The urban heat island’s share of observed city warming: a three-term decomposition of raw station records for 553 cities in 37 countries

Author:
Preprint
Publisher website
Google Scholar
Abstract
Cities warm faster than the global mean, and how much of that excess is the city’s own heat island decides how much of it city-level measures can reach. Estimates range from negligible to a third of a country’s warming, because no station study separates the regional background from the urban term, and because global assessments rest on homogenized records that remove the gradual urban trend. Using raw, non-homogenized GHCN-Daily records for 553 cities in 37 countries, this paper decomposes observed warming from a 1951–80 baseline to 2001–20 into global warming, regional amplification measured at rural stations, and the urban excess. The share is reported for six reference windows, and covariates are tested over the same years on two settlement records. The urban heat island supplies about a tenth of observed city warming: two-fifths in North America and a few percent in Europe and Japan. The difference lies not in the cities but in how far the countryside kept pace with them. Population growth explains the share, whereas size, density or built surface at one moment does not, and separating the regional term reconciles global and single-country studies. For city planning and for economic assessments of urban heat, the result separates two quantities. Cooling measures act on the heat island’s level, which is nocturnal and close to a degree in North American cities. Growth management and urban form act on its change, the tenth measured here. The remaining nine-tenths is regional and global climate beyond the reach of city-level measures alone.

This publication has 26 references indexed in Scilit: