CFSv2模式对东北地区夏季各月降水的预测性能评估

Prediction performance of CFSv2 model for monthly precipitation in summer in Northeast China

利用距平相关系数(ACC)、距平符号一致率(Pc)、趋势异常综合检验(Ps)等3种方法,对CFSv2模式在东北地区开展的1983-2010年夏季各月降水回报试验结果、2011-2014年业务应用结果进行综合评估。结果表明,CFSv2模式在东北地区夏季各月降水的效果相对较好。尽管目前应用的业务评分办法的评分都相对较高,但年际差异明显,预测效果不稳定,需要进一步提高模式的预测技巧。

Three methods of anomaly correlation coefficient （ACC）, anomaly sign consistency rate （Pc） and trend anomaly comprehensive test （Ps） are used to evaluate the CFSv2 pattern in the northeast region of 1983-2010 for each month of summer precipitation return test results and 2011-2014 business application results. The results show that, The effect of CFSv2 model on monthly precipitation in summer in Northeast China is relatively good. Despite the relatively high scores of the current business scoring methods, the inter-annual variability is signifi- cant and the predictive effect is not stable, so further improvement of the forecasting skill is needed.