A Globally Available Tropospheric Scintillation Model Using Numerical Weather Prediction Techniques

Abstract

This dissertation discusses the development and validation of a method for predicting the statistical distribution of tropospheric scintillation attenuation using data from the Global Forecast System (GFS). This model uses open-access tools and data and provides global availability. The model considers the physical structure of the atmosphere as opposed to other models that often depend on climatic averages of local parameters and measurements from propagation experiments. By considering the physical structure of the atmosphere, the model’s performance is location-independent, as opposed to empirically derived models which may show bias towards a single region. The model discussed in this work was arrived at by modifying scintillation models intended for use with high resolution radiosonde measurements. This work details the model’s operation as well as the methods used to overcome issues inherent to the relatively coarse GFS data. Scaling functions were developed for gradient-dependent parameters based on data resolution and altitude. These functions were derived independently of GFS data using high-resolution radiosonde datasets which were downsampled to various resolutions. This work also details the design and characterization of a terminal used to measure tropospheric scintilla- tion on an 18.444 GHz link in Auburn, AL. This data collection effort has been used to validate the model but not in its development to assure their independence. Comparison between scintillation predictions using the modified GFS model and a model using high-resolution radiosonde data shows that the modified model has good agreement at times and universally outperforms radiosonde models using unmodified GFS data. Comparison with measured scintillation values shows good agreement between the modified GFS model, traditional models using wet index of refractivity calculated from measured values, and traditional models using wet index of refractivity calculated from GFS surface parameters. These results indicate that surface parameters and profiles from GFS are independently capable of providing quality scintillation predictions, and a combination of the two may hold potential for higher quality predictions in future work.