基于LAPS分析的WRF模式逐时气温精细化预报释用

Product Interpretation of Refined Hourly Temperature Based on the Assimilation of WRF Model in LAPS

为了满足湖北电力和华中电力系统对逐时气温预报的需求,开展气温精细化预报服务,基于局地分析预报系统(LAPS)分析融合的WRF模式精细化数值预报产品,对2011年4月1日至7月20日湖北省内及周边区域站点的逐时气温进行精细化预报释用。比较几种模式释用方案,结果表明:WRF模式预报趋势较好,预报准确率评分50%左右,日间08-20时预报误差小于夜间,最高气温(T_(max))评分54%,最低气温(T_(mm))评分为44%。通过模式释用,提高了预报准确率,可使预报评分提高到60%～70%,对T_(max)评分达到57%,对T_(mm)评分最高达到74%。分时刻建立MOS方程能够有效降低夜间20-07时的预报误差。将实况最高气温引入Kalman滤波方程能进一步提高预报准确率。

Based on the assimilation of refined WRF model products in Local Analysis and Prediction Sys- tem （LAPS）, in order to meet the needs of the power system temperature forecast, several product inter- pretation forecast methods of hourly temperature were designed in Hubei Province and surrounding sta- tions from 1 April to 20 July 2011. Compared with other several forecast models, results show that the WRF model has a good simulation of temperature trend, the temperature forecast accuracy score is around 50~, the forecast error in the daytime of 08--20 BT is less than that in the nighttimes, the accuracy score of maximum temperature is 54 ~ and the accuracy score of minimum temperature is 44 ~. The forecast ac- curacy can be further improved through product interpretation of model output, so that the accuracy score reached to 60^--70~, that of maximum temperature got 57~ and that of minimum temperature reached 740//oo. In addition, the MOS equation established hourly time can effectively reduce the forecast errors in nighttime of 20--07 BT. Furthermore, bringing maximum temperature factor into Kalman system equa- tions can be further improved forecast accuracy.