Sensing

The Sensing research thrust is engaged in developing and testing highly connected networks of Doppler radar hardware that transmit short-range microwave pulses into the lower atmosphere and measure numerous meteorological conditions at smaller scales and greater frequency than ever before. Our research will define, design, fabricate, and field-test solid-state, low-cost, rapid-scan, multi-beam radars, along with the unique collaborating configurations that connect these units to create DCAS systems.

The key parameters of these small-scale radars are transmitter power, beam requirements, antenna size, operating frequency, and bandwidth. The Sensing thrust is conducting research into the interaction of electromagnetic waves with the atmosphere in both stormy environments and in clear weather. The team will study the physical mechanisms by which DCAS networks measure precipitation, detect winds, temperature, and humidity. View a multimedia presentation on the Sensing thrust and radars by Sandra Cruz Pol, CASA Associate Director, University of Puerto Rico Mayaguez. (Note: RealPlayer is required to view this presentation.) To see the results of our research, please visit the Sensing Thrust publications webpage.

The Sensing thrust involves projects being conducted at all four of CASA's partner campuses. Sensing is divided into three sub-thrusts, Radar Network, Solid State Radar, and CLEAR.

• Radar Network Sub-Thrust
  This sub-thrust is responsible for designing, fabricating, and field-testing CASA's first generation radars; defining critical design parameters such as range/Doppler ambiguity structure, waveform design, polarization basis; establishing the fundamental scientific basis for network-based retrieval of Doppler and reflectivity in the presence of clutter, range/Doppler ambiguity, and attenuation; and single radar calibration/validation prior to deploying radars in CASA's system-level test beds.



• Solid State Radar Sub-Thrust
  The goal of this sub-thrust is to research, design and fabricate, and test prototype electronically scanned, low-cost radars.
  
• CLEAR Sensing Sub-Thrust
  This research thematic area's focus is exploratory or "look ahead" research in probing the clear (non-precipitating) atmosphere. The research challenge here is exploiting sources of scattering or other perturbations of electromagnetic waves to determine atmospheric parameters such as boundary layer wind fields, humidity fields, and temperature fields that may be used to aid in such applications as the onset of severe storms, air quality prediction, and boundary layer transport of pollutants or other hazardous agents.

Research Projects
Our research contributions in the Sensing Thrust range from designing retrieval algorithms to developing fundamental precipitation observation techniques to investigating longer-time horizon problems such as choice of a scattering mechanism for clear-air sensing.

Dual-polarization Rain Profiling Algorithm for Attenuation Mitigation
This research activity is critically important for enabling X-band radar observations of precipitation. Attenuation correction at X-band is significantly more challenging than at C-band, where the influence of differential phase on backscatter is negligible. Our improved X-band attenuation mitigation algorithms incorporate differential reflectivity measurements into rain profiling by taking the variation of raindrop distribution into account. The consistency of our dual-polarization rain profiling algorithm is assured by incorporating differential propagation phase measurements as an optimization constraint. We have tested and validated our attenuation correction algorithms through a microphysical model-based simulation in which X-band radar observations are obtained from high-resolution S-band precipitation measurements.

Sensing Node Processing Software System
A high performance signal handling software system has been implemented in the Sensing Node Store and Processing Computer (SNSPC). The software processes digitized I/Q samples in real-time into reliable Doppler parameters and dual polarization radar moments. I/Q data from a wide-bandwidth link are converted into processed data for distributing through a medium-bandwidth link. The algorithms in the SNSPC implement functions such as clutter filtering, second-trip suppression and recovery, velocity unfolding, parameter estimation, and attenuation mitigation. The SNPC software package also provides for radar data display, storage, and transport.

Waveform Design Tables Developed for IP1 System for Range-Velocity Mitigation and Clutter Filtering

Range-velocity ambiguity is a well-known problem in weather radars. This problem becomes even more challenging as we move up in frequency to X-band. We have designed a waveform of varying PRFs with random phase coding to address range-velocity resolution. The waveform design must also address additional constraints based on hardware compatability, dwell time, clutter filtering, and accuracy requirements. Since the scan strategy of an individual sensing node changes adaptively in a DCAS system, the waveform must also change accordingly. For a given scan strategy, we have developed a set of range-velocity ambiguity resolution solutions that are optimized in terms of the parameters of dual PRFs to meet the accuracy requirements.

Choice of Scattering Mechanism for Clear Air Sensing
This research activity is an exploratory effort in anticipation of the CLEAR phase of CASA research. We have evaluated the comparative effects of Bragg scattering from refractive index fluctuations and of Rayleigh scattering from insects as a means to probe boundary layer winds. Our initial focus has been the exploitation of bistatic enhancement for Bragg scattering to "see" refractive index turbulence in the atmospheric boundary layer with short-wavelength radar. Using data collected by the UMass S-Band FMCW radar, we simulated the reflectivity fields that would be detected by monostatic and bistatic radars (at both S-band and X-band) including both the influences of insects and Bragg scattering. Our main conclusion is that reliance on bistatic scattering to access otherwise invisible clear-air refractive index turbulence at short wavelengths is not warranted for the purpose of wind estimation in the boundary layer. Simulations of bistatic scattering derived from existing S-band monostatic observations indicate that, when present, insect scattering dominates for the majority of forward scatter geometries at S-band and all but the most extreme forward scatter geometries at X-band. These particular forward geometries yield poor Doppler sensitivity. In addition, the conditions in which insect scatter is unlikely to dominate correspond to cold, dry, shallow boundary layers, where the clear-air signal is weak, intermittent, and competes directly with ground clutter. In continental environments, we believe that improved understanding of insect scattering for wind estimation is needed.

LMS Retrieval of Surface-Layer Refractivity from One or Multiple Radars
The potential to retrieve surface layer refractivity fields, which is a function of moisture and temperature, from networks of radars is being investigated for clear air sensing. With a single radar, the refractivity field can be obtained through observation of echoes from stationary ground targets. In a DCAS system, multiple radars can be steered towards a common region to retrieve the surface-layer refractivity field with minimum mean-square error. We have demonstrated that an accurate discrete refractivity field model enables classical signal processing approaches to be employed to improve the desired inversion and to extend it to the multiple-radar scenario. In the inversion process, "phase unwrapping" is particularly critical for the CASA radars operating at X-band. A novel algorithm has been developed that improves the estimation of unwrapping parameters jointly with the refractivity field, especially in regions of sparse targets. A statistical Bayesian approach has been developed to further improve the refractivity inversion. Preliminary results on small fields indicate the improvements that will be possible using this technique with a network of radars.