Creation and Evaluation of the Freezing Rain Accumulation National Analysis (FRANA) in Preparation for NWS Operations Weather and Forecasting

In response to a National Weather Service (NWS) requirement for a gridded analysis of ice accumulation, the Freezing Rain Accumulation National Analysis (FRANA) is created for nowcasting and postevent verification of freezing rain/drizzle (FZRA/FZDZ) ice accumulations. This product generates CONUS-wide hourly flat and radial ice accumulations for 1-, 3-, 6-, 12-, and 24-h windows. The core science algorithm within FRANA that converts liquid to ice accumulation is the Freezing Rain Accumulation Model (FRAM). The inputs to FRANA come from the High-Resolution Rapid Refresh analyses of 2-m wet-bulb temperature and 10-m wind speed and the Multi-Radar Multi-Sensor (MRMS) hourly multisensor quantitative precipitation estimate. An additional trace ice footprint was added to FRANA to help forecasters identify areas of light FZRA/FZDZ. A 3-yr retrospective analysis reveals that FRANA can overestimate or underestimate the icing footprint due to NWP temperature uncertainty, precipitation-type error, radar overshooting, precipitation evaporating below the lowest radar tilt, or MRMS quality control. For accumulating ice, FRANA has a modest RMSE of 1.27 mm (0.05 in.) and event-maximum errors typically less than 2.54 mm (0.1 in.). In fact, 88% of the events in the study had less than 2.54-mm (0.1 in.) absolute error which is within the tolerance range for many operational applications. Additionally, lower storm-total accumulations are found to be associated with lower absolute error. Example events show how FRANA can provide guidance to forecasters on the spatial extent of icing as well as “hotspots” where higher accumulations may cause significant socioeconomic disruptions. Grant no. NA22OAR4590169

为响应美国国家气象局（National Weather Service, NWS）对积冰累积量格点分析的需求，冻雨累积量全国格点分析（Freezing Rain Accumulation National Analysis, FRANA）应运而生，用于冻雨/冻毛毛雨（FZRA/FZDZ）积冰累积量的临近预报与事后验证。该产品可生成覆盖美国本土（CONUS）的逐小时平面积冰与径向积冰累积量，覆盖时段涵盖1、3、6、12及24小时。FRANA中将液态降水转换为积冰累积量的核心科学算法为冻雨累积模型（Freezing Rain Accumulation Model, FRAM）。其输入数据来自高分辨率快速刷新模式（High-Resolution Rapid Refresh）的2米湿球温度、10米风速分析场，以及多雷达多传感器系统（Multi-Radar Multi-Sensor, MRMS）的逐小时多传感器定量降水估计产品。FRANA额外增设了微量冰影响范围识别模块，以帮助预报员识别弱冻雨/冻毛毛雨影响区域。一项为期3年的回溯分析显示，受数值天气预报温度不确定性、降水类型误差、雷达波束超探、最低扫描仰角下方降水蒸发，或MRMS质量控制问题等因素影响，FRANA可能会高估或低估积冰影响范围。针对累积性积冰，FRANA的均方根误差（Root Mean Square Error, RMSE）为1.27毫米（0.05英寸），误差表现适中，事件极大值误差通常小于2.54毫米（0.1英寸）。事实上，本研究中88%的事件绝对误差小于2.54毫米（0.1英寸），这一精度符合多数业务应用的容忍范围。此外，研究发现风暴总累积量越低，对应的绝对误差也越小。典型案例展示了FRANA可为预报员提供积冰影响空间分布范围的指导，以及可能引发显著社会经济中断的积冰高值“热点区域”。资助编号：NA22OAR4590169