河西走廊东部沙尘暴气候特征及短时预报

沙尘暴是河西走廊东部多发的灾害天气之一。利用河西走廊东部5个气象站1961-2014年月沙尘暴资料,运用统计学方法分析了河西走廊东部沙尘暴日数的时空分布特征及变化趋势;选取1990-2009年逐日NCEP再分析资料,依据气流的南北配置对沙尘暴天气进行了环流分型;采用诊断方法、因子组合和日常的经验预报等方法对不同层次、不同物理量进行分析和计算,构建了具有经验性的预报因子库,利用线性相关、经验预报和最大靠近原则等诊断分析方法建立了沙尘暴诊断预报模式。结果表明:受天气系统、地形以及海拔等影响,低海拔地区沙尘暴日数多于高海拔地区。年、年代沙尘暴日数呈显著减少趋势,递减率为民勤 >凉州 >永昌 >古浪 >天祝。年沙尘暴日数的时间序列存在着5~6 a的准周期变化,北部年沙尘暴日数发生了突变,而南部没有发生突变。沙尘暴日数月变化也比较一致,峰值出现在4月,谷值出现在9-10月。 沙尘暴日数均为春季最多,秋季最少。沙尘暴的环流形势分为西北气流型、西南气流型和西风气流型。确定了各型沙尘暴的预报指标和阈值,诊断预报模式的预报准确率在73.3%以上,达到了较高的预报水平,填补了沙尘暴精细化预报的空白,可为沙尘暴的业务预报预警提供客观有效的指导产品。

Climatic Characteristics and Short-term Forecast of Sandstorm in East of Hexi Corridor

Sandstorm is one of the most frequent disaster in eastern of Hexi Corridor. In this paper, using monthly sandstorm data of five meteorological stations in eastern of Hexi Corridor during 1961-2014, temporal and spatial distribution characteristics and change trend of sandstorm days were systematically analyzed by statistical methods. Besides, with daily NCEP reanalysis data with 1°×1° resolution during 1990-2009, the sandstorm weather circulation pattern were formed according to north-south airflow configuration, different levels and different physical quantities were analyzed and calculated by methods of diagnosis, factor combination and daily experience forecast. Further empirical prediction factor database were composed, the diagnostic forecast model of sandstorm weather was set up by diagnostic analysis methods which include linear correlation coefficient, empirical prediction and maximum proximity principle. Results showed that sandstorm days in low altitude area were more than that in high altitude area because of weather system, landform and altitude. Sandstorm days showed obvious reducing trend at year and decade scales, the decrease rate was Minqin > Liangzhou > Yongchang > Gulang > Tianzhu. Annual sandstorm days had 5-6 a quasi-periodic variation, and mutation occurred in north, but no mutation in south. Monthly variation of sandstorm days was consistent for different areas, and the peak occurred in April but the trough was in September and October. Sandstorm days were the most in spring, but the least in autumn. Sandstorm circulation pattern was divided into three categories, including northwest airflow, southwest airflow and westerly airflow. The forecast indexes and thresholds of different sandstorm types were identified, and the forecast accuracy rate of diagnostic model was above 73.3% with a high prediction level.