Tests of the performance of Analgorithmic scheme of the hourly urban heat island

The central areas of cities are on average warmer than their surrounding countryside - this is the so-called Urban Heat Island (UHI) effect. The genesis of this thermal phenomenon is attributable to the modification of the surface heat balance, which in turn is due to changes in surface and atmospheric properties, accompanying urban development. Current approaches to predicting the magnitude of the UHI at screen-level (within the urban canopy layer) are either complex energy balance models with onerous input requirements, or simple multiple linear regression relations that lack universality. Here we present tests of the simple algorithmic scheme of Oke (Preprints Second AMS Symposium on the Urban Environment, 1998) that takes an intermediate path. It uses empirical algorithms relating the UHI to measures of urban structure, weather, rural surroundings and time of day to estimate the magnitude of the hourly UHI in most weather conditions and all seasons for a city in the temperate latitudes. At present it does not account for the influences of frontal passage, coasts, topographic form or large anthropogenic heat fluxes that also affect temperature differences across the landscape.

Performance of the scheme is tested using three data sets. The first consists of semi-continuous hourly UHI and standard weather station data from an observational programme in Uppsala, Sweden by Taesler in 1976-77. The second is a classic set of vehicle traverse data, also from Uppsala. It was collected by Sundborg in 1948-49 on more than 200 occasions, at all times of day and night, and under a wide range of weather conditions. The third set consists of 3 years (1997-99) of continuous hourly data from Lodz, Poland gathered by Klysik and Fortuniak. Preliminary results of the performance of the scheme are presented.