Radiosonde errors and spurious surface-based ducts

Substantially more frequent occurrences of surface-based super-refractive and ducting layers have been noted in daytime than in night-time refractive climatological data derived from standard United States radiosondes. Evidence is presented which indicates that the diurnal bias in these data is largely fictitious. Although the problem is fundamentally a consequence of vulnerability of the carbon humidity sensor to temperature effects, several factors contributed to this bias. These factors are discussed, and steps suggested which could improve the quality of future measurements for refractive applications.