Long time-series of climatic data are subject to several types of uncertainties. Systematic or random errors in the measurement instruments make up only part of the total uncertainty. Changes in instrumentation, measurement practices, reporting, location, and station environments also contribute to the total uncertainty and inevitably lead to heterogeneous data sets. In order to assess long-term climate trends reliably, systematic errors must be corrected to make different parts of a series comparable, which is termed homogenization. For this process, sufficient background information on the measurements (often termed metadata) is necessary. Mostly, data are homogenized with respect to a time average such as monthly means. Additional efforts are necessary to address changes in climate extremes. With daily climate data, for instance, it is necessary to homogenize the shape of the distribution function, as negative extremes might be affected differently by an inhomogeneity (such as change in location) than positive extremes.
Another type of uncertainty relates to the changing representativeness of the measurement location over time. Meteorological stations erected a century ago at the outskirts of a city are now often in the city center, and, therefore, represent a more local climate that is subject to urban effects. Depending on the application, this bias is undesired and must be accounted for (for example, when calculating long-term global climate trends).
the history of climate observations
Climate observations were performed during the classical epoch. However, instrumental measurements started only in the 17th century with the advent of the Enlightenment. The longest continuous temperature time series is the Central England temperature, which reaches back to 1659. Early meteorological measurements were disseminated through scientific journals and private correspondence, and were studied by the scientists of the Enlightenment.
Around the 1850s, the number and spatial distribution of meteorological stations became such that a global view became possible. Many of the widely-used global climate data sets for the Earth's surface reach back to the second half of the 19th century, comprising surface air temperature, pressure, pre cipitation, and other variables. There are, however, changes within these data sets that affect the quality of the analyses. National meteorological networks, the Brussels Maritime Conference of 1853 and, after 1873, the International (now World) Meteorological Organization established standards for weather observations. The data quality and coverage over both the terrestrial and marine domains have increased.
Climate and weather data near the Earth's surface only provide a very limited view of the large-scale atmospheric circulation. Upper-air measurements started in the late 19th century, but were rather experimental in the beginning. In many countries, operational upper-air networks using aircraft, kites, and pilot balloons (free-flying balloons tracked from the ground to derive wind profiles) were gradually established during the 1900s to 1920s, while radiosonde networks (weather balloons measuring temperature, pressure, and humidity and transmitting the data to a receiver at the ground) began in the late 1930s and 1940s. The quality of atmospheric observations and international coordination was strongly improved during the International Geophysical Year (IGY) 1957-58. The IGY led to the establishment of meteorological stations in Antarctica, improved networks of weather balloon soundings, atmospheric ozone observations, and to measurements of carbon dioxide (CO2) in the atmosphere.
Another important step in the history of climate observations was the onset of space-borne Earth observation in the 1970s. Satellites provide a near-global coverage of numerous climate variables, such as sea ice or snow coverage, cloud cover, and the vertical temperature structure, though radiosonde data still form the backbone of the upper-air network. In particular, satellites provide detailed information on the concentration of trace gases and on the amount and properties of aerosols. At least as important as for surface data, quality remains a fundamental issue for satellite data. In particular, the overlap between different sensors is often too short to obtain reliable transfer functions.
A huge amount of meteorological data is measured each day and is processed to produce meteorological analyses and weather forecasts. Although observations are irregular in time and space (with large gaps), these applications require three-dimensional data sets on a regular space-time grid. Such data sets are produced from all atmospheric observations (from the Earth's surface, ships, aircraft, balloons, and satellites) using a filtering and interpolation scheme combined with a numerical weather prediction model. The procedure is termed data assimilation, or analysis, and essentially consists of short-term forecasts that are corrected according to the incoming observations, providing new initial fields for the next forecasting time step, and so on.
In large reanalysis projects, this procedure is applied to past observations, thus providing global gridded meteorological data sets for the past approximately 50 years. Only a subset of the global meteorological network is needed for climate observations, but with particular requirements concerning quality and long-term stability. Ensuring the existence of suitable networks is the task of the Global Climate Observing System (GCOS), an institution that is co-sponsored by the WMO and other intergovernmental organizations. Climate data are used for climate monitoring, but also for climate research, such as detection and attribution of climate change or the validation of climate models.
SEE ALSO: Climate Models; Climatic Data, Historical Records; International Geophysical Year (IGY); University Corporation for Atmospheric Research; Vostock Core; World Meteorological Organization.
BIBLIOGRApHY. P. Brohan, et al., "Uncertainty Estimates in Regional and Global Observed Temperature Changes: A New Dataset from 1850," Journal of Geophysical Research (v.111, 2006); P.W. Thorne, et al., "Revisiting Radiosonde Upper-Air Temperatures from 1958 to 2002," Journal of Geophysical Research (v.110, 2005).
Stefan Bronnimann ETH Zurich, Switzerland
Was this article helpful?
There's no magic bullet that will make you slim down without trying. No particular diet that lets you eat a big amount of food and drop pounds quickly. No ab-machine or exercise bike that you see at three fifteen in the morning on an infomercial is truly going to make that much difference to you.