brower's "back" button to return to text]
Table 2. Measurement periods, number of radiosonde soundings,
and file sizes for each North Pole station in the archive.
Click here for Table 2.
Return to text.
Figure 1. The tent that housed the four scientists aboard the
first drifting station, North Pole 1, during ten months of drift
in 1937-38. The tent was made of seven layers of canvas, down
and fur. Note the human-powered generator and the meteorological
instrument shelter to the right. These items are on display at
the Arctic and Antarctic Museum in St. Petersburg, Russia.
Click here for figure (260K - don't worry, only
the first two figures are large files).
Return to text.
Figure 2. Radiosonde releases were particularly difficult in the
harsh environment of a drifting ice station. On-site production
of hydrogen gas, inflation and release of the balloon, and preparation
of the sonde instrument were among the most challenging tasks
faced by aerologists aboard the drifting stations. This painting,
on display at the Arctic and Antarctic Museum in St. Petersburg,
Russia, depicts a radiosonde launch aboard a drifting station.
Click here for figure (260K - don't worry, only
the first two figures are large files).
Return to text.
Figure 3. Minimum (lower line) and maximum (upper line) radiosonde
altitudes for drifting station NP5.
Click here for figure.
Return to text.
Figure 4. Sounding positions (one dot per sounding) for station
NP4. The positions contained in our archive are plotted alongside
monthly positions provided by the Arctic and Antarctic Research
Institute (crosses) for verification. The archive contains 2,414
soundings for station NP4 during the period 20 April 1954 - 14
April 1957. Similar position plots for all stations are available
on the World Wide Web at http://www.uwm.edu/~kahl/Arctic/np.html.
The map background is oriented such that the Greenwich meridion (0 degrees E) is the vertical line
at the bottom-center of the plot, and the International Dateline (180 degrees E) is the vertical line at
the top-center of the plot. The 70 degree N and 80 degree N parallels are also shown. Longitude
intervals are 10 degrees.
Click here for figure.
Return to text.
Figure 5. As in Figure 4, but for station NP6. The archive contains
1,780 soundings for station NP6 during the period 16 May 1956
- 31 March 1959.
The map background is oriented such that the Greenwich meridion (0 degrees E) is the vertical line
at the bottom-center of the plot, and the International Dateline (180 degrees E) is the vertical line at
the top-center of the plot. The 70 degree N and 80 degree N parallels are also shown. Longitude
intervals are 10 degrees.
Click here for figure.
Return to text.
Figure 6. As in Figure 4, but for station NP16. The archive contains
1,910 soundings for station NP16 during the period 1 June 1968
- 21 March 1972.
The map background is oriented such that the Greenwich meridion (0 degrees E) is the vertical line
at the bottom-center of the plot, and the International Dateline (180 degrees E) is the vertical line at
the top-center of the plot. The 70 degree N and 80 degree N parallels are also shown. Longitude
intervals are 10 degrees.
Click here for figure.
Return to text.
Figure 7. As in Figure 4, but for station NP22. The archive contains
4,530 soundings for station NP22 during the period 1 May 1974
- 28 February 1982.
The map background is oriented such that the Greenwich meridion (0 degrees E) is the vertical line
at the bottom-center of the plot, and the International Dateline (180 degrees E) is the vertical line at
the top-center of the plot. The 70 degree N and 80 degree N parallels are also shown. Longitude
intervals are 10 degrees.
Click here for figure.
Return to text.
Figure 8. Percentages of available meteorological data for each
North Pole station in the sounding archive. Non-missing pressure,
geopotential height and temperature data (not shown) are present
in > 99% of all measurement levels. Cloud information is available
only for stations NP22, NP26, NP28 and NP31.
Click here for figure.
Return to text.
