The major concern of this study was to relate spatial and temporal variations of atmospheric infra-red radiation (IR↓) in Christchurch to peculiar attributes of the urban atmospheric environment, and to attempt to determine the relative role of IR↓ in contributing to the urban heat island. This was approached in several ways including energy balance simulation of urban surface temperature to assess the likely effects of IR↓ variations on the heat island, urban-rural observations of radiation and other variables to establish variations in radiation transfer, and measurements of temperature, humidity and pollution in the vertical dimension to allow theoretical calculation of IR↓ and assessment of pollution effects on emissivity. Data collection was undertaken during 1978 and 1979 under clear-sky conditions to reduce the complicating effects of cloud cover. The energy balance model, validated against field observations and making no direct allowance for IR↓ variation across the city, predicted Christchurch's nocturnal heat island with a reasonable degree of success indicating that the phenomenon was related more to surface effects of specific landuses rather than any atmospheric radiative effect. Urban-rural differences in incoming radiation showed an average clear-sky depletion of total solar radiation (SW↓) at the urban site of 3.5%, with individual daily depletions ranging above 10% and clearly related to atmospheric pollution. The urban SW↓ deficit appeared to be more than compensated for by an average clear-sky urban excess IR↓ of 5.3% over the study period, and substantially more during winter and at certain times of day. While daytime excess IR↓ at the urban site was almost exactly balanced by a SW↓ deficit, the nocturnal continuation of urban IR↓ excess resulted in an excess radiant energy receipt at the urban site at that time. However. overall urban radiant energy excesses did not appear sufficient to account for the observed urban temperature excess. Detailed investigations of the nature and composition of the urban atmosphere initiated to identify the source of observed urban IR↓ excess showed that the urban temperature excess present in the horizontal was also evident in the vertical, where warmer temperatures were consistently observed to approximately 300 m. The phenomenon was most marked at night and under a low windspeed regime. Atmospheric moisture regimes at the representative sites also showed distinctive urban and rural characteristics, with a nocturnal urban moisture excess extending approximately 350 m into the boundary-layer and a daytime urban moisture deficit existing to about 200 m. Significant urban-rural differences in air quality were also observed, with measurements emphasising the low level high concentration characteristics of Christchurch urban pollution, particularly during winter. An emissivity approximation model, making no allowance for pollutant emissivity, showed that only a small proportion of observed urban-rural IR↓ difference in Christchurch could be accounted for by variations in temperature and water vapour across the city. Analysis of data for individual sites revealed a substantial component of atmospheric emissivity unaccounted for by profile characteristics, and this excess emissivity was observed higher at the urban site and during daytime. Detailed theoretical consideration of possible causes of the excess emissivity at the urban site indicated the probable importance of particulate emission, and to a lesser degree emission by gaseous pollutants. Accordingly a mechanism to explain the urban-rural divergence in atmospheric infra-red radiation over Christchurch was proposed.