Statistical depictions of climate

Climate Averages for Barrow, Northern Alaska

Climate conditions have been adequately measured and recorded for many years, making it possible to define what is “normal” and what is an “extreme” event for any part of the world. The image to the right shows the “normal” climate for Barrow, Alaska.

Data gathered over the 30-year period from 1961 to 1990 are currently used to define the latest global “Normals” used for climate reference. At the end of 2020, the 1991 – 2020 will be used as the next reference time frame replacing the 1961-1990 values.

At any given time of the year, an extreme high temperature might be defined as one that occurs only once in every 30 years. A cold winter or hot summer can be specified in a similar way, or in the number of days below or above defined exceptional values. This means that when there is a succession of extremes, or more extreme events occur over a period (e.g. season), it is possible to estimate whether they seem to be part of the normal expectation for the locality, or are so unlikely that they can only be explained in terms of some more radical shift in the climate. The basic properties of any data series, for example temperature, can be defined in terms of the mean over time and the amount of variance about the mean.

Other meteorological variables have more complicated statistical properties. For instance
rainfall. Rainfall is episodic. In many parts of the world, much of the annual rainfall falls in a short rainy season. In addition, most of that rain may be concentrated in a few heavy falls and small shifts in the large-scale weather patterns from year to year may significantly alter the amount and distribution of seasonal rainfall. More complex techniques are usually needed to interpret variations in rainfall.

[Technical information] on the 1961-1990 “normals” can be found on the World Climate Data and Monitoring Programme (WCDMP) website.

Problems with climate statistics

Handling the statistics of rare events is of particular importance to society because we are so vulnerable to extreme weather. There is, for example, a special need to have reliable measurements over as long a time as possible, in order to build up a reliable set of statistics for what are inherently rare events. In order to make sure that the statistics and records are as accurate and reliable as possible several factors need to be taken into account.

One complicating factor can arise from changes at the site where the measurements were made and in the instruments used.  For example, a well-known site problem results from the expansion of urban areas around an observatory. Hence, measurements at such an observatory can be influenced by the so called the Heat Island Effect which is illustrated in the image below. Typically weather stations are established outside the town or city, however as a city grows it can end up surrounding the station.  Urbanization changes many characteristics of the local climate, notably the replacement of cooling trees with concrete and asphalt that heat up during the day but cool only slowly at night.  Buildings and structures also change the ground-level wind flow and create eddies. Urbanization can also lead to rapid rainfall runoff and increases in flash flooding.

These changes can also occur to weather stations that are still in rural locations and are often harder to detect. For instance, the growth of trees around a farmstead that maintains a weather station alters the local wind flow and temperature patterns, and so reduces extreme wind speeds and the incidence of frosts (where they occur). The trend in the observations reflects the changes in the microclimate of the farmstead while the general climate may not have changed.

Heat Island effect (Image: NASA )

Greater challenges arise when many different types of measurements are used for the same event. This is particularly the case for destructive storms. Earlier records depended on shore-based and shipping observations, plus an increasing number of aircraft measurements from the 1940s, all of which relied on different types of equipment and provided only partial mapping of storms around the world. It was only with the advent of weather satellites that a reasonably complete and consistent record of tropical storms could be maintained.

All of this means that any long climatic series must be subjected to close scrutiny to ensure that what appears to be significant changes is a real part of the larger climate and not due to changes in the equipment, observing practices or the site itself.

The history of the study of climate variability and change has been principally a matter of painstaking and scholarly detective work to establish the reliability and applicability of the data.

[Technical information] on how climate statistics are collected and maintained can be found in the 3rd Edition of the Guide to Climatological Practices

[More in depth information] on past observations and how the climate has changed, the IPCC working group I, has a chapter on the topic: Historical Overview of Climate Change Science