中尺度数值预报模式预报水汽与GPS观测的比较研究

A Comparison Study of Precipitable Water Simulated by NWP and GPS Observations

利用建立在长江三角洲地区的GPS观测网中 4个站点资料 ,进行 2 0 0 2年梅雨期( 6月 1 8日～ 7月 1 1日 )MM5模式 2 4小时预报结果与GPS观测资料的比较研究。研究结果表明 :2 0 0 2年梅雨期 ,不同的GPS站点可降水量的变化与梅雨带的移动和走向有直接的关系。提高模式分辨率能减小模式初始场对湿度的分析误差 ,但对可降水量的预报能力改善不明显。尽管MM5模式积分一开始对可降水量的预报就存在 5 %～ 1 0 %的偏差 ,但积分前1 0小时模式对可降水量表现了较好的预报能力。积分 2 0小时后 ,模式对可降水量的预报偏差明显增长。研究还表明 :GPS对可降水量的测量与探空观测有较高的一致性。相距 43km时 ,两者仍有一定的可比性。GPS测量的可降水量资料可能存在系统误差 ,同化GPS资料时应关注GPS资料本身的可靠性。

Based on the GPS data at 4 sites of GPS networks in the Yangtze delta over the Meiyu period of 2002 ( June 18～July 11), precipitable water (PW) evaluated by MM5 24h simulation is compared with GPS observation to validate the ability of MM5 to simulate water vapor as depicted by PW The results show that PW features observed at GPS sites directly relate to the movement and position of the Meiyu band during Meiyu period The humidity error of MM5 reanalysis fields can be minimized by increasing the horizontal resolution of the model, but it has a little positive impact on PW prediction Although a 5%～10% relative bias is found from the beginning of simulation, MM5 has a ability in PW prediction over 1h～10h integration period with the PW prediction error increasing obviously after 20h integration It is also found that the bias of PW between GPS observations and radiosondes is small GPS data is comparable to radiosondes even if GPS site is 43 Km far from radiosonde site The quality of GPS data must be considered before the data is fed into NWP model because it may have systematic errors