CAJ | 학술논문

基于“动力-统计”预报方法的MEOFIS（精细化气象要素客观预报）平台以相关模式预报结果为基础，结合历史实况资料建立预报模型，实现站点的精细化预报．利用2009～2011年的T639模式产品和渤海湾北部相关观测站的数据积累统计建模，并对2012～2013年海面4个季节的气温和风速进行预报统计，对比分析该平台在海面气温和风速预报中的适用性．经客观检验，1℃误差范围内，海面各季节的气温和风速预报准确率均高于陆上的预报；海面日最高、日最低和逐3h气温预报准确率均超过68％，秋季的日最高气温、逐3 h气温和冬季的日最低气温预报最为理想，准确率分别达86.8％、75.2％和78.9％，春季的气温预报整体不理想；显著性检验结果显示：和T639直接输出的结果相比，MEOFIS在各季节的气温预报中具有明显的订正能力．2 m/s误差范围内，过渡性季节春、秋季的日最大风速预报准确率均超过75．0％，夏季的预报效果较差，但逐3 h风速预报准确率最高，达78.0％，冬季的风速预报效果整体不佳；利用总体平均经验模态分解法（EEMD）对各月逐3 h的海面气温和风速预报误差做滤波处理，结果显示MEOFIS平台对这两要素的预报误差均存在明显的双周震荡波，通过滤波可以提高二者预报的准确率，且气温预报准确率的提高更大．预报偏差和方差小的季节，预报准确率的改善更为理想．
기우“동력-통계”예보방법적MEOFIS（정세화기상요소객관예보）평태이상관모식예보결과위기출，결합역사실황자료건립예보모형，실현참점적정세화예보．이용2009～2011년적T639모식산품화발해만북부상관관측참적수거적루통계건모，병대2012～2013년해면4개계절적기온화풍속진행예보통계，대비분석해평태재해면기온화풍속예보중적괄용성．경객관검험，1℃오차범위내，해면각계절적기온화풍속예보준학솔균고우륙상적예보；해면일최고、일최저화축3h기온예보준학솔균초과68％，추계적일최고기온、축3 h기온화동계적일최저기온예보최위이상，준학솔분별체86.8％、75.2％화78.9％，춘계적기온예보정체불이상；현저성검험결과현시：화T639직접수출적결과상비，MEOFIS재각계절적기온예보중구유명현적정정능력．2 m/s오차범위내，과도성계절춘、추계적일최대풍속예보준학솔균초과75．0％，하계적예보효과교차，단축3 h풍속예보준학솔최고，체78.0％，동계적풍속예보효과정체불가；이용총체평균경험모태분해법（EEMD）대각월축3 h적해면기온화풍속예보오차주려파처리，결과현시MEOFIS평태대저량요소적예보오차균존재명현적쌍주진탕파，통과려파가이제고이자예보적준학솔，차기온예보준학솔적제고경대．예보편차화방차소적계절，예보준학솔적개선경위이상．
The MEOFIS(meteorological element objective forecast integrated system)platform is based on the statis-tical dynamical forecasting method.According to the model prediction and historical observational data,the refined forecasting elements are obtained.The T639 model products and data from buoy stations of Bohai Bay,coastal auto-matic weather stations of the year 2009~20 1 1 are used to build the equations.The temperature and wind speed on the bay in the year 2012~2013 is predicted and used to analyze the applicability of the platform of the northern Bo-hai Bay forecasting.Objective verification of the forecast elements shows that within 1℃error range,the sea surface temperature is forecasted better than that of onshore within 1℃error range.MEOFIS grasps the trend of sea surface temperature better.The accuracy rates of daily maximum,daily minimum and 3h temperature are all over 68%.The forecasting of daily maximum,3 h temperature in autumn and minimum temperature in winter is ideal with accuracy rate of 86.8%、75.2% and 78.9%,respectively.The temperature forecasting in spring is unsatisfactory overall. The significance test shows that compared with the results of T639,the temperature-correction capability of MEOFIS is significant in 4 seasons.Within 1℃ error range,the maximum wind speed accuracy rate in transitional seasons, spring and autumn,is over 75.0%.the worst forecasting occurs in summer though the 3h forecasting accuracy rate is the biggest(78.0%).The wind speed forecasting in winter is unsatisfactory overall.EEMD(Ensemble Empirical Mode Decomposition)is utilized to filter 3h temperature and wind speed forecasting error range.The results show that temperature and wind speed forecasting errors from MEOFIS have remarkable biweekly oscillation waves.Filte-ring can improve the prediction accuracy and has better result for temperature forecast.The accuracy rates for fore-cast seasons of small deviation and variance are improved better.