Abstract
Different reanalysis data and physical parameterization schemes for the Weather Research and Forecasting (WRF) model are considered in this paper to evaluate their performance in meteorological simulations in the Ili Region. A 72-hour experiment was performed with two domains at the resolution of 27 km with one-way nesting of 9 km. (1) Final Analysis (FNL) and Global Forecast System (GFS) reanalysis data (hereafter, WRF-FNL experiment and WRF-GFS experiment, respectively) were used in the WRF model. For the simulation of accumulated precipitation, both the WRF-FNL (mean bias of 0.79 mm) and WRF-GFS (mean bias of 0.31 mm) simulations can display the main features of the general temporal pattern and geographical distribution of the observed precipitation. For the simulation of the 2-m temperature, the simulation of the WRF-GFS experiment (mean warm bias of 1.81 °C and correlation coefficient of 0.83) was generally better than that of the WRF-FNL experiment (mean cold bias of 1.79 °C and correlation coefficient of 0.27). (2) Thirty-six physical combination schemes were proposed, each with a unique set of physical parameters. Member 33 (with the smallest mean-metric of 0.53) performed best for the precipitation simulation, and member 29 (with the smallest mean-metric of 0.64) performed best for the 2-m temperature simulation. However, member 29 and 33 cannot be distinguished from the other members according to their parameterizations. For this domain, ensemble members that contain the Mellor–Yamada–Janjic (MYJ) boundary layer (PBL) scheme and the Grell–Devenyi (GD) cumulus (CU) scheme are recommended for the precipitation simulation. The Geophysical Fluid Dynamics Laboratory (GFDL) radiation (RA) scheme and the MYJ PBL scheme are recommended for the 2-m temperature simulation.
Highlights
Precise flood and drought forecasts are difficult due to the scarcity of observation data and the complex topography over the Ili Region [1]
The Weather Research and Forecasting (WRF) model is simulated using different reanalysis data, and the results show that the WRF model has significant differences in precipitation and temperature
The results showed that the simulations of the WRF model have some discrepancies from the observation data
Summary
Precise flood and drought forecasts are difficult due to the scarcity of observation data and the complex topography over the Ili Region [1]. No dynamic downscaling model studies have been conducted to record high-resolution climate data in the region. Finding the optimal set of physical parameterization schemes (and selecting appropriate model grid resolutions and LBCs) to simulate rainfall events and understanding the impact of different parameterization schemes on rainfall simulation in the Ili Region would be a valuable study.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
