Multi-scale Features And Predictability Of South China Sea Summer Monsoon

With The National Centers for Environmental Prediction (NCEP) Climate Forecast System Reanalysis (CFSR) data, the South China Sea summer monsoon (SCSSM)onset index Uscs defined by Wang et al(.2004)is selected to define the onset date of individual years during 1979-2009. The convection differences and contribution of intraseasonal Oscillation of SCSSM in early and late onset years have been examined. Under the different combinations of the El Ni?o/Southern Oscillation (ENSO) warm and cold phase with the Pacific Decadal Oscillation (PDO) positive and negative phase, the possible impact of the Pacific SST decadal and interannual variability on SCSSM onset time and the possible physical mechanism have been analysed. At last, define the predictive indices of SCSSM onset, making the predict of 2010 SCSSM onset to test its effectiveness.In early onset years, a couple of cyclones appear on both sides of the equator over eastern tropical Indian Ocean before the SCSSM onset, while the active convection occurs over tropical western Pacific. When westerly associated with anomalous cyclone in Bay of Bengal connects the westerly associated with cyclone east of the Philippines, the southwest enhanced over South China Sea, hence the SCSSM onset. In late onset years, the two cyclones absent over the ETIO, while the easterly originates from equatorial western Pacific extend to Indian Ocean which means the weakened monsoon circulation. A large anomalous anticyclone occupies the western north Pacific, which will enhance the intensity of the subtropical high in early summer. As a result, the SCSSM will later onset.The intraseasonal variability of the SCSSM has two preferred time scales (30-60 days and Quasi-Biweekly Oscillation). The 30-60-day mode in early onset years, which originates from convection over equatorial Western Pacific, exhibits an westward propagation, while in late onset years the 30-60-day mode originates from convection over equatorial Indian Ocean exhibits an eastward propagation. In early onset years, the 30-60-day mode convection and a couple of cyclones appear on both sides of the equator over central and eastern tropical Indian Ocean before the SCSSM onset. Then the couple of cyclones propagate poleward, while the active convection propagates eastward to Maritime Continent and tropical western Pacific. When SCSSM onset, the convection evolves into a large northwest-southeast belt from Bay of Bengal and Indochina to western equatorial Pacific. In late onset years, beside couple of cyclones appear on both sides of the equator over eastern tropical Indian Ocean, the 30-60-day mode convecton appears in western tropical Indian Ocean and extend eastward and northward. So, convection evolves into a large northwest -southeast rain belt from Arabian Sea, Bay of Bengal and Indochina to western equatorial Pacific when SCSSM onset. Contrast to the 30-60-day mode, the Quasi-Biweekly Oscillation mode convection over equatorial region exhibits an westward propagation in both early and late onset years. In early onset years, cyclonic vortex across the equatorial eastern Indian Ocean, and turns to a couple of anticyclones when the monsoon onset, which is similar with the convective heating of the Rossby waves. In late onset years, the Quasi-Biweekly Oscillation convection show a north-south propagation in South China Sea. The convection from China South propagate southward and the convection from equator, both converge to the South China Sea, which may trigger the onset of SCSSM.Interdecadal variation and interannual variation in the onset times of SCSSM are examined by analyzing their relationships with Nino3.4 index and PDO index. The results show that the ENSO (Nino3.4 index) could couple with the PDO to affect the SCSSM onset. It is found that when the Nino3.4 index and PDO index are in phase, i.e. high Nino3.4 index/ high PDO phase, or low Nino3.4 index/ low PDO phase, the SCSSM tends to later or earlier onset than normal respectively more often, and hence the SCSSM is more predictable. However an out of phase Nino3.4 index-PDO index relationship gives a less SCSSM predictability. A reasonable mechanism is as follow. In an in phase scenario, the sea surface temperature anomalies (SSTA) associated with the Nino3.4 index and PDO index over the equatorial central and eastern Pacific and North Pacific are of the same sign. The SSTA then modifies the intensity of the subtropical high over western North Pacific. An enhanced subtropical high would then extend westward into the South China Sea while a weakened one would not, which then affect the SCSSM onset day. On the other hand, for an out of phase scenario, whether the SSTA over western North Pacific is positive or negative depends on whether the Nino3.4 index or the PDO index dominates. Furthermore, in such a scenario, the associated circulation over western North Pacific will be canceled, which results in a low predictability of the strength of the subtropical high and hence the SCSSM onset.According to the correlation between the SCSSM onset time and the SSTA, two SSTA Indies have been defined. The area mean SSTA over tropical western Pacific (0o-20oN,125o-155oE) has been defined as WP SSTA index, while the area mean SSTA over equatorial central Pacific (0o-20oN,125o-155oE) has been defined as the ECP SSTA index. Define the objective standard (1) WP SSTA index and ECP SSTA index must keep opposite phase from Sep.(1) to Mar.(0), and no transformation happened to WP SSTA index; (2)the absolute value of the average monthly (from Sep.(1) to Mar.(0)) WP SSTA index must greater than or equal to 0.2; (3) when satisfy (1) and (2) conditions, the SCSSM may onset later when WP SSTA index and ECP SSTA index distribute in the second quadrant, the SCSSM may onset early when they distribute in fourth quadrant; (4) when two indices is located in the first and third quadrant the SCSSM may onset normally, satisfying (1) but not (2) the SCSSM may onset more normally. The two indices have been used to predict 2010 and 2011 SCSSM onset time reasonably.