Cloud Optical Thickness And Effective Particle Radius: The Method Developing Of Ground Based Retrieval And The Impact Of Cloud Inhomogeneity On The Satellite Retrieval

Cloud optical thickness and effective particle radius are the important parameters to parameterize the cloud optical properties. Satellite and ground based retrievals are the main techniques to obtain these two parameters. However, more theoretical studies and observations are still required to validate the current methods. In this paper, we present a ground based retrieval method for determining the cloud optical thickness and effective particle radius and discuss the impact of cloud inhomogeneity on the satellite retrieval. The main results are summarized as follows:1. Develop a ground based retrieval method. A method is presented for determining the optical thickness (τc) and effective particle radius ( re ) of stratiform cloud layers from transmitted solar radiation measurements. A detailed study shows that the cloud optical thickness and effective particle radius of water clouds can be determined from transmission function measurements at 0.75 and 2.13μm , provided that the scaled optical thicknessτ0' .75>1 and re >5μm . The wavelengths adopted by our study are similar to the channels of the moderate resolution imaging spectrometer (MODIS). The proposed method is invalid for optically thin clouds since transmission at 2.13μm is less sensitive to re . The retrieval errors ofτ0' .75 and re monotonically decrease with increasingτc. For clouds havingτ'0.75≥2,the retrieval errors ofτ0' .75 and re are below 10% and 20%, respectively. Transmissions at 0.75 and 1.65μm can also be used to retrieveτc and re .2. Study the impact of cloud inhomogeneity on the satellite retrieval simultaneously. Two 3D cloud fields, cumulus and stratocumulus fields are used in our study. The two cloud fields are generated with LES (Large Eddy Simulation) model and are adopted by the I3RC phase II. The BRDF at the wavelengths of 0.65 and 2.13μm are calculated with SHDOM, and then the cloud optical thickness and effective particle radius are simultaneously retrieved using the MODIS algorithm. We studied the impact of cloud inhomogeneity on the satellite retrieval at high and medium resolutions. Some conclusions have been made, such as the BRDF abnormal phenomenon, the reverse relationship betweenτc and re , the clustering effect of over and under estimation in the retrieval ofτc and re .3. Present new methods for determining the true value of effective particle radius of 3D cloud field and for scale averaging of high resolution 3D cloud filed. We used theτc as the weight to determine the true value of re of each grid column of 3D cloud field. This method is more appropriate than using single re value for the whole cloud field, which is adopted by most studies. Thus, it is more valuable for the quantitative study. A drifting box method is developed to average the BRDF at different scales. This method can make sure that all the pixels involved in the averaging are filled with cloud.4. Found the clustering effect of over and under estimation in the satellite retrieval. By detailed numeric simulations, we found that the scale averaging cannot affect the overall tendency of the retrieval. The over and under estimation of the results have clustering effect. This effect shows that the cloud inhomogeneity can affect the retrieval even under low resolution. Correction scheme must be taken into account for current satellite methods, including the MODIS algorithm.