Development of Processing Algorithms for the Detection and Estimation of Multi-Layered Structures of a Windfield with Polarimetric Doppler Weather Radar

Abstract

The Technical University of Delft has started a project in the form of a Master Thesis to improve their existing and develop new algorithms for processing the measured data with the PARSAX radar for 2D-Windfield monitoring. A new algorithm is developed that detects and estimates the multi-layered structure of a windfield using the measured polarimetric Doppler weather data. It is studied which variables can be used and within which boundaries they will be for the measurement of precipitation targets. The processing of the measured data is improved and higher quality data is obtained. The high quality data is used to make a Velocity Azimuth Display (VAD) and study the pattern of a windfield. An existing algorithm is studied and tested to obtain the direction and strength of the windfield from the VAD. A model is developed that uses the radar configuration and a user divined weather model as input to simulate the Doppler velocity pattern that the radar would measure with this configuration. This model is used to simulate the pattern that would be observed in case of a multi layered windfield, to be used as input for the testing of the newly developed algorithm. The developed algorithm is finally tested on a real dataset. The obtained windfield parameters with the existing and developed algorithm are used as input for the developed Doppler velocity pattern simulation model to compare the simulated Doppler velocity patterns with the measured one. It can be concluded that the existing algorithm has problems and gives wrong results when more windfields are present. The developed algorithm gives more realistic results, retrieving wind characteristics for every vertical layer, and producing high resulting correlation between simulated and observed patterns on VAD.