摘要
面雨量是流域防汛抗洪调度的重要依据,单模式的确定性预报无法反映面雨量的多种可能性,为了解决面雨量预报存在不确定性的问题,本文基于中国气象局区域集合预报系统(China Meteorological Administration-Regional Ensemble Prediction System, CMA-REPS)降水结果对海河流域集合预报面雨量开展了适用性评估和分析,结合高分辨率格点实况资料对其2020—2022年5—8月面雨量进行检验,并依检验结果开展了集合预报面雨量产品及跨度预报产品研发。检验表明CMA-REPS集合预报对海河流域面雨量预报有改进:1)海河流域集合平均面雨量预报的绝对误差明显优于控制预报,在空间分布上表现为南部河系预报误差较大,中部次之,北部最小。2)模糊评分表明小雨和中雨量级集合平均预报更贴近实况,而暴雨量级则需进一步参考集合成员的极端性。3)TS及Bias评分表明集合平均在小雨量级预报效果较好,中雨及以上量级在预报后期改进更加明显。4)概率预报评分发现CMA-REPS对海河流域中部河系中雨以下面雨量有较好的表现,而南部河系的评分则低于中部河系。开发集合预报面雨量产品并对海河流域两次致洪暴雨过程检验发现:1)集合成员对24 h大暴雨等级面雨量的预报有较高的概率,与实况量级相当,对极端降水有较好的提示。2)24 h内集合平均具有较好的评分,而24~48 h, 75%分位数产品对于强降水过程预报效果更好;本文依据检验结果开发的集合预报混合百分位产品及跨度预报产品对于强降水有较好的参考性。3)依据概率预报随时间的变化曲线对面雨量进行平滑拟合,可以得到更贴近于实况的逐小时面雨量预报产品。
Area rainfall plays a crucial role in flood prevention within river basins.However,deterministic forecasts from single models often fail to capture the full range of possibilities,leading to uncertainties in area rainfall predictions.To address this issue,we employ the China Meteorological Administration-Regional Ensemble Prediction System to study ensemble forecasted area rainfall in the Haihe River basin.We evaluate and analyze the applicability of ensemble area rainfall forecasts from May to August of 2020 to 2022,utilizing high-resolution observed grid data.Subsequently,we develop ensemble area rainfall products and span forecasts based on our findings.Our analysis reveals several key insights:(1)The ensemble mean forecast of area rainfall in the Haihe River basin exhibits lower absolute error compared to the control forecast,with larger spatial distribution errors observed in the southern regions,followed by the central and northern regions.(2)Fuzzy scores indicate that the ensemble mean area rainfall forecasts for light and moderate rain closely align with observations,while forecasts for rainstorms require further consideration of ensemble extremes.(3)The ensemble mean demonstrates higher TS and Bias scores for light rain across all forecast times,with more improvements in moderate rain and higher categories during later forecast periods.(4)Probability forecasts suggest that CMA-REPS performs better for lower rainfall predictions in the central rivers of the Haihe River basin,while scores for the southern rivers are lower than those for the central rivers.We develop ensemble forecast area rainfall products and span forecast products and test them on two flood-causing rainstorm events in the Haihe River basin.Results indicate:(1)Ensemble members show a higher probability of predicting 24-hour intense area rainfall equivalent to observed magnitude,serving as warnings for extreme rainstorms.(2)Ensemble mean forecasts within 24 hours exhibit better scores,while 75%quantile products are more significant for hea
作者
王静
陈宏
徐威
颜京辉
徐姝
WANG Jing;CHEN Hong;XU Wei;YAN Jinghui;XU Shu(Tianjin Key Laboratory of Marine Meteorology,Tianjin 300074,China;Tianjin Meteorological Observatory,Tianjin 300074,China;Tianjin Weather Modification Office,Tianjin 300074,China;China Meteorological Administration Hydro-Meteorology Key Laboratory,Beijing 100081,China;Center for Earth System Modeling and Prediction of China Meteorological Administration,Beijing 100081,China)
出处
《大气科学学报》
CSCD
北大核心
2024年第4期653-668,共16页
Transactions of Atmospheric Sciences
基金
国家自然科学基金资助项目(42205166)
中国气象局青年创新团队项目(CMA2024QN05)
中国气象局复盘总结专项(FPZJ2023-005)
中国气象局水文气象重点开放实验室开放研究课题(23SWQXM005)
海河流域气象科技创新项目(HHXM202401)
环渤海区域海洋气象科技协同创新项目(QYXM202314)
天津市气象局科研项目(202312ybxm07)。
