An end-to-end emulation of the CASA radar network

An End-to-End Emulation of the CASA Radar Network

In an effort to test the critical software components being developed for the initial radar network to be deployed by the Engineering Research Center for Collaborative Adaptive Sensing of Atmosphere (CASA), we have designed and implemented an end-to-end software-based emulator. Adaptive sampling of the atmosphere in response to the weather and end-user needs is an important theme of the Center and for its network of small, low-cost X-band radars. The initial deployment of 4 radars is scheduled for early 2006.

The system emulates the sampling of a model simulated atmosphere by the radars, the collection of the data at the operations center, the detection of severe weather features from the radar data, the prioritization of tasks based on the detected features and the optimization of scan strategies for those tasks, and finally the issuance of commands to direct the radars to sample during the subsequent time period.

The tests reported in this paper use numerical weather prediction model simulations at horizontal resolutions up to 25 meters for input data. In one realization the model simulates the life cycles of multiple tornadoes, with the strongest one having an F5 intensity. A radar emulator samples this model atmosphere in a realistic way from the four planned radar sites of the Oklahoma test bed. The radar emulator is configurable for different scan strategies and different radar operating characteristics, including gate-spacing, pulse repetition frequency, attenuation, rotation rate, and scan sector limits. The sampled radial velocity and reflectivity data are relayed to the central processing location via the same protocols anticipated for use in the test bed. Meteorological algorithms are being developed to detect and diagnose the severe weather signatures (localized rotational shear and storm cells) using data from the emulated CASA radars and other platforms. Important weather features recognized by the detection algorithms, the application requirements (e.g., for initializing a numerical weather prediction model, quantitative precipitation estimates, and feature pinpointing) as well as requests from end-users, are input into the optimization algorithms, which use a set of policies, formulated in consultation with end users, to determine the optimal radar scan strategy. In the case of the system emulator, this optimal decision is used to 'operate the virtual radars'.

The goal of this experiment is to fully test all pieces of our modular software before real data are available and to perform initial validations of the networked system in simulated real-time settings, before the actual radars come online.