基于漂浮式测风激光雷达观测数据的海上预报风速订正

Correction of Offshore Forecast Wind Speed Based on Floating LiDAR Observation Data

本文阐述了海上漂浮式测风设备对数据的测量方式和校准方法后,利用三部漂浮式测风激光雷达50、80和100m三个高度的风速数据,对同地点和高度的GFS全球天气预报模式的风速数据进行预报订正。结果表明,直接采用滑动自适应权重的Kalman滤波方法产生累积的滞后平均偏差对风速进行订正的结果并不理想,统计发现通过降低预报订正公式中的滞后平均偏差项的比值,可以获得更好的标准偏差结果。滞后平均偏差比值减小至0.1~0.5后经Kalman滤波方法订正后的风速标准偏差整体要低于订正前的风速标准偏差,其中最佳的比值为0.3。利用这一滞后偏差比值对三部雷达不同预报时刻和预报时效的风速样本组进行订正后发现,除了个别数据样本组外,超过90%数据样本组的风速标准偏差较订正前是减小的。经改进的算法订正后各高度层风速预报值的改善率大致在5~20%范围内。

This study focused on correcting short-term wind speed forecasts extracted from GFS global forecast data sets at corresponding locations and heights.Data measurement mode and calibration method of the LiDAR data were discussed,and wind speed at three different heights(50m,80m,and 100m)from floating LiDARs were employed.The results indicated that directly using the Kalman filter method with sliding adaptive weights to generate the average hysteresis deviation to correct the wind speed did not yield satisfactory results.Statistical analysis suggested that better standard deviation results can be obtained by lowering the ratio of the average hysteresis deviation in the forecast correction formula.When the average hysteresis deviation ratio was reduced to 0.1—0.5,the wind speed standard deviation after correction by the Kalman filtering method was generally lower than the wind speed standard deviation before correction,and the best ratio was 0.3.Using this hysteresis deviation ratio for wind speed correction at different forecast times and periods for these three LiDAR devices revealed that,except for individual data sets,more than 90%of the wind speed data sets exhibited reduced standard deviation after correction compared to the uncorrected data sets.After the introduction of this improved algorithm,the wind speed forecast improvement rate at each height ranged from 5%to 20%.