National Oceanic and Atmospheric Administration Nineteenth Century U.S. Climate Data Set Project

• Objectives and Strategy
• Project Background and Accomplishments
• Time Periods Covered
• Maps of Station Distrubution
• Request Preliminary Data and/or Metadata
• Data and Metadata File Format Descriptions
• Future Plans and Current Efforts
• References
• Bibliography of Data Sources

The National Climatic Data Center (NCDC) , by virtue of an inter-agency agreement with the Department of Energy (DOE), has developed a number of data sets to document climate fluctuations and detect long-term climate change in the United States and the world. These long-term data sets include the U.S. Historical Climatology Network (USHCN) and the Global Historical Climatology Network (GHCN) and contain monthly-scale data for several thousand stations covering the last 150 to 200 years. Most of these data, however, are concentrated in the Twentieth Century, with the Nineteenth Century under-represented.

In these data sets (and in global change data sets compiled by other researchers), large areas of the United States have no data for the Nineteenth Century simply because it has not previously been available in digital form. For example, a national-scale data set (the Climate Division Data Set), which NCDC uses to document temperature, precipitation, and drought variations over the contiguous U.S., begins in 1895.

Incorporation of the data from this project into the USHCN and GHCN data sets would yield a more accurate assessment of climate variations over the last two centuries which can be utilized in global change reports such as those issued by the Intergovernmental Panel on Climatic Change (IPCC) and in analyses and reports generated by the Global Climate Perspectives System (GCPS). Incorporation of these data into the U.S. Climate Division Data Set would extend the national perspective backwards in time to cover climatic periods in the Nineteenth Century that were comparable to the 1930's drought and early 1980's wet period. This expanded perspective would enhance the review of U.S. climate anomalies made on a monthly basis by NCDC in the Climate Variations Bulletin publication.

to the top...

Overall Objectives and Strategy

• Overall Strategy
• Rules Followed
• Gross QC Methods

to the top...

The objectives of this project have been to:

• Digitize published data from the Nineteenth Century and
• Transform the data into a form usable by the global change research community and United States policy makers.

Once the data have been validated and adjusted for inhomogeneities and adequate metadata have been compiled, these data may be incorporated into other national and global data sets (USHCN, for example) for use in climate monitoring and research activities.

The overall strategy has been to...

to the top...

• Identify the data (states, station names, locations)
• Remove gross errors (documented and detected)
• Determine the value/use (ex., spatial coverage)
• Expand coverage (ex.,locate period of record for stations now included)
• Enhance metadata (ex., document means computation method; published footnotes)
• Identify/assign station history numbers; Append to TD3220

to the top...

There are, of course, certain rules we are following throughout this effort. They are to:

• Standardize Station Names to the most recent and document aliases by year
• Document Latitude, Longitude, and Elevation by year and encode the source:
  • Published with the data
  • Published elsewhere with same or similar data
  • Anywhere else (atlas, etc.)
• Remove (to another file) and replace questionable values and encode reason
• A Mean Temperature with days missing is not acceptable
• If monthly and annual (compilation) tables (in the publications) disagree...
  • Keep the compilation latitude/longitude
  • Keep the monthly data values
• Accept published corrections (errata) and indications of inaccuracy

to the top...

The gross QC methods we have used are:

• State-wide Period of Record time series - were able to detect:
  • Errors in location
  • Gross keying errors
  • Published errors
• Map plots - identified:
  • Co-located (renamed) stations
  • Errors in Latitude or Longitude
  • Station moves
  • Multiple stations with same name
• Random (and in some subsets, comprehensive) spot checks - detected keying errors
• Comparative (with other sources of same data) QC - detected published errors

Project Background and Accomplishments

to the top...

- as of 2/10/97:

• Monthly station data (1822-1900) have been keyed
• Station name lists (1822-1900) have been compiled
• Station name corrections made and duplications (1822-1889) removed
• Metadata files (1822-1889) have been compiled: latitudes, longitudes, and elevations; aliases; coop station numbers; errors and suspect values)
• Data (1822-1900) and Metadata (1822-1889) are available from NCDC Customer Service

Comprehensive station lists have been compiled for the entire period. Other metadata files include station location (latitude and longitude and elevation) as well as station name aliases, known errors and questionable values. These metadata are used to determine the uniqueness of the station and to aid in the assignment of station numbers. The data from 1822 through 1889 have been corrected for station name changes and duplication. These preliminary data are available from NCDC Customer Service. Descriptions of the data and metadata files' record formats are available in a readme file.

Future Plans and Current Efforts

• Complete metadata files (1890-1900)
• Rework Gross QC for Period of Record (1822-1900)
• Data Analysis/Bias Detection

The dataset will be evaluated concerning its inclusion into the USHCN. Data will be provided to researchers for possible inclusion in their respective long-term research data sets. Data will be placed on-line for Internet access.

The evaluation is required to determine how the Nineteenth Century data need to be adjusted to be comparable to Twentieth Century data. For example, typical Nineteenth Century instrument exposure and practices consisted of attaching the thermometer to the outside north wall of a building and taking temperature readings three times daily (usually in the morning, early afternoon, and evening). These thrice-daily measurements were then averaged to compute daily mean values, which were averaged to compute a monthly mean value.

The implementation of maximum/minimum thermometers and Stevenson and Cotton Region shelters late in the Nineteenth Century may have introduced a non- climatic change into the data. Other researchers (Chenoweth, 1992) have determined independently that some Nineteenth Century data may have a warm dry bias.

Chenoweth, M., 1992, personal communication.

Karl, T.R., C.N. Williams, F.T. Quinlan, and T.A. Boden, 1990:  
     United States Historical Climatology Network (HCN) serial
     temperature and precipitation data, Oak Ridge National
     Laboratory, Carbon Dioxide Information Analysis Center,
     Environmental Sciences Division Publication No. 3404, Oak
     Ridge, TN 37831-6285.

Back to the National Climatic Data Center Home Page