Turkish Cloud-Radiation Database (CRD) and Its Application with CDR Bayesian Probability Algorithm

ABSTRACT It is rather a very difficult task to determine ground rainfall amounts from few Special Sensor Microwave Imager/Sounder (SSMI/S) channels. Although ground rainfall cannot be observed from the space directly, but knowledge about the cloud physics helps to estimate the amound of ground rainfall. SSMI/S includes so much information about the atmospheric structure, however it cannot provide cloud micro-physical structural information. In such a situation, in the rainfall algorithm, besides the SSMI/S data, it is necessary to incorporate cloud micro-physical properties from an external data source. These properties can be obtained quite simply by the help of Cloud Resolving Model (CRM). Later, in addition to all available data, also micro-physical properties obtained from Radiative Transfer Model (RTM) help to determine the SSMI/S brightness temperatures (Brightness temperatures - TBs), which can then be correlated with Cloud-Radiation Database (CRD) data generation. SSMI/S satellite data and CDR provide a common basis for rainfall prediction procedure through CDR Bayesian probability algorithm, which combines the two sets of data in a scientific manner. The first applications of this algorithm, which is being used up today, is due to various researchers. In this work, in order to establish a reflection of available data processing CDR CRM University of Wisconsin - Non-hydrostatic Modeling System (UW-NMS) model is employed, which is first developed by Prof. Gregory J. Tripoli. It is also used by Turkish Meteorological Service by benefiting from radar network data, and finally 14 simulations are realized in this study. Moreover, one case study is fulfilled by using a 3X3 spatial filtering, and then radar data and result of CDR Bayesian probability algorithm are compared with each other. On 9 September 2009 at 03:40 GMT rainfall event on comparatively flat area matches far better with the retrieval values and hence the spatial rainfall occurrence extent and numerical values are very close to each other.Statistical features (9 September 2009 at 03:40 GMT) Rainfall occurrence indices (9 September 2009 at 03:40 GMT)