Observations from the University of Alabama in Huntsville campus and groundbased scanning radar for over 140 total spring, summer, and autumn cases are studied to contribute to the relative scarcity of long-term datasets documenting the afternoon-to-evening transition of the planetary boundary layer. A sunset relative frame of reference is employed, focusing on the period 3 h before to 2 h after astronomical sunset, and several findings are consistent with previous investigations. Fluctuating components of wind and temperature computed from nearly collocated surface, Doppler wind profiler, and vertically pointing Doppler lidar measurements show a consistent decline as turbulence intensity diminishes through the transition. When normalized by their initial values, a pattern emerges: temperature variances decline slowly at first then quite abruptly after about 90 min before sunset. After the temperature variances begin to wane, vertical velocity fluctuations decrease, and the rate of their decay increases as vigorous thermal structures diminish. The fastest decline of the horizontal wind variance occurs after an accelerated vertical wind variance decrease, and the horizontal wind fluctuations display the slowest rate of decrease among these quantities. Near-surface humidity measurements show increases in mean water vapour mixing ratio as a steady rise generally beginning about 80 min prior to sunset. Composites of mean lidar vertical motion show final convective-type towers of upward motion occur about an hour before sunset and are coherent through 800 m (all heights a.g.l.). Lidar vertical motion variance at 195 m decreases by more than an order of magnitude approaching sunset, then remains below 0.01 m 2 s −2 for the rest of the studied time frame. Subtle, but steady, increases in both horizontal wind speed and radar-derived horizontal wind convergence above the surface layer (at 300 m) span the entire 5-h time frame. While the convergence results show a broad range, an increase in the mean is clear and found to be statistically significant. Implications