An experiment study of cloud cover detection by ground based sky thermal infrared brightness temperature measurements.

Measurements of sky thermal infrared brightness temperature (TB) were conducted during the period of April to August in 2007 at the Beijing Meteorological Observatory, using an automatic scanning infrared thermometer （SIRIS 1） developed by the Key Laboratory of Middle Atmosphere and Global Environment Observation (LAGEO), the Institute of Atmospheric Physics, Chinese Academy of Sciences. Cloud cover in the sky was calculated by using the data of TB and the real time surface weather data provided by the Beijing Meteorological Observatory (No.54511). The comparison was made between the two kinds of cloud cover, one was observed by observer in the Beijing Meteorological Observatory, and another was the averaged cloud cover observed by SIRIS1 over the half hour before each round hour. The results show that (1) for middle and low cloud situations, if the cloud in the sky does not change very fast during the half hour before round hour, the two kinds of cloud cover are very consistent with each other; if the cloud changes rapidly, the difference is great. The reason is because the method and the time period of observation are different. (2) For the sky situation dominated by cirrus, the difference between the two kinds of cloud cover is great due to the limitation of infrared sensor (-50℃) used for measurements and the poor ability of threshold determination to distinguish thin cirrus from clear sky at present, especially for thin cirrus, very high in the sky. (3) For complex cloud situations with poor visibility (e. g. cloud cover 10- as is recorded by meteorological observer), the difference of the two kinds of cloud cover is also great. The difference was caused either by the characteristic of instrument, or by observer's ability of visual observation and experience. As far as the instrument is concerned, the lowest measurement temperature (-50℃) limits the ability of high cloud sensing, and the uncertainty of the cloud retrieval is also a key factor. In addition, the poor horizontal visibility may restrict observer's ability of visual observation. The cloud cover measurements by the SIRIS1 demonstrates the distinct superiority of high temporal resolution and allsky automatization but the ability of infrared sensor currently used to sense high cirrus is insufficient.