The USU Rayleigh Lidar (41.74°N 111.81°W) has been regularly used to measure temperatures in the middle atmosphere from 45 to 90 km. It is well suited for nightly observation; provides excellent vertical temperature resolution; and does not need external calibration. It began operation in August 1993 and a dataset spanning more than ten years has been collected. The analysis here includes 593 nightly temperature profiles from September 1993 through July 2003.
With many sources of variation in the atmosphere, all temperature effects cannot be easily detected. The largest source of temperature variation, and the easiest to measure, is the annual variation. Other effects, such as the semiannual variation, solar cycle radiation, and secular trends are also important but more difficult to detect at every altitude. Our model includes these effects, some of which are significant at some altitudes while others are not. The linear model used in this analysis included variables for the annual and semiannual variations, solar effects, average temperature, and secular trend. The MgII index, averaged over 81 days, was used as a solar proxy instead of F10.7 because it yielded a marginally better fit.