Prediction skills of the CAMS-CSM Climate Forecast System on the South China Sea Summer Monsoon establishment

Based on a set of hindcast experiments from 2011 to 2020, the prediction skills of the Chinese Academy of Meteorological Science Climate System Model (CAMS-CSM) climate forecast system on the features of abrupt change, characteristic process around the South China Sea (SCS) Summer Monsoon (SCSSM) establishment and the interannual variations are evaluated by using the ERA5 reanalysis and NOAA OLR data. The predictions are classified into three categories according to the length of lead time, that is, 1-30-, 31-60- and 61-90-day. Skillful prediction on the SCSSM onset pentad (pentads 27–28) are found at three lead times. The prediction skills decrease with the increasing lead times. Further investigation suggests that the transitions of the large-scale circulation around the SCSSM establishment at 200-hPa and 850-hPa can be well captured at three lead times. In the SCS, the transition times of the meridional temperature difference and zonal wind shear at 1-30-day lead time coincide with those of observation. The prediction skills on the characteristic process of the SCSSM onset, including the transition from easterly to westerly, the burst of convection, and the occurrence of rainfall over the SCS, are the best at 1-30-day lead time. The interannual variations of the SCSSM onset can only be reasonably reproduced at 1-30-day lead time. In addition, accurate predictions of the eastward retreat of the West Pacific subtropical high, the establishment of the trough in the Bay of Bengal, and the cyclone in the equatorial western Pacific are critical to the predictability of SCSSM onset.