SR-50A超声雪深仪在西藏高原的适用性研究

Applicability Study of SR-50A Ultrasonic Snow Depth Sensor for Snow Measurement in Tibetan Plateau

为满足应急气象服务需求,2013 2014年在西藏自治区强降雪和雪灾易发及重点积雪区域气象站安装了4套SR-50A超声波雪深观测系统,首次实现了西藏高原雪深自动观测和数据实时传输。利用12:30加密和08:00(北京时)常规人工雪深观测数据对4个站SR-50A雪深观测数据进行了评估和对比分析。结果表明:(1)SR-50A与人工观测的平均雪深偏差范围在±2 cm之内。雪深越大,平均均方根误差越小,观测精度越高。SR-50A传感器更为适合雪深较大地区的积雪监测。(2)SR-50A对西藏高原的雪深具有较好的监测能力,与人工观测雪深具有较好的一致性,4个观测点的线性相关系数在0.81~0.97,呈现极为显著的线性关系。(3)大风、局地太阳光照条件、气温和地表特征等因素通过风吹雪和融雪引起观测场内积雪分布不均匀,加之仪器是固定点观测,人工观测是观测场内3个点的雪深平均值,这些是SR-50A与人工观测雪深差异较大的主要原因。

Snow depth is one of main surface meteorological observation elements and currently is mainly measured by manual observation. Due to poor real-time performance, it is difficult to meet requirements of emergency meteorological service and response. Previous study indicated that ultrasonic technology has excellent potential for snow depth measurement. Therefore, the SR-50A ultrasonic snow depth sensors developed by Campbell Scientific Company were installed from September 2013 to November 2014 at 4 meteorological stations where have often heavy snow and are prone to snow disaster and key regions of snow accumulation in Tibetan Plateau. In this study, the applicability evaluation of SR-50A snow depth in these stations was made using daily 12：30 intensive and 08：00 （Beijing Time） manually observed snow depth as ground truth data. The following are results. （1） Mean biases between SR-50A and manual snow depth are within ＋2 cm. SR-50A sensors are more suitable for the measurement of higher snow depth since the larger the depth of snow, the smaller the mean root-mean-square error and the higher the observation precision. （2） SR-50A shows a good performance for monitoring snow depth in Tibetan Plateau. There is a very good agreement and highly significant linear correlation between manual and SR-50A sensor observation for snow depth. The correlation coefficients at 4 stations range from 0. 81 to 0. 97. （ 3 ） Wind scour, local solar radiation condition, environmental temperature and surface feature will cause spatial variability of snow cover at observation sites through wind blowing and snow melting, and manual observation in the study is not taken just adjacent to sensor installation, which are main factors resulting in these biases.