Impact Of Initial And Sst Forcing On Temporal And Spatial Distribution Of Extended-Range Predictability

Wang Qiu-Liang,Zhang Li-Feng,Guan Ji-Ping
DOI: https://doi.org/10.6038/cjg20131205
2013-01-01
Abstract:Numerous extended-range numerical ensemble forecasting experiments are performed with a global spectral model (T106L19), which are initialized by Breeding Generated Modes, and each experiment has two sets of model integrations forced by climatic SSTs and predicted SSTs, respectively. Based on model results, an analysis of variance approach is utilized to investigate the impact of initial and SST forcing on temporal and spatial distribution of extended-range predictability and relevant mechanisms. Four indexes are defined that are index of initial error influences (II), index of SST influences (EI), potential predictability index (R), and wave activity index (WI). It was shown that the distribution of II has significant regional and seasonal differences, suggesting that the influences of initial error conditions on mid-high latitudes are greater than that on tropics for extended-range forecast. EI has similar distribution to II in same seasons. R exhibits zonal distribution obviously, and larger values are located in tropical regions. Within low-latitude belts, potential predictability is larger at high layers than that at low layers, and it is larger in summer and autumn seasons than that in spring and winter seasons. Results show that the impact of initial error conditions and SSTs forcing on temporal and spatial distribution of extended-range predictability relies on general circulation strongly. Further investigation indicates that the regions of significant II and EI almost overlap the active atmospheric long-wave regions and westerly jets within extended-range. Since strong vertical wind sheer close to westerly jets that result in baroclinic instability and provides plenty of baroclinic instable energy for long wave activities. Consequently, evolution of long-wave activities modulates the influences of initial error and SSTs forcing, and it indicates that extended-range predictability is flow-dependent clearly, and the role of external forcing has close relationship with atmospheric internal dynamical process within the extended range.
What problem does this paper attempt to address?