Abstract
This paper presents an evaluation study of the Regional Climate Model version 4.3 ( RegCM4 .3) over the Loess Plateau in northern China, which is a semi‐arid region characterized by complex topography. During recent years, a series of reforestation programmes have been implemented across the region that might influence the local climate in complex ways. To better understand the local climate, the RegCM4 .3 was applied to simulate the present‐day conditions over the Loess Plateau. The simulation was carried out from 1990 to 2009 at the 50‐km horizontal resolution, with lateral boundary conditions taken from the ECMWF ‐Interim reanalysis. A series of climate variables and processes were evaluated during the winter and summer seasons, such as 2‐m air temperature, precipitation, wind circulation, surface energy balance, full moisture budget, and cloud radiative forcing (CRF). The possible origins of the simulation bias and the physical linkages with other model processes were examined. In general, RegCM4 .3 is able to reproduce both the spatial and temporal features of the regional climate over the Loess Plateau. However, there are still biases in some meteorological variables including precipitation and 2‐m air temperature. In particular, the model tends to produce cold biases during winter and underestimate precipitation during summer. Further analyses indicates that the cold biases in winter may have resulted from the deficiency of the downward longwave radiation fluxes, excessive ground heat fluxes, and negative temperature advection by the seasonal mean circulation. These processes are primarily triggered by deficiencies in CRF and excessive northwesterlies over the plateau. The underestimated precipitation during summer is associated with a weak southerly monsoon in the model. A full moisture budget analysis reveals that the dry bias in this region can be mainly attributed to model deficiency in moisture advection and convergence, and to a lesser extent to that in surface evaporation.