Tropical cyclones (TCs) simulated at gray-zone resolution (9km) in the Advanced Research version of the Weather Research and Forecasting Model (WRF-ARW) are performed and evaluated over the Western North Pacific (WNP). Two sets of experiments are carried out to test the effect the cumulus parameterization (CPS), driven by the European Center for Medium-Range Weather Forecasts (ECMWF) fifth-generation global atmospheric reanalysis (ERA5), over 11-year (2008-2018) TC seasons (June–November). Through comparisons with the observation and reanalysis, it is shown that the ICPS (involving CPS) experiment yields good skills in simulating TC frequency due to more realistic large-scale mean states, including stronger low-tropospheric circulation, wetter mid-tropospheric environment and more active ascending motion over TC main development region (MDR). On the other hand, the NICPS (not involving CPS) experiment is found to better reproduce intense TCs (Saffir-Simpson hurricane categories 3, 4 and 5) in terms of intensity, accumulated cyclone energy (ACE) and inner-core size. Although NICPS can capture TC inner-core structure well with stronger radial inflow in the boundary layer, rising motion around the eyewall and outflow in the mid to upper troposphere, NICPS significantly underestimates TC outer size due to rapidly decay of TC outer wind field, suggesting limited severe convective activities in the outer region.

High-resolution regional reanalysis is one of the most powerful tool to study local and regional high-impact weather events, climate extremes and climate change impact. In this study, two high-resolution Chinese Regional Reanalysis (CNRR) datasets with a resolution of 18km over 1998-2012, which are produced using the Gridpoint Statistical Interpolation (GSI) data assimilation system and spectral nudging (SN) technique, were assessed. The reliability of the surface and upper air variables of the CNRR datasets was evaluated by comparing with in-situ observations and the European Centre for Medium-Range Forecasts (ECMWF) ERA-Interim (ERAIN) global reanalysis dataset. The results show that the CNRR can provide more accurate near-surface variables than the driving ERAIN global reanalysis. However, special care should be taken when using the CNRR precipitation dataset, especially for heavy rainfall cases. CNRR datasets are also able to generate high-quality upper atmospheric products especially in CNRR-GSI experiment, which assimilates long time series of local observations. By using the three-dimensional variational data assimilation (3D-Var) method, CNRR-GSI outperforms the CNRR-SN and ERAIN. With the increase of the computing resources, the potential opportunities for improving CNRR can be expected by applying more advanced methods.

Based on the object-based tracking algorithm, the precipitation simulation ability of the convection-permitting (CP) regional climate models (RCMs) is evaluated from the viewpoints of the precipitation systems in this work. The characteristics of precipitation systems over eastern China during 1998-2007 derived from the Weather Research and Forecasting model (WRF) with the horizontal grid spacing of ~ 4 km are compared with CMORPH. On the whole, CP RCMs can capture the average duration and eccentricity of all precipitation systems reasonably. However, precipitation systems tend to be stronger but with smaller coverage area in CP RCMs, which leads to the wet biases and dry biases of accumulated precipitation amount in the longer-duration (>= 48 hr) and shorter-duration (< 48 hr) systems, respectively. Such deficiencies in accumulated precipitation amount of precipitation systems with various durations can be made up by employing spectral nudging technique in CP RCMs to a certain degree. This work further indicates that, to improve the capability of precipitation simulation in CP RCMs, the relationship between intensity and coverage area of precipitation systems should be described properly in CP RCMs, especially for those with shorter-duration.