AREM模拟云参数与卫星观测的比较研究

Comparison of Cloud Parameters Between AREM Simulation and Satellite Retrieval

数值模式模拟计算云量与卫星观测云量的定义不同,难以进行定量比较,从而限制了对云模拟误差的检验和修正。作者介绍了可把数值模式输出云参数转换成ISCCP卫星反演云参数的云模拟器,并将其试用于AR-EM区域模式中,定量分析了模拟云参数和ISCCP反演云参数的差异。结果表明,AREM模式可以再现中国东部冬季云的基本分布特征,但与观测相比,四川地区云量模拟偏大20%,华东地区则偏小20%左右,云量中心的云顶高度比ISCCP反演值偏低一个高度段,云光学厚度偏大一个厚度段。模拟与观测云参数的定量比较,为评估和改进模式性能提供了依据。

Clouds play an important role in the Climate System. It is one of the key works in the model validation to verify and improve the model simulation of the clouds. However, The high, middle and low cloud amounts derived from the model, the surface observation and the satellite-retrieved data have different definitions, and cannot be compared with each other quantitatively, which restricts the model improvement for a long time. In recent years, Klein, Jakob, and Webb developed a method called International Satellite Cloud Climatology Project （ISCCP） cloud simulator, to convert the cloud products from the model to those defined by the ISCCP satellites. Using the simulator, some simulated cloud parameters, such as cloud amount （42 types）, cloud top pressure and cloud optical thickness, can be compared separately and quantitatively with the ISCCP data, providing us a further probability to correct the model errors. In this paper, the ISCCP cloud simulator is introduced firstly, and added into a regional numerical model AREM （Advanced Regional Eta-coordinate Model） to convert the simulated cloud parameters to those defined by the ISCCP data. Under the NCEP initial fields and time-varying lateral boundary condition forcing, the AREM model is integrated continuously from 31 January 2001 to 28 February 2001, outputting every 3 hours. The horizontal and vertical distributions of the cloud parameters converted by the simulator such as cloud amount, cloud topical pressure, and cloud optical thickness are compared with the ISCCP data quantitatively. The results show that the AREM model can basically simulate the distribution of the horizontal cloud amount and the variation of the vertical cloud amount over the eastern part of China, but the monthly mean total cloud amount is smaller, cloud top height is lower and the cloud optical depth is thicker than observation over Sichuan and Guizhou area （28°N- 32°N, 102. 5°E- 110°E ）. The authors speculate that the errors come from the diagnostic cloud amount scheme and the optical thickness calculation in the model. The cloud amount of the 42 cloud types from the simulator is also compared to those from the ISCCP data, and it is concluded that the simulated stratiform cloud is relatively agreed with the data, while the convective cloud is not. It may be resulted from the small convective cloud amount in winter and the limited capability of the model simulation over the Tibetan Plateau. The ISCCP cloud simulator has many priorities and may play an important role in the model validation. However, it has some uncertainties and should be used carefully. The comparison of the single cloud type should not be carried out unless the model system errors are corrected. Some technical problems such as the number of the columns in the grid and the method defining the cloud top pressure should be further confirmed. Some day, the surface observation also should be coupled into the simulator to make the comparison more effective.