On The Variability Of The Western Pacific Warm Pool And Its Mechanism

The Western Pacific Warm Pool (WPWP) is the primary mode of sea surfacetemperature (SST) in the tropical Pacific Ocean and plays an essential role in regionaland global climate system. The WPWP variability and its mechanism have attractedtremendous attention in the oceanographic and meteorological community as one ofhot topics in the ocean/climate studies. In the present paper, variations of the WPWPintensity and displacement are investigated on intraseasonal, interannual, interdecadalto centurial time scales, and the physical mechanisms of the multi-scale WPWPvariabilities are also explored. A dozen newly available observational datasets (orreanalysis) including NOAA AVHRR SST, WHOI OAFlux and JAMSTEC MOAAGPV combing model outputs (e.g., IPCC AR5) and assimilations (say SODA) areemployed in the present study. The main conclusions are as follows.1. It is discovered that the WPWP area has strong intraseasonal oscillation(comparable to its interannual variability), of which the interannual modulationand long-term decrease trend, and their corresponding mechanisms are alsoexamined.1). It is revealed that the WPWP size (area) possesses drastic intraseasonaloscillation (ISO). The amplitude of WPWP ISO is1.71012m2, which is close tothat of interannual variation (1.81012m2). However, there is no regular periodpertaining to WPWP ISO, except a quasi-period around60days. Further study showsthat the Madden-Julian Oscillation (MJO) is the source of the WPWP ISO. On the onehand, the MJO induces the ISO of convection and sea surface heat flux in the WPWP,especially in the region of warm pool Northern/Southern Boundary (NB/SB), leadingto the intraseasonal meridional movement of NB/SB, and hence the intraseasonalvariation of WPWP size. On the other hand, the MJO sets off the ISO of advection inthe central-eastern Pacific Ocean by producing eastward propagating Kelvin waves,triggering the intraseasonal east-west migration of the Eastern Boundary (EB) and the ISO of WPWP size. It is also shown that the SST ISO in the EB region is determinedby the MJO induced intraseasonal variation of advection and sea surface heat flux,whereas in the NB/SB region, the intraseasonal sea surface heat flux is dominant.Since previous studies focus on seasonal to interdecadal time scales, this paperproposes another view on the WPWP variability in terms of its size.2). MJO-induced interannual modulation of the WPWP ISO is exposed. It issuggested that the WPWP ISO itself displays noteworthy interannual variation.Specifically,1982,1992,1995,1996,1997,2002,2009and2010are strong WPWPISO years, while1991,1993,1994,1998,2000,2001and2004are weak years.Composite analysis suggests that the interannual variability of the MJO is themodulating factor for the interannual variability of WPWP ISO. In detail, the evidentinterannual variation of the MJO drives the interannual change of the sea surface heatflux ISO surrounding the WPWP, and hence generates the modulation of WPWP areaISO on an interannual time scale.3). Significant descending trend of WPWP ISO and its mechanism. Between1982and2011, the WPWP ISO (amplitude) has a remarkable long-term descendingtrend. During the30years, the WPWP ISO index was decreased by0.71012m2(about41%of its average), i.e., by1.4%yr-1. Diagnostic analysis shows that the MJOindex in the Pacific Ocean was decreasing during that period. As a result, both the seasurface heat flux ISO and advection ISO are suppressed, leading to the weakened SSTISO and WPWP ISO.2. WPWP Split and its physical mechanism.1). A phenomenon of WPWP split is pointed out for the first time. Theevolution of the WPWP during1982–2011is investigated (every20days) and it isrevealed that the WPWP is divided into two parts by cool water in certain periods, i.e.,WPWP split takes place. In general, the split zone occurs within0–16°N and120–180°E. The split WPWP has two cores: one is located between15°N and20°N,and another is around5°S.2). Physical mechanism: cyclonic circulation anomaly (enhanced upwelling) is the main cause of WPWP split. Statistical results indicate that the WPWP splitevents usually comes during the developing phases of El Ni o events(July–September) with few exceptions. It is concluded that ENSO-related oceancirculation anomaly accounts for the WPWP split. During El Ni o developing phasesboth the North Equatorial Current (NEC) and North Equatorial Countercurrent(NECC) are strengthened. Consequently, a cyclonic circulation anomaly andanomalous upwelling emerges in the region mentioned above. Then SST is decreasedover there, and hence the split of WPWP is formed.3. Interannual-Interdecadal variability and long-term trend of the WPWPand the corresponding mechanisms.1). In the past the interannual variability of WPWP was almost always related toENSO. But the question is: are the interannual variations of the WPWP all relatedto ENSO? It is suggested in the present paper that yes the zonal displacement ofWPW is closely related to ENSO, but the interannual variability of the WPWPintensity (say heat content, OHC) is not. A new variable named WPWP heat center(WPHC) is proposed and defined to describe the3-D characteristics of WPWP.WPHC is indeed closely related to ENSO. However, the WPWP OHC is quitedifferent. It is found from comparison between the WPWP OHC with Ni o index thatthey are inconsistent (not corresponding), even out-of-phase in many cases,suggesting ENSO is not the only reason for the interannual variability of WPWPintensity. Based on composite analysis, it is concluded that meridional wind fieldconvergence/divergence in the equatorial Pacific Ocean inter-annually, can induce thesea water and heat convergence/divergence in the WPWP, and ultimately can alsocause interannual variability of the WPWP intensity. Therefore, not all the interannualvariation of WPWP is attributed to ENSO.2). A new method named “Running Trend Method”(RTM) is designed toestimate long-term trend of a time series. RTM is utilized to calculate the trendof the WPWP in the past130years. The long-term trends of the WPWP assessed inthe previous studies are usually disturbed by sizeable interdecadal variation. RTM caneffectively remove the influence of short-time oscillation (e.g., interdecadal time scale) in figuring out a long-term trend. The trend of two-dimension WPWP HC and surfacearea during1874–2005is computed using RTM. It is found that the WPWP area hasbeen increasing with a speed of3.21010m2yr1and the WPWP HC has beenmoving eastward with a speed of0.007yr1in the132years. Particularly, thevariation trend of WPWP area during1955–2003recalculated via RTM is about111010m2yr1, which eliminates about30%relative error induced by theinterdecadal oscillation in the results reported by Cravatte et al.(2009).In addition, the Empirical Mode Decomposition (EMD)-Hilbert-Huang Transform(HHT) is employed to examine the intrinsic mode of interdecadal variability of the2-dimentional WPWP HC and surface area. From1870to2009, the interdecadalvariability of the WPWP HC longitude is dominated by the period of30–40years,with relatively weak variability before1950s. The WPWP area possesses interdecadalperiods of20–30years and has lower frequencies during1870–1890.