Radar echo classifier algorithm development using Python

Our research to improve the quality of NEXRAD products requires us to efficiently process large volumes of meteorological radar data. We have developed a radar data processing environment that processes time series data into Level 2 products (base data), classifies radar echoes, and compensates reflectivities for meteorological echoes that have been attenuated by the radar's clutter filters. This processing environment was built using Numeric Python, an interpreted, interactive, object-oriented programming language. Using Numeric Python allows us to build and test complex algorithms using an interactive interpreter. We can easily build small test cases that test each portion of our algorithms. We can quickly build new algorithms by connecting the output of one algorithm to the input of another algorithm. Python's object-oriented features makes it easier to organize complex algorithms. Since Numeric Python is quite efficient, we can also build "production", real-time processing systems. We use various extensions to Python to plot our data and build graphical user interfaces to set algorithm parameters. Numeric Python has allowed us to build flexible software that efficiently processes large data-sets.