Western Station Data
This Atlas contains monthly means of meteorological observation data from 24 western coastal and island stations for a period that generally includes the early 1950s through 1990. The data are in uniformat files. (See Table 9 for the WMO station names and other information for these stations, and for the starting month and year for which data are available. Use the Atlas interface to see the length of the data record at any station for any of the parameters included in the Atlas.)
Stations were selected based on length of record and coastal location, and to complement the Russian coastal station data and provide circumpolar coastal data. The western station data are from National Center for Atmospheric Research (NCAR) data set 570.0, World Monthly Station Data Climatology. For two stations, Vardo and Isfjord Radio, data are from the NCDC Global Historical Climate Network (GHCN).
Table 9. Western coastal stations on the Atlas. Station location and number are from a WMO station listing rather than from information in data set 570. Starting month and year is for data on this Atlas.
Station Name | WMO Station Number | Latitude (deg. N) | Longitude (deg. E.) | Elevation (m) | Data starting month, year |
AKLAVIK | 71968 | 68.2 | 225.2 | 56 | Jul.1926 |
ALERT | 71082 | 82.5 | 297.4 | 63 | Jan.1951 |
BARROW | 70026 | 71.4 | 203.7 | 13 | Jan.1921 |
BARTER ISLAND | 70086 | 70.1 | 216.4 | 2 | Jan.1948 |
BJORNOYA | 1028 | 74.5 | 19.0 | 16 | Jan.1951 |
CAMBRIDGE BAY | 71925 | 69.1 | 255.0 | 23 | Jan.1941 |
CLYDE | 71090 | 70.5 | 291.4 | 25 | Jan.1943 |
COPPERMINE | 71938 | 67.8 | 244.8 | 22 | Jan.1931 |
EGEDESMINDE | 4220 | 68.7 | 307.2 | 43 | Jan.1951 |
EUREKA | 71917 | 80.0 | 301.1 | 10 | May 1947 |
ISACHSEN | 71074 | 78.8 | 256.5 | 58 | May1948 |
ISFJORD RADIO | 1005 | 78.0 | 14.2 | missing | Jan.1912 |
JAN MAYEN | 1001 | 71.0 | 351.6 | 10 | Jan.1921 |
KOTZEBUE | 70133 | 66.9 | 197.4 | 3 | Oct.1928 |
MOULD BAY | 71072 | 76.3 | 240.5 | 15 | Jun.1948 |
MYGGBUKTA | 99900 | 73.5 | 338.5 | 899 | Jan.1932 |
NOME | 70200 | 64.5 | 194.6 | 11 | Jan.1907 |
NORD | 4310 | 81.6 | 343.3 | 36 | Apr.1952 |
RESOLUTE | 71924 | 74.7 | 265.1 | 67 | Jan.1948 |
SACHS HARBOUR | 71051 | 72.0 | 235.3 | 88 | Nov.1955 |
SCORESBY SUND | 4339 | 70.5 | 338.0 | 65 | Nov.1980 |
THULE | 4202 | 76.5 | 291.2 | 77 | Sep.1946 |
UPERNAVIK | 4210 | 72.8 | 303.9 | 120 | Sep.1873 |
VARDO | 1098 | 70.4 | 31.1 | 14 | Jan.1941 |
NCAR data set 570.0, World Monthly Station Data Climatology, is a product of the Data Support Section at NCAR. Most of these data came from NCDC's Monthly Climate Data of the World. Other data sources were World Weather Records from the Smithsonian Institution, the U.S. Weather Bureau, and the Department of Commerce. Standard parameters are sea level pressure, station pressure, temperature, and precipitation. After 1960, moisture parameters and percent sunshine are available. Documentation by W. Spangler and R. Jenne (available from ftp://ncardata.ucar.edu/datasets/ds570.0/docs/format) provides information on the entire data set, which contains data from more than 3900 stations. Data were quality controlled at NCAR by looking for deviations of greater than four or five standard deviations from the long period monthly mean. These extreme values were then manually inspected and either accepted, or set to missing.
The Global Historical Climatology Network (GHCN, http://www.ncdc.noaa.gov/ol/climate/research/ghcn/ghcnoverview.html) contains monthly temperature and precipitation data. Version 2 contains temperature data from over 7000 stations. They have been quality controlled by a variety of methods, as documented in "Quality Control of Monthly Temperature Data: the GHCN Experience", by T.C. Peterson, R. Vose, R. Schmoyer, and V. Razuvaev (http://www.ncdc.noaa.gov/ol/climate/research/ghcn/ghcnqc.html). Extensive metadata are available. In addition to the quality controlled monthly data, a quality controlled and homogeneity adjusted version is available. This version makes historical data homogeneous with present day observations by adjusting for non-climatic discontinuities [Peterson and Vose, 1997]. The GHCN incorporates data from data set 570.0, and from many additional sources. The GHCN was used for the records on this Atlas from Vardo and Isfjord Radio.
For any given station, the GHCN may have more than one time series [Peterson and Vose, 1997] and each time series taken singly may not cover the entire history of that station (the average number of time series for the stations selected for the Atlas is three). Data set 570.0, World Monthly Station Data Climatology, has only one time series for each station that generally covers the entire history of the station. Therefore we used data set 570.0 for the Atlas, but to obtain some of the benefit of the extensive quality control that GHCN data have been subjected to, we compared the temperature data from data set 570.0 with both homogeneity adjusted and non-homogeneity adjusted GHCN temperature data.
For almost all stations, the data set 570.0 records were identical, or very close, to the non-homogeneity adjusted data in the GHCN. In comparing the homogeneity adjusted GHCN data with data set 570.0 records, we found that the two data sources were often offset by 1 °C to 3 °C early in the record. Data from the two sources agreed for most stations beginning in the late 1950s. For five stations, data set 570.0 values closely agreed with GHCN homogeneity adjusted data for the entire period of record (these stations were Jan Mayan, Kotzbue, Myggbukta, Nord, and Resolute).
To avoid having to choose and piece multiple time series from a single station together, we used data set 570.0 rather than GHCN for temperature (and other parameter) records. Two exceptions are station records for Vardo and Isfjord Radio, where we used GHCN because the GHCN records are considerably longer than that in data set 570.0. Those interested in regional, rather than arctic-wide, analysis of climate trends may wish to use homogeneity adjusted GHCN data from NCDC.
Parameters included on the Atlas are two-meter air temperature, sea level pressure, and precipitation. Precipitation coded as "trace" in data set 570.0 is reported as 0.1 mm in the Atlas uniformat files. Relative humidity is generally available only in the early 1960s, with vapor pressure available after the late 1960s. Time series of other parameters such as cloud amount were not included because they are not available in daily or monthly summarized form. Some hourly data are available from various sources, but summarizing them would require resources beyond those available to us. These sources include a U.S. Air Force data base (DATSAV2) with hourly data from airport reporting stations starting in the 1970s; International Surface Weather Observations, a CD-ROM product available from NCAR, with hourly data from 1982-1997, and NCDC data set 464.0, NCEP ADP Global Sfc Obs, daily Jul 1976-con (although data set 464.0 does not include two-meter air temperature and relative humidity). Other data sets of synoptic observations can be found at NCAR (http://www.ncar.ucar.edu/) and NCDC (http://www.ncdc.noaa.gov/).
The method of observation used at the stations, and the accuracy of the synoptic observations from which the monthly means were compiled, cannot be documented without station histories that include instrumentation types and details of instrument placement. Compiling this documentation is beyond the scope of this project. In general, observation methods would have followed those outlined in WMO publication 8, Guide to Meteorological Instruments and Methods of Observation, and, for NOAA National Weather Service stations in the United States, the Federal Meteorological Handbook vol. I. Station inventories are available from NCDC (http://www.ncdc.noaa.gov/ol/climate/surfaceinventories.html). Unfortunately, the method of calculating the monthly mean temperature at a station is not readily available, and as Peterson and Vose [1997] point out, differences in temperature attributable to calculating the mean using two different methods at a particular station can be greater than the temperature difference between two neighboring stations.
(Thomas Peterson, NCDC, Scott Stephen, NCDC, and Dennis Joseph, NCAR, kindly provided information on data sources)