甘肃省主要地质灾害精细化气象风险预警预报

Fine meteorological risk early warning forecast of main geological disasters in Gansu Province

甘肃省是我国地质灾害多发的省份之一,降水型地质灾害占比重。利用甘肃省2013-2021年4-10月地质灾情数据、加密降水观测资料和地面-雷达-卫星三源融合降水分析产品(CMPAS),选取有效雨量代表降水致灾因子,建立黄土高原和陇南山地有效雨量致灾概率方程,以有效雨量致灾概率、地质灾害潜在危险度、易损度为因子构建甘肃省地质灾害气象风险精细化网格预警模型,并应用实况降水资料和兰州中心气象台精细化网格定量降水预报产品,建立地质灾害气象风险精细化网格预报试验,检验模型的预警效果。结果表明:(1)基于有效雨量致灾概率分别给出了黄土高原和陇南山地地质灾害气象风险蓝色、黄色、橙色和红色预警的临界雨量阈值,其中,陇南山地蓝色和红色预警临界雨量阈值分别为40.6 mm和113.5 mm,远高于黄土高原的18.0 mm和73.6 mm。(2)确定了甘肃省地质灾害气象风险蓝色、黄色、橙色和红色预警的风险度判别指标,数值介于0.004~1.000,其中0.336~1.000表示红色预警。(3)甘肃省地质灾害气象风险精细化网格预警模型能够较好预警地质灾害事件,各等级预警的比例合理,且能有效降低高等级预警率和空报率,该模型具有较强的地质灾害气象风险预警能力。

Geological disasters frequently occur in Gansu Province,China,and the proportion of precipitationtype geological disasters is significant in this region.Based on the geological disaster data,encrypted precipitation observations and the CMA multi-source merged precipitation analysis system pertinent to April to October of each year from 2013 to 2021 in Gansu Province,the effective rainfall data were selected as the precipitation factor;and further the geological disaster probability fitting equations of effective rainfall for the two regions were established.A refined grid geological disaster meteorological risk early warning model is constructed using the disaster probability of the precipitation factor,potential risk of geological disaster,and vulnerability.Using real precipitation data and the fine gridded prediction forecasts(QPF)of the Lanzhou Central Meteorological Observatory,a mesh refinement forecast test of the risk model was established to test the geological disaster events that occurred in October 2021 in Gansu Province.The study results show the following:(1)Based on the disaster probability caused by effective rainfall,the critical effective rainfall thresholds of blue,yellow,orange,and red warning levels of geological disasters in the Loess Plateau and Longnan Mountains,respectively.Among them,the critical effective rainfall thresholds for blue and red warning levels in Longnan Mountains are 40.6 mm and 113.5 mm,respectively,which were significantly higher than the blue and red warning levels of 18.0 mm and 73.6 mm,respectively,in the Loess Plateau.(2)The risk discrimination indices of blue,yellow,orange,and red early warning levels of geological disaster meteorological risk in Gansu Province were determined.The index values ranged from 0.004 to 1.000,with values ranging from 0.336 to 1.000,indicating an early red warning level.(3)The refined grid geological disaster meteorological risk early warning model in Gansu Province can effectively provide a warning of geological disaster events,the proportion of each level of early warning is reasonable,and it can effectively reduce the high-level early warning rate and false alarm rate.Thus,the model shows a strong ability to provide geological disaster meteorological risk early warnings.