Impact Of Heat Content Over Western Pacific Warm Pool On Spring Persistent Rain And Its Mechanism

By using Scripps and GODAS(Global Ocean Data Assimilation System) monthly mean sea temperature, NOAA ERSST V3b,China753stations precipitation and NCEP/NCAR daily reanalysis data, investigation on precipitation and onset date of SPR(spring persistent rain) and their general features are performed. Respective correlations between heat content over Pacific warm pool with SPR intensity and onset time have been discovered, their possible processes and mechanisms are interpreted with anomalous circulation in upper and low level. BP neural network predicting model has been constructed with HC (heat content) in preceding seasons as the fundamentally predictable factor. Main conclusions can be summarized as follows:(1) Major modes of SPR spatial distribution are dry-wet consistency, meridional-opposite phase and zonal-opposite phase. Evident features after SPR onset are the abrupt increasement in upstream southwest flow and connection in powerful southwest moisture channel with eastly moisture channel, through which, moisture over warm pool are transported to Jiang Nan via South China Sea and Indochina Peninsula. Meanwhile, intense divergence center is formed over Jiang Nan with vigorous convective activity. Anti-WP pattern in middle and low troposphere, anomalous southwest wind in middle and low latitudes, anomalous meridional circulation in Jiang Nan-Warm Pool, and anomalous divergence over Jiang Nan in upper level are vital circulation features in rainy years. Circulations in drought years are basically the opposite. Advanced establishment of SPR is promoted by low-level steering vapor transport of westward flow through South China Sea and Indochina Peninsula as well as the high-level anomalous divergence center in early rainy years, however, anomalous convergence in high level barely lags the SPR onset time.(2) SPR and HC are more correlated compared to SST (sea surface temperature) over Western Pacific Warm Pool. It is suggested then, that HC over Western Pacific Warm Pool be selected as the superior factor predicting SPR when taking the centrality, stability, significance and predictability of the key regions into consideration comprehensively. Spring circulation stimulated by HC (opposite sign) in preceding-summer over warm pool are consistent with anomalous field of rainy years in spring.(3) HC over key region (150°E～166°E,9°N～16°N) in preceding-summer was anti-correlated with the intensity of SPR. Significantly low HC in preceding-summer strongly indicates surplus rainfall in spring. Anomalous low HC over key region in preceding-summer induces low-level anomalous anti-cyclone in northwest, which persists to current spring from preceding-summer. In spring, driven by the anomalous anti-cyclone, powerful southwest flow carries abundant moisture to Jiang Nan. Meanwhile, anomalous cyclone of upper level over mainland of China is stimulated and lasts from preceding-autumn to current spring. The anomalous cyclone leads to the blocked northward of South Asian High, southward replacement of westerly jet axis and formation of the abnormal divergence center with intense pumping effect accompanied by "bell-mouthed flow"in the upper level, hence resulting in surplus rainfall in spring. Abnormally circulation may attribute to continuous convective anomaly stimulated by heat content. Anomalous latent heat contributes to the consistency and stabilization of circulation in low and upper level. Abnormally low heat content in preceding-summer stimulates low-level anti-cyclonic anomaly, which steers and gathers water vapor to Jiang Nan and thus promotes the advanced onset of SPR. Meanwhile anomalous cyclone has not retreated westward, and pumping effect has hardly been formed. Vapor transport in low level is the main cause of SPR onset in advance. Process of abnormally less rainfall in spring and delayed onset caused by anomalous high HC in pre-summer is on the contrary.(4)A SPR forecast model is established in terms of BP neural network using HC as fundamental input factor. Results of the operational experiments show superior fitting accuracy compared to multiple linear regression equation.lt is equitable and feasible to forecast SPR using HC as fundamental predictable factor based on BP neural network.