An averaging of data in space or time, designed to compensate for random errors or fluctuations of a scale smaller than that presumed significant to the problem at hand. Thus, for example, a thermometer smooths the temperature reading on the scale of its time constant; the analysis of a sea level weather map smooths the pressure field on a space scale more or less systematically determined by the analyst by taking each pressure as representative not of a point but of an area about the point. See consecutive mean, curve fitting, filtering, bloxam.