ABSTRACT In this study, two planetary boundary layer (PBL) schemes in the National Center for Atmospheric Research Community Atmosphere Model version 5 (NCAR‐CAM5) have been evaluated over India and its adjacent regions during winter (DJF), pre‐monsoon (MAM), monsoon (JJA), and post‐monsoon (SON) seasons. Most of the PBL schemes are designed to compute vertical mixing considering two sources of turbulence, namely buoyancy and shear. Here, the two PBL schemes, namely the HB (Holtslag and Boville, 1993) and UW (University of Washington scheme; Bretherton and Park, 2009) schemes, are considered for the evaluation. One is based on the non‐local approach, while the other employs a 1st‐order local TKE (turbulent kinetic energy) closure approach, respectively. The evaluation reveals that the UW PBL is preferable for precipitation throughout the winter and monsoon seasons. In contrast, HB PBL demonstrates better performance during the pre‐monsoon and post‐monsoon seasons over the whole study domain (20S‐40N, 40E‐140E). For 2‐m temperature, UW PBL is beneficial in NCAR‐CAM5 during all four seasons. In the case of surface and upper air wind patterns, both schemes are comparable with respect to observed monsoon flow and location of jets, with a slight difference in magnitude. Notably, over the Indian land and associated different climatic zones, significant differences have been observed between simulated variables utilising the two schemes. The performance is strongly dependent on the considered season and region, and a scheme should judiciously be chosen for the simulation. Furthermore, the sensitivity of the HB PBL scheme to the critical bulk Richardson number (Ri cr ) suggests that the simulated variables are highly sensitive to the utilised value of Ri cr . The mean planetary boundary layer height and total precipitation averaged over the whole domain increase linearly with increasing Ri cr during all four seasons. The surface sensible (latent) heat flux decreases (increases) as Ri cr increases. Thus, an appropriate value of Ri cr in a PBL scheme could enhance the performance to some extent.
Namdev et al. (Sun,) studied this question.