PyDDA: A New Pythonic Multiple-Doppler Analysis Package

Current packages for multiple Doppler analysis of radar winds are difficult to use and are based off of legacy optimization algorithms do not perform well on modern computational infrastructure. To overcome these issues we have designed a new multiple Doppler package written in Python, relying only on the scientific Python ecosystem, that is based off of the 3D variational technique for retrieving winds from multiple radars (i.e. Potvin et al 2012). We call this Python Direct Data Assimilation or PyDDA. The initial aim of PyDDA is to provide an easy to use multiple Doppler package in which a retrieval can be done with just 4 lines of code of Python, and visualization done on a fifth line. We will demonstrate its ease of use by showing an example Jupyter Notebook on retrievals from multiple X-band Scanning Atmospheric Radiation Measurement (ARM) Precipitation Radar (X-SAPRs) over the ARM Southern Great Plains (SGP) site. Due to its simplicity, PyDDA can be used for classroom instruction on multiple Doppler retrievals. Finally, given that all dependencies are available by package managers like Conda, users of Windows, Linux, or OS X can use PyDDA covering most users in the radar community.

Another advantage of PyDDA is its performance. We use SciPy’s limited memory BFGS-B optimization code to perform the optimization. A typical wind retrieval can be performed in 30 to 90 seconds on a laptop. Compared to its predecessor, Multidop, a Python wrapper around an external C code for Dual Doppler analysis, PyDDA is typically a factor of 4 to 5 faster for the same grids. Finally, PyDDA scales with Dask better than Multidop as it does not depend on any calls to external programs. Ultimately we are looking expand this into a framework that encompasses more sensors besides radars as well as models. We welcome collaborations and contributions from the community at large. PyDDA is available at: https://www.github.com/rcjackson/PyDDA.

References:

Potvin, C.K., D. Betten, L.J. Wicker, K.L. Elmore, and M.I. Biggerstaff, 2012: 3DVAR versus Traditional Dual-Doppler Wind Retrievals of a Simulated Supercell Thunderstorm. Mon. Wea. Rev., 140, 3487–3494, https://doi.org/10.1175/MWR-D-12-00063.1