ANALYSES OF ERRORS IN MEDIUM-TERM NUMERICAL FORECAST PRODUCTS FOR THE SUBTROPICAL HIGH 1998

1INTRODUCTIONItiswellknownthattheaccuracyoftheforecastingofwesternPacificsubtropicalhighcontributesmuchtothesuccessofforecastsofmedium-termprecipitationduringtherainingseasonofeasternChina.Forthepastfewyears,NWPproductsarebecomingindispensablefoundationsforoperationalforecastersmakingmedium-termforecasts,withtherapidmodernizationofmeteorologicalmeansandon-goingimprovementoftheNWPproducts.ModelscapableofproducingNWPproductscannotyieldresultsthatentirelyrepresentthe搕rueatmosphere?butha…

ANALYSES OF ERRORS IN MEDIUM-TERM NUMERICAL FORECAST PRODUCTS FOR THE SUBTROPICAL HIGH 1998

By statistic and dynamic analyses, we have come to the following concl usions: (1) The ECMWF medium-term numerical forecast can forecast medium-term ac tivity of subtropical high, and the accuracy rate of forecast cannot have large improvement by translational corrections. (2) The important cause for the ECMWF medium-term numerical forecast to have errors in 1998 is that the astronomical t ide is not included in the model. (3) Two indexes are found from which it can be judged that ECMWF medium-term numerical forecast will have errors if the astro nomical tide is ignored in the model : ① When the 54.7° line under the moon of the nodical month astronomical singularities coincides with the trough-line of the subtropical jet flow from 50°E to 150°E on the 500 hPa level at 2000 L.T. of the same day, and is approximately vertical (α>60°) with the isotherm, then the day 0 - 2 days after the appearance of the nodical month astronomical singu larities is defined as the initial day. Then in three successive days after the initial day, ECMWF medium-term numerical forecast of the northern latitude of th e 588 line at 120 °E will have continuous errors as large as two latitudes (7/9 ). Otherwise, it won't have continuous errors (13/13). ② Otherwise, if the 54.7 ° line is in the range of a low pressure between two high pressures, then ther e is a dispersive error on the day of the nodical month astronomical singulariti es (5/7). There is not any error (6/6) otherwise.