The paper makes a case for the adoption of common protocols in urban heat island (UHI) research and its use in applied climatology. Strict attention to issues of scale, experimental design, site classification, instrument exposure and metadata could greatly improve the present situation. Too often lack of simple scientific rigour leads to sloppy or erroneous measurement or modeling efforts, and misdirected attempts to ameliorate urban thermal climates.

UHI research and its application has suffered because of complacency arising from several commonly held perceptions. These include the facts that temperature is an easy variable to measure, the warmth of cities relative to their surrounding countryside is a well established and clear example of anthropogenic influence on climate, and potential reasons for its existence and relevance to human activities are easy to comprehend. Such perceptions have attracted an enormous number of investigations from all over the world. Most studies focus on the existence, magnitude, dynamics and form of the UH, either in the atmosphere or at the surface for a particular settlement.

The scientific value of much of this UHI activity is however compromised by lack of rigour in the design and conduct of experimental work. Most critical is explicit recognition of scale because that underpins conceptualization of the study question and sets limits on the placement of instruments, guides interpretation of the data, establishes the time and space scales of any modeling framework and sets the usefulness of the results in practical applications.

A review of the large UHI literature shows that the experimental design, choice of sites, exposure of instruments and lack of sufficient metadata often leave much to be desired. This leaves observational studies sufficiently idiosyncratic that their results cannot be confidently used in comparative work or as data to test model output. The abundance of descriptive studies is not matched by research into UHI processes that are directly relevant to the formulation of UHI models at the different scales. Hence models may be missing relevant processes or ignore significant interactions with other scales. Some models have been tested against data from the wrong type (scale) of UHI. Finally, the practical application of UHI results also faces difficulties because users who are not UHI specialists may misconstrue the applicability of the data to their needs. Again, mismatching the scale of the data with that of the target of interest (e.g. pedestrians, houses, neighbourhood) can lead management agencies to erroneous interpretations and misdirection of funds.

It will be shown that resolution of these problems is surprisingly uncomplicated. It requires those involved in UHI work to truly appreciate the significance of scale in urban climates and to carry that understanding through all aspects of a study uniformly.