Urbanization and its Effect on Temperature in Phoenix Arizona

Phoenix, Arizona, is one of the most rapidly growing cities in the Nation, both in terms of area and population. In 1950, Phoenix covered 17 square miles and had a population of 107,000; by 1995, Phoenix covered 456 square miles and had a population of 1,079,900. Such dramatic urbanization has had a significant impact on temperatures recorded in the metropolitan area, especially on minimum temperatures.

An examination of maximum, mininum and average daily temperatures for Phoenix, collected since 1896, reveals that: 1) 18 of the warmest 20 years, expressed in terms of a yearly mean, occurred during the 20 year period 1980-1999; 2) since the mid 1960s, average minimum temperature increased markedly, while average maximum temperature increased much more slowly; 3) fewer record low temperatures were established in the 1990s than in any previous decade; 4) a disproportionately large number of record warm minimum temperatures were established during the 1980s and 1990s; and 5) new maximum temperature records were established for the year, as well as for the months of June and July, during the 1990s.

NOAA's Climate Prediction Center (CPC) monthly and seasonal (three month) temperature anomaly forecasts are generated from 30-year means (as of February 2000, the 1961-1990 period of record was used). During the 1990s, CPC temperature anomaly forecasts for Phoenix have been heavily weighted toward "above climatology". The authors suggest the CPC forecasts are, to a great degree, influenced by the observed upward trend in minimum temperatures at Phoenix during the period 1961-1990, a period of very rapid areal expansion and population increase. To mitigate the effects of urbanization on the climatological record and to provide more meaningful temperature anomaly forecasts for rapidly growing metropolitan areas such as Phoenix, the authors recommend that one, or both, of the following be considered: 1) compare mean maximum temperature, not mean daily temperature; and 2) shorten the period used to generate climatological means.