NCDC Frequently Asked Questions
What is a climate normal?
[ Return to FAQs ]
The term climatic normal has faced a dilemma since its introduction a century and a half ago. A climate normal is defined, by convention, as the arithmetic mean of a climatological element computed over three consecutive decades (WMO, 1989). "... a normal value is usually not the most frequent value nor the value above which half the cases fall." The casual user, however, tends to (erroneously) perceive the normal as what they should expect. Dr. Helmut E. Landsberg, who became Director of Climatology of the U.S. Weather Bureau in 1954 and, later, Director of the Environmental Data Service, summarized the dilemma quite well over four decades ago (Landsberg, 1955):
The layman is often misled by the word. In his every-day language the word normal means something ordinary or frequent. ...When (the meteorologist) talks about 'normal', it has nothing to do with a common event..... For the meteorologist the 'normal' is simply a point of departure or index which is convenient for keeping track of weather statistics..... We never expect to experience 'normal' weather.
It might be "normal" for the weather to swing radically between extremes from day to day and year to year, but the "climatic normal" is simply an arithmetic average of what has happened at such a "swinging" place. This is why it's important to use a measure of the variability of climate (such as the standard deviation and extremes) in conjunction with the climatic normal when studying the climate of a location (Guttman, 1989).
In accordance with national and international convention, the official climate normals computed for U.S. stations by NCDC consist of the arithmetic average of a meteorological element over 30 years. The "official" normals are provided solely by NCDC, which should be noted in light of other non-official means computations from a myriad of sources.
Is a "normal" the climate you would "expect"?
Climate normals are a useful way to describe the average weather of a location. Several statistical measures are computed as part of the normals, including measures of central tendency (such as the mean or median), of dispersion or how spread out the values are (such as the standard deviation or inter-quartile range), and of frequency or probability of occurrence.
Over the decades the term "normal", to the lay person, has come to be most closely associated with the mean or average. In this context, a "climatic normal" is simply the arithmetic average of the values over a 30-year period (generally, three consecutive decades). A person unfamiliar with climate and climate normals may perceive the normal to be the climate that one should expect to happen.
It's important to note that the normal may, or may not, be what one would "expect" to happen. This is especially true with precipitation in dry climates, such as the desert southwestern region of the United States, and with temperature at continental locations which frequently experience large swings from cold air masses to warm air masses.
Is it appropriate to use normals for predictive purposes?
Normals are best used as a base against which climate during the following decade can be measured. Comparison of normals from one 30-year period to normals from another 30-year period may lead to erroneous conclusions about climatic change. This is due to changes over the decades in station location, in the instrumentation used, in how weather observations were made, and in how the various normals were computed. The differences between normals due to these non-climatic changes may be larger than the differences due to a true change in climate.
What are the differences between normals, standard normals, and long–term means?
Normals refer to the official thirty-year normals computed by the National Climatic Data Center. Standard normals are official normals computed for the World Meteorological Organization based on submission of data by its members. Standard normals are computed every thirty years (e.g., 1931-1960, 1961-1990, etc.). Long–term means are mean values of meteorological elements that are computed for a myriad of reasons by organizations and individuals. Even if long–term means are computed for the normals period (e.g., 1971-2000), only the NCDC values are appropriately called normals.
What is the period of record of the official normals and how often are they updated?
Normals cover a 30–year period of record, and are updated through the end of each decade ending in zero (e.g., 1951-1980, 1961-1990, etc.). Normals are generally computed shortly after all data for the period has been received by NCDC and quality control processing has completed.
Are there differences in the values for the same time period between the 1961–1990 and 1971–2000 normals?
Differences in monthly values during the overlap period (1971–1990) in temperature and precipitation for a given station are possible due to differences in adjustment methods between the 1961–1990 and the 1971–2000 normals. Generally, the 1971–2000 time series has been subjected to adjustments with combined spatial and statistical quality control, whereas the 1961–1990 time series was based less on spatial comparisons and more on limited quality control. See the methdology section under the products page for more information on the 1971–2000 methodology.
Guttman, N.B., 1989: "Statistical descriptors of climate," Bulletin of the American Meteorological Society, Vol. 70, pp. 602-607.
Landsberg, H.E., 1955: "Weather 'normals' and normal weather," Weekly Weather and Crop Bulletin, 1/31/55, pp. 7-8.
World Meteorological Organization, 1989: Calculation of Monthly and Annual 30-Year Standard Normals, WCDP-No. 10, WMO-TD/No. 341, Geneva: World Meteorological Organization.