The impact of atmospheric dynamics on vertical cloud overlap over the Tibetan Plateau

Abstract. The accurate representation of cloud vertical overlap in atmospheric models is particularly significant for predicting the total cloud cover and for the calculations related to the radiative budget in these models. However, it has received too little attention due to the limited observation, especially over the Tibetan Plateau (TP). In this study, 4 years (2007–2010) of data from the CloudSat cloud product and collocated ERA-Interim reanalysis product were analyzed to examine the seasonal and zonal variations of cloud overlap properties over the TP region, and evaluate the effect of atmospheric dynamics on cloud overlap. Unique characteristics of cloud overlap over the TP have been found. The statistical results show that the random overlap assumption slightly underestimates the total cloud coverage for discontinuous cloud layers over the TP, whereas the overlap parameter α for continuous cloud sharply decrease from maximum to random overlap with an increase of layer distance, eventually trending towards a minimal overlap (e.g., negative α values) as the cloud layer separation distance exceeds 1.5 km. Compared with the global averaged cloud overlap characteristics, the proportion of minimal overlap over the TP is significant high (about 41 %). It may be associated with the unique topographical forcing and thermos-dynamical environment of the TP. As a result, we propose a valid scheme for quantifying the degree of cloud overlap over the TP through a linear combination of the maximum and minimum overlap, and further parameterize decorrelation length scale L as a function of wind shear and atmospheric stability. Compared with other parameterizations, the new scheme reduces the bias between predicted and observed cloud covers. These results thus indicate that effects of wind shear and atmospheric stability on cloud overlap should both be taken into account in the parameterization of overlap parameter to improve the simulation of total cloud cover in models.