The performance of Florida State University atmospheric boundary layer model (FSU1DPBL) in predicting the daily maximum temperature drop is investigated with respect to sensible and latent heat fluxes, Turbulent exchanges in the surface boundary layer are taken into account in the model. Nocturnal atmospheric boundary layer heights obtained from the FSU1DPBL model are compared with those obtained from other model formulations, and stability conditions of the ABL are discussed to derive information concerning the prediction of minimum temperature, It is explained that a growth in the ABL at night can occur under cloudy and/or windy conditions in the model, causing a bias as high as 5 degrees C in predicting the daily minimum temperature. The usual similarity theory for the stable boundary layer leads to a significant overestimation of surface cooling in the model. From the model perspective, the success of the 12-h model forecast in predicting minimum temperature compared to that of the 24-h and 36-h forecasts is explained. The results show the prediction of nocturnal ABL height under the extremely stable and moderately stable cases by the model yields acceptable values in comparison with other model formulations. The ABL height from the model can be replaced by an interpolation formula to yield more realistic values under very stable conditions at night. (C) 1999 Elsevier Science B.V. All rights reserved.