Comparisons of monthly mean nighttime temperature profiles observed by the sodium lidar at Colorado State University and TIMED/SABER overpasses are made. In the altitude range from 85 km to about 100 km, the two observations are in very good agreement. Though within each other's error bars, important differences occur below 85 km in the entire year and above 100 km in the summer season. Possible reasons for these difference are high photon noise below 85 km in lidar observations and less than accurate assumptions in the concentration of important chemical species like oxygen (and its quenching rate) in the SABER retrieval above 100 km. However, the two techniques both show the two-level mesopause thermal structure, with the times of change from one level to the other in excellent agreement. Comparison indicates that the high-level (winter) mesopause altitudes are also in excellent agreement between the two observations, though some difference (2--3 km) may exist in the low-level (summer) mesopause altitudes between ground-based and satellite-borne data. In addition, the difference in local time dependency between lidar and SABER is investigated; this difference makes the comparison of mesopause altitudes difficult.