Russian Ice Island Upper Air Data (DSI-6355)
Entry ID: gov.noaa.ncdc.C00567

[ Get Data ]
[ Update this Record ]

Abstract: USSR Ice Island Rawindsonde is historical digital data set DSI-6355, archived
at the National Climatic Data Center (NCDC). This is meteorological upper air data from ice
islands in the Arctic Ocean, collected by the old Union of Soviet Socialist Republics (USSR).
The period of record is 1950-91. Major parameters in upper air data sets are: geopotential
height, pressure, temperature, relative humidity, and wind speed and direction.

Purpose: To make
a wide range of climatic data available to researchers and the public. SUPPLEMENTAL INFORMATION:
data set CURRENTNESS REFERENCE: Ground Condition Pricing is dependent on customer order
specifications. Please contact NCDC for information on fees and terms for retreiving the Data
Set or Product.

Related URL
Description: Data from the Integrated Global Radiosonde Archive can be accessed using FTP and
HTTP access methods.


Description: View Source Metadata document from the Metadata Manager and Repository



Geographic Coverage
 N: 90.0 S: 65.0  E: 180.0  W: -180.0

Data Set Citation
Dataset Originator/Creator: National Climatic Data Center, NESDIS, NOAA, U.S. Department of Commerce
Dataset Title: Russian Ice Island Upper Air Data (DSI-6355)
Dataset Release Date: Unknown
Data Presentation Form: tabular digital data
Online Resource:

Temporal Coverage
Start Date: 1950-01-01
Stop Date: 1991-12-31

Location Keywords

Science Keywords

ISO Topic Category

BALLOONS    [Information]
RADIOSONDES    [Information]

ALTIMETERS    [Information]
ANEMOMETERS    [Information]

ESIP >Earth Science Information Partners Program    [Information]

ATTRIBUTE ACCURACY REPORT: The National Center for Atmospheric Research (NCAR) Data
Support Section (DSS) received a version of the Russian North Pole Raobs in November 1999. We
had processed previous versions of this collection, but in earlier versions each sounding was in
two separate parts, a lower part with data at 3km and below and an upper part with data above
3km. It was nearly impossible to insure that the two parts were combined correctly due to
differences in the way times were recorded and the lack of any overlapping data. This new
version included the complete soundings in one part and DSS ran comparisons to other data
sources to see if the parts seemed to be combined correctly. Additional checks on the data were
run by DSS to investigate some of the problems we had noticed in earlier versions. These checks
involved checking to insure that the times on the observations seemed correct after converting
to GTS. The original times on the soundings were a local time that seemed to cause some
confusion with these moving platforms. We checked data by seeing if times looked reasonable, by
comparing times on the upper air data to times on the surface data from these same stations, and
by comparing the data against these same observations from other sources where we could find
them. Time of Observations For the data to be useful for Reanalysis, we must know the time of
the observation in UTC. The information we had on this latest shipment was that the times are a
derived local time, based on using the current longitude and adjusting from Moscow time. For our
adjustments we used a longitude of 37deg 34min for Moscow. To get UTC we used 3 hours as the
Moscow time difference. The actual longitude difference is about 2.5 hours. We assumed the most
likely time for observations based on Russian practice is 1/2 hour before the standard synoptic
hours. We made plots of date versus time of ob to identify the preferred ob time. The plots show
the time recorded with the data (local time) and the time after a conversion to Moscow time.
Separate plots were made adjusting directly to GMT (using 0 longitude) rather than Moscow time.
In general the Moscow plots give better-looking results, but there are a number of questions. A
sample plot is shown in figure 1, and the full set of plots is available in a separate document.
For Raobs prior to July 1957, the main upper air synoptic ob times were 03Z and 15Z. The times
derived from the local times for data in this period seem to be more like 00:30Z and 12:30Z.
Station 04 seems to change ob times to about 13:30Z and 01:30Z in September 1956. Station 06
starts reporting at 13:30Z and 01:30Z (Jun 1956) and switches to 11:30Z and 23:30Z after July
1957. Stations 07 through 11 seem to report at the expected 1/2 hour before the synoptic hour.
Stations 12 through 22 and 28 seem to all report a full hour before the synoptic hour. It is
interesting to note that if you convert local time to Greenwich Time directly for stations 12
through 22 and 28 you do get ob times of about 1/2 hour before the synoptic hour. Stations 26,
30, and 31 have varying ob times and include 1 hour before, 1/2 hour before and right on the
synoptic hour. These plots seem to confirm that the original times are a local time in most
cases. The variations are hard to sort out and probably do not have a major impact on the
usefulness of the data. We accepted the conversions and used the Moscow time as the basis for
converting to GMT. Date-Time Checks The pressure from surface level of the soundings and the
surface pressure from the separate surface observations were plotted versus date-time for each
station. These plots usually show good agreement but they also show periods when there seems to
be a time shift, periods where there is an apparent offset, and periods of general poor
agreement. Plots of the differences in pressure and plots of the pressure values from both
sources on the same graph were made. The former show offsets and other problems while the latter
show the time offsets. Sample plots are given in figures 2 and 3 and the complete set of plots
is available in a separate document. Information on stations with apparent problems follows. For
station 04, there seems to be a number of short periods with one-day offsets. These seem to be
related to dateline crossings, but not all crossings cause a shift. This station has many
dateline crossings. For station 08, observations beginning on 23May1960 seem to have a 1-day
offset. The upper air observations are assigned a date which is one day later than the surface
observations. There is a one-day gap in the upper air data at the beginning of this period which
would indicate the problem is probably in the upper-air data. The problem disappears on
12June1960 and then begins again 22July1960 and ends on 25 August1960. These dates correspond to
dates when the station crossed the dateline so the problem is almost certainly related to the
convention used at the dateline. The local times show no jumps when the station crosses the
dateline so it looks as if the dateline was ignored and only the time-of-day was adjusted to
Moscow time. It is interesting to note that at a later dateline crossing for this same station
the date 17SEP1960 appears twice as one would expect when crossing the dateline moving eastward.
Station 09 seems to have a pressure offset between surface and upper air observations. Beginning
on about 17August1960, the upper air pressure tends to be 1.5 to 2 mb lower than the surface
pressure. The problem continues until 16 October1960. This station also shows a time shift
starting 26FEB1961 and running to the end of the period of record for this station on
11March1961. This again appears to be related to a dateline crossing. Station 14 from 01JUN1965
to about 10JUN1965 seems to have a one-day shift in ob time. It does seem to be related to a
dateline crossing. Station 15 shows a time shift at the beginning of its record, 11May1966 and
continuing through 30June1966. In this case, the upper air appears to have a date one day
earlier than the surface data and there is no related dateline crossing. There is also no gap
near the end so it is unclear how to handle this case. Station 16 shows a time shift starting on
4September1968 and ending on 29 October 1968. Again the shift is linked to a dateline crossing,
however this station has a number of other crossings which don't seem to lead to time shifts.
Two more that do seem to cause shifts are 30June1969-18July1969 and 29-30November1969. Station
26 seems to have a one-day shift, but only for a short period of about 26-28Jul1984. It appears
they crossed the dateline and then corrected for the crossing a few days after the actual
crossing. This could be caused by uncertainty about the exact position. Station 30 has a short
period 20-22May1989 where the upper-air pressures are up to 5mb above the surface pressures.
This occurs after a several day gap in the data and pressures before the gap were in the range
of the ones that look bad. The four or five observations should probably be removed. Station 31
has very poor agreement with the surface pressures for much of its period of record. In general
the pressures are 4 to 5 mb higher in the upper- air observations in the beginning of its record
with agreement improving to about 1-2mb on about 15APR1990. More investigation is needed on this
station. We did some spot compares to the data from GTS and the surface agrees with surface GTS
and upper-air agrees with upper-air GTS. After applying corrections for the shifts detected
above, further checking on poor agreement between surface pressure from the surface observations
and the lowest pressure in the Raob it was discovered that there were many short periods
involving most stations where the dates were out of order after conversion to Greenwich time.
This was traced to the dateline problem, but many of the periods were too short or pressure
changes too gradual to be noticed on the plots. It appears that the longitude with the data may
not be the same as the longitude used when local date-time was determined and stored with the
data. This results in many cases where the day is either one day earlier or one day later
because the conversion to local time assumed the station was on one side of the dateline and our
conversion to Greenwich from local time using the longitude with the data assumed the station
was on the other side of the dateline. Station NP02 required no day corrections. Station NP03
required no day corrections. Station NP04 required 24 day corrections. Station NP05 required no
day corrections. Station NP06 required 1 day correction. Station NP07 required 5-day
corrections. Station NP07 seems to have a number of observations either out of order or with
improper ID lines. This seems to usually occur when the ob frequency changes from twice daily to
four times daily or vice-versa. A day change and moving one ob seems to fix this problem.
Station NP08 required 112 day corrections. Station NP09 required 22-day corrections. This
station also had two identical observations on 18June1962. Also on this same date, a period of 3
mb offset between surface and upper air pressures begins. It ends on about 15 Oct 1960.
Observations with this offset were left in the archive. Station NP10 required no day
corrections. Station NP10 has at least two occurrences where the levels in two consecutive
observations have identical values at 3km and below but different values above 3km. This would
seem to indicate that there was some merging of upper and lower parts and it was not always done
correctly. The observations most likely in error were removed from the set. Station NP11
required no day corrections. Station NP12 required 1 day correction. Station NP13 required no
day corrections. Station NP14 required 10 day corrections. Station NP15 required 1 1/2 day
correction. Station NP16 required 171 day corrections. Station NP16 had two identical obs on
20July1968 and one was removed Station NP17 required no day corrections. Station NP19 required 5
day corrections. Station NP21 required 2 day corrections. Station NP22 required 3 day
corrections. Station NP26 required 10 day corrections. Station NP26 had one ob on 09JUL1983 that
was out of order. It was moved. Station NP28 required 12 day corrections. Station NP28 had one
ob out of order on 20APR1988. It was moved. Station NP30 required no day corrections. Station
NP30 had four observations with very poor agreement with the surface pressure. These
observations, 23Z19May1989 - 11Z21May1989 were removed from the record. Station NP31 required no
day corrections. Station NP31 had one ob out of order on 11Z10FEB1991. It was moved. NP31 has
problems with offset from surface pressures. See earlier comments. Additional information is
provided with the data in file "6355_more_info.doc". LOGICAL CONSISTENCY REPORT: Unknown
COMPLETENESS REPORT: Completeness information not available. LINEAGE/PROCESS STEP: PROCESS

Access Constraints
to ensure that these data are accurate and reliable within the limits of current NOAA quality
control procedures. NOAA can only certify that the data provided to its customers is an
authentic copy of the records which were accepted for inclusion in NOAA archives. NOAA cannot
assume liability for any damages caused by any errors or omissions in the data, nor as a result
of the failure of the data to function on a particular system. NOAA makes no warranty, expressed
or implied, nor does the fact of distribution constitute such a warranty.

Use Constraints


Data Set Progress

Data Center
National Climatic Data Center, NESDIS, NOAA, U.S. Department of Commerce    [Information]
Data Center URL:
Dataset ID: ______DSI-6355 ____

Data Center Personnel
Phone: +1 828-271-4800
Fax: +1 828-271-4876
Email: ncdc.orders at
Contact Address:
151 Patton Avenue
Federal Building, Room 468
City: Asheville
Province or State: NC
Postal Code: 28801-5001
Country: USA

Distribution_Format: ASCII and GIF Files
Fees: Please See Summary

Phone: +1 828-271-4800
Fax: +1 828-271-4876
Email: ncdc.metadata at
Contact Address:
151 Patton Avenue
Federal Building, Room 468
City: Asheville
Province or State: NC
Postal Code: 28801-5001
Country: USA

Creation and Review Dates
DIF Creation Date: 2006-06-27
Last DIF Revision Date: 2011-07-21

[ Update this Record ]

Link to Web Site