绍兴市水利测雨雷达试点定量降水估测算法与效果评估

Quantitative Precipitation Estimation (QPE) algorithm and eff ectiveness evaluation for Shaoxing hydrological precipitation radar pilot project

摘要:

为突破传统雨量站网空间覆盖不足与分辨率较低的局限，提升致洪暴雨监测能力，评估绍兴市水利 X 波段测雨雷达试点采用定量降水估测（QPE）产品的实际效果，获取雷达基数据并经过严格质量控制后，基于双偏振参量反演得到 QPE 产品。利用相关系数、相对平均误差、相对平均偏差和均方根误差等统计指标，对雷达 QPE 产品与地面雨量站实况数据进行定量对比分析。结果表明：选取的 4 场典型降水过程中，雷达 QPE 产品数据与雨量站实况数据保持高度的一致性，相关系数不低于 0.87，相对平均误差小于 0.11，表明系统性偏差较小。QPE 产品对不同类型降水的空间分布和强度均展现出优异的捕捉能力，尤其对强降水中心量级和落区估测准确。绍兴市水利 X 波段测雨雷达组网采用的“先组网再反演”QPE 算法有效且可靠，成功消除 X 波段雷达单站探测的固有衰减缺陷， 对长三角河网地区致洪暴雨的精准识别与预警具备重要的应用价值和推广潜力。

Abstract:

To overcome the limitations of traditional rainfall station networks in terms of insufficient spatial coverage and low resolution and to enhance the monitoring capability for flood-inducing heavy rainfall，this study aimed to systematically evaluate the performance of Quantitative Precipitation Estimation （QPE） products derived from the Shaoxing hydrological X-band hydrological precipitation radar pilot project. By acquiring radar base data and implementing rigorous quality control，QPE products were subsequently retrieved based on dualpolarization parameters. A quantitative comparative analysis between radar QPE and ground-based rain gauge data were conducted using statistical metrics including the Correlation Coeffi cient （CC），Relative Mean Error （RME）， Relative Mean Bias （RMB），and Root Mean Square Error （RMSE）. The evaluation results indicated that across four typical precipitation events，the radar QPE products maintained high consistency with rain gauge observations， with CC values no lower than 0.87 and RME less than 0.11，indicating limited systematic bias. The QPE products demonstrated strong capability in capturing the spatial distribution and intensity of diff erent precipitation types， particularly in accurately estimating the magnitude and location of heavy rainfall cores. The “network-first， retrieval-later” QPE algorithm adopted by the Shaoxing hydrological X-band hydrological precipitation radar network proved effective and reliable. This technical route successfully overcomes the inherent attenuation limitations of single X-band radar detection and holds signifi cant application value and promotion potential for the precise identifi cation and early warning of fl ood-inducing heavy rainfall in the river-network region of the Yangtze River Delta.