You are here: Home / FAQ / About Statistics

About Statistics

Time-series data statistics on Water Data for the Nation vs. USGS publications

The statistics generated on this web site may not match values published by the USGS in various interpretive or data reports for a number of reasons. In general, these relate to the fact that the statistics provided on the USGS Water Data for the Nation site are generated from the daily values upon request. As such, it is not possible to exactly mirror the decisions made by the authors of USGS publications. Some of the most common reasons for differences between statistical values are as follows.:

  • The period of record is different between the statistics from the USGS Water Data for the Nation site and the published statistics.
    • The USGS Water Data for the Nation site allows the user to select the period of record for generating long-term statistics. Thus the user-selected period may not match the period used for the published values. Published values are typically generated on the entire period of record. However, in some cases only part of the period or record is used. For example, published discharge statistics may be computed using only the period after a significant change in the regulatory practices or storage of a river system.
  • The year used may be different
    • The USGS Water Data for the Nation site allows the user to select either water year (October 1 to September 30) or calendar year (January 1 to December 31) statistics. Published statistics may be based on either of these years, or sometimes on an entirely separate year such as a climatic year (April 1 to March 31).
  • The treatment of partial record may be different
    • The USGS Water Data for the Nation site automatically removes periods of partial record from statistical computations. For example, months that are not complete are not used for long-term monthly statistics and years that are not complete are not used for long-term annual statistics. In general, this matches the common practice with published statistics. However, at times published statistics may be derived from partial periods.
  • The precision may be different
    • The precision (rounding) shown on Water Data for the Nation for the daily values themselves will usually match any published values. However, the precision of calculated statistics may not exactly match. Water Data for the Nation rounds calculated statistical values to one place greater than the standard precision of any given parameter. The published statistics may be based on a different rounding standard, particularly when the daily values data were not collected to the normal standard precision for that parameter.
  • The statistics themselves are different
    • The USGS Water Data for the Nation site currently has a limited set of available statistics. It may be that the published values are actually totally different statistics. For example, the USGS Water Data for the Nation site only shows statistics based on the daily mean values, whereas the published values may be based on daily median.
  • The published statistics may combine continuous time-series and discrete data
    • Some parameters have historically long periods where only discrete data were collected combined with a more recent period where continuous time-series data are available. Some USGS publications may combine these data using various methods in order to provide a longer period of record for statistical analysis than the continuous time-series data alone provides. A common example of this is when a long historical period of weekly or monthly groundwater level data is combined with a shorter period of continuous data. Future versions of the USGS Water Data for the Nation site should allow for data display and statistics for combined discrete and continuous data, but currently the USGS Water Data for the Nation site requires retrieval/display of each data set independently and displays statistics on continuous time-series data only.