资源环境科技发展态势分析平台

The purpose of this paper is to elucidate a potential use of the large samples of seasonal means that hindcasts provide for investigating different aspects of climate variability. This use of hindcasts complements their traditional uses in bias correction, real-time forecast calibration, and prediction skill assessment. For seasonal hindcast data from NCEP CFSv2 we show that a sample size 5208 for each target season is achievable. To demonstrate the utility of the proposed concept, we use this large sample dataset to illustrate how it could be used in documenting spatial variability in various moments of seasonal mean precipitation PDF over the US, and further, quantify nuances in the variations in precipitation PDF at different geographical locations with the amplitude of ENSO SSTs. It is our hope that analysis presented in this paper will accelerate utilization of seasonal hindcast datasets in furthering our understanding of different aspects of climate variability. With the advantage of the large sample size, we demonstrated that the precipitation PDF at the each grid of the CONUS can be represented by gamma distribution for a more concise and effective way to summarize precipitation variability. The availability of the large sample dataset also allowed us to analyze the statistical characteristic of the precipitation responses to the different amplitudes of ENSO SSTs. The results show that for strong warm events, enhancement in precipitation has larger amplitude than decrease in precipitation for cold events in the regions of Southern California and southeastern US. The variation of the precipitation signal over the other sub-regions including the southwestern US, mid-northwest, and mid-east shows more linear relationship with the ENSO SSTs. In response to anomalous ENSO SSTs, although the PDF of December-January-February seasonal mean precipitation anomaly is shifted from its climatological PDF, there is still a large overlap between precipitation PDFs for ENSO and its climatological counterpart. This uncertainty in seasonal mean outcomes of precipitation, therefore, limits the seasonal prediction skill.