Doppler Weather Radar Research and Development
NSSL Special Project – Optimal Use of Phased Array Radar for Multi-Mission Weather Surveillance and Aircraft Tracking
Funding Type: CIMMS Task III
Objectives
Develop a real-time data acquisition system for the storage and processing
of I/Q data on the phased array radar; develop the theory and implement
several important techniques related to multi-mission phased array radar,
including: refractivity retrieval, cross-beam wind, advanced tracking,
adaptive scanning, sidelobe canceling, pulse compression, and scattering
experiments.
Accomplishments
This project represents a unique collaboration between scientists at
the University of Oklahoma and the NSSL. The research is targeted to
investigate several facets of system performance and optimization with
the phased array radar (PAR) system. Ultimately, it is anticipated
that this work will assist in the scientific justification of the multi-mission
PAR (MPAR) project. The following paragraphs provide an overview of
the various aspects of the research being conducted for this project.
In each case, a strong collaboration with NSSL scientists is emphasized.
Refractivity retrieval (Palmer). An investigation of the potential of rapid refractivity retrieval has been conducted. The retrieval technique utilizes radar phase measurements of ground clutter to derive nearsurface refractivity, which has been commonly used as a proxy for humidity given its close relation to vapor pressure. Surface humidity is an important meteorological parameter and has been known to play an important role in convective initiation. In the present work, the refractivity retrieval technique is exploited by using smaller numbers of samples for phase calculation, which is a fundamental process in refractivity retrieval. The impetus for this study is to explore the possibility of rapid refractivity retrieval by exploiting the rapid beam steering capability of phased array radar. Using the National Weather Radar Testbed (NWRT) in Norman, Oklahoma, a 64-pulse per radial raw dataset was collected for conventional refractivity processing. Then, subsets of the 64 samples were extracted to emulate shorter dwell periods and the corresponding more rapid experiments. The test cases that were considered are 2, 4, 8, 16 and 32 samples. Refractivity fields retrieved using smaller numbers of samples are compared against the reference field, which was obtained using the entire 64-samples dataset. It will be shown that statistically significant refractivity fields can be obtained from as short as a 2-sample dwell.
Boon Leng Cheong is supported part-time by this project and is in charge of the development of the radar simulator and the refractivity retrieval work. Working with NSSL scientists, we have just recently developed the real-time processing code on the PAR. By handling any scanning strategy, it is anticipated that we will have real-time refractivity measurements under any conditions. The new OU data acquisition node on the PAR is instrumental in this effort. We have recently submitted a journal article on PAR refractivity to the IEEE. Over the next year, we anticipate moving more to the meteorological applications working with CIMMS scientist at NSSL Pam Heinselman.
Target detection/tracking (Yeary). This work has evolved from target tracking to a more general area of adaptive scanning and storm tracking. Developing kinematic models, adaptive state estimation techniques, uncertainty measures and volumetric target libraries are crucial for the adaptive beam steering of phased array radar. In the characterization of real world dynamic systems, it is often required to estimate unknown quantities, such as future positions, based on given measurements. State space models are often used to approximate real world phenomena as systems, where the states (either observable or not) represent the internal variable that govern how systems evolve with time. In this framework, estimating the unknown quantities can be formulated in the problem of state estimation. The Kalman filter and the emerging particle filter are suitable for this task, and many open issues are present. Such as how can improved kinematic models be incorporated into statistical state estimation techniques to control the beam steering of modern-day multifunction radars? This will create precise beam location instructions for the real-time controller and one-step-ahead prediction algorithms for new phased array systems. Severe weather observations to be tracked, such as convective storm cells, may also be classified as a maneuvering target in some cases, which presents additional tracking challenges. When tracking a target which can perform maneuvers, the estimations generated by a conventional singlemodel method are not always accurate enough. The problem of maneuvering target tracking is often referred to as a jump Markov process, in which the system is assumed to operate according to one model from a finite set of hypothetical models, known as regimes or modes. This assumption is made at each iteration of the filtering process, as radar measurements are received. Since the current posterior probability distribution of a storm's trajectory can be computed, via a Kalman filter or other state estimation technique, future positions may be predicted with a high fidelity which may be used to adaptively steer radar resources or yield improved warnings. The long-term goal is to focus the resources of NWRT on a convective cell's trajectory, without making unnecessary scans. Enhancements to storm tracking and short-time prediction will be gauged against (Cheng, et al, 1996) and (Johnson, et al, 1998), with an eye to the future of advanced nowcasting (Jankowski, et al, 2005) and (JAG report, 2006).
Crossbeam wind measurement (G. Zhang). The theory of measuring crossbeam wind, shear, and turbulence within the radar's resolution volume V6 is developed. Spaced antenna interferometry is formulated for such measurements using phased-array weather-radar. The formulation for a Spaced Antenna Interferometer (SAI) includes shear of the mean wind, allows turbulence to be anisotropic, and allows receiving beams to have elliptical cross sections. Auto- and cross-correlation functions are derived based on wave scattering by randomly distributed particles. Antenna separation, mean wind, shear, and turbulence all contribute to signal de-correlation. Crossbeam wind cannot be separated from shear and thus crossbeam wind measurements are biased by shear. It is shown that SAI measures an apparent crossbeam wind (i.e., the angular shear of the radial wind component). Whereas the apparent crossbeam wind and turbulence within V6 cannot be separated using monostatic Doppler techniques, angular shear and turbulence can be separated using the SAI. The antenna patterns for the sum and difference channels have been measured (by NSSL) and calculated, as shown in the attached figure. The sub-array discretizations have been taken into accounted. It shows good agreement between measurement and theory for the main beam and side lobes, including the large side lobes at 12-17 degrees due to the sub-array discretization. These antenna patterns allow the antenna beam balance factor and effective separation calibrated for SAI applications.
Pulse compression (Chilson). Whereas the MPAR offers many exciting opportunities, there are still several technological challenges to overcome. One of these challenges has been precipitated through the reduction in available transmit power associated with the need to use solid-state modules. In order to maintain the same levels of signal-to-noise ratios enjoyed by the NEXRAD systems without compromising in range resolution, it will be necessary to implement one of several pulse compression strategies. Pulse compression is a process that provides for enhanced range resolution in radar systems. This is achieved through transmitting a coded signal, which is then decoded to extract the desired high-resolution data. Coding of signals is implemented by altering the phase, frequency, or both during its transmission while decoding is achieved through filtering. An inherent problem of pulse compression systems is the selfcluttering effect whereby targets at nearby ranges corrupt the desired data. As weather systems are distributed in nature, the performance metric used to measure the amount of self-clutter is the Integrated Sidelobe Level (ISL). The ISL is a measure of the total amount of power corrupting the desired data so that minimizing this metric means a cleaner signal is obtained. Minimizing ISL is realized through the use of different code and filter combinations, each with their own pros and cons. The study being conducted has been focusing on the use of binary phase codes whereby the phase of the transmit signal is flipped at specified intervals throughout transmission. Decoding of the return signals is then processed through a matched filter that maximizes the signal-to-noise ratio but has relatively large ISL. Near term efforts will focus on incorporating other types of filters that will suppress the side lobes corrupting the desired data.
Adaptive array processing (R. Palmer). Phased array radars are attractive in weather surveillance primarily because of their capability to electronically steer. When combined with the recently developed beam multiplexing technique, these radars can obtain very rapid update scans that are useful in monitoring severe weather. A consequence is that the small number of contiguous samples of the time series obtained is problematic for temporal/spectral filters used for clutter mitigation. As a result, the accurate extraction of weather signals can become the limiting performance barrier for phased array radars that employ beam multiplexing in clutter-dominated scattering fields. By exploiting the spatial correlation of the auxiliary channel signals, the effect of clutter contamination can be reduced in these conditions. In this paper, three spatial filtering techniques that used low-gain auxiliary receive channels are presented. The effect of clutter mitigation has been studied using numerical simulations of a tornadic environment for changes in signal-to-noise ratio, clutter-to-signal ratio, number of time series samples, varying fading regimes, and maximum weight constraints. Since such data are not currently available from horizontally-pointed phased array weather radar, experimental validation was applied to an existing data set from the Turbulent Eddy Profiler, which is vertically-pointed phased array radar. Although preliminary, the results show promise for clutter mitigation with extremely short non-uniform sampling.
Adaptive scanning (T. Yu): Phased array radar can simultaneously and dynamically position its beam to perform adaptive scanning and multiple tasks, for example. Unlike WSR-88D that only limited Volume Coverage Patterns (VCP) are available, a phased array system offers abundant possibilities for weather sensing. Therefore, new scanning strategies, which take full advantage of the phased array system, should be developed. At the same time due to the complex nature of the scanning strategy and limited radar resources, a dynamic resource manager is needed to maximize the radar performance. The goal of this project is to develop a framework of resource management, including two major components of priority assignment and scheduler. This framework will be demonstrated through simulations. We have been conducting literature survey on this topic, mainly for military applications. We will then modify their approaches for meteorological observations considering various types of weather such as single-cell storm, multi-cell storm, squall line, and supercell. Moreover, we have been experimenting different approaches for priority assignments based on severity of the region, which is currently defined by the reflectivity and radial velocity. Note our framework is developed using a modular approach that is flexible for incorporating additional requirements if necessary.
Scattering experiments (Y. Zhang and G. Zhang): To date, the following has been accomplished: (1) furbishing the lab room with broadband absorber materials; (2) scatterometer system: the NWA is now ready and system calibration with continuous wave mode and 12 calibration sphere is being conducted. Even before the new EM absorbers arrive, we are trying to control the in-door clutter scattering by searching the strong reflection spots and moving target locations; (3) components and parts: we have submitted all the purchase orders for the scatterometer experiments. We are working on the microphysics sample assembly parts and materials; (4) recruiting graduate student: is still in progress and we expect an EM-major graduate student come in spring 2008. Before that, using part-time student lab assistant is very possible; and (5) some natural hail-storm samples have been collected and stored in a freezer (with UPS battery supplied). For fall 2007, the following will occur: 1) EML will have its first external user from Intelligent Automation, Inc. for their SBIR project, which will perform time-domain test an X-band receiver assembly box during a 3-day period; 2) all absorbing materials will arrive and be installed; 3) and a better reflection-controlled room will be delivered at that time. Measurement will be extended to wider-bandwidth, longer range and pulsed waveforms.
These projects are ongoing.
Publications
Cheong, B. L., R. D. Palmer, and M. Xue, 2007: A time-series
weather radar simulator based on high-resolution atmospheric model simulations. J.
Atmos. and Oceanic Technol., in press.
Cheong, B. L., R. D. Palmer, C. Curtis, T.-Y. Yu, D. S. Zrnic, and D. Forsyth, 2007: Refractivity retrieval using the phased array radar: First results and potential for multi-mission operation. IEEE Trans. on Geosci.. and Remote Sensing, submitted.
Le, K., R. D. Palmer, B. L. Cheong, T.-Y. Yu, G. Zhang, and S. Torres, 2007: On the use of auxiliary receive channels for clutter mitigation on phased array weather radar. IEEE Trans. on Geosci. and Remote Sensing, submitted.
Zhang, G., and R. J. Doviak, 2007: Spaced antenna interferometry to measure crossbeam wind, shear, and turbulence: Theory and formulation. J. Atmos. Oceanic Technol., 24, 791-805.
Yu, T.-Y., M. B. Orescanin, C. D. Curtis, D. S. Zrnic, and D. E. Forsyth, 2006: Beam multiplexing using the phased array weather radar. J. Atmos. Oceanic Technol., 24, 616-626.
Alberts, T., P. Chilson, B. L. Cheong, R. D. Palmer, and M. Xue, 2007: Evaluation of binary phase coded pulse compression schemes using a time-series weather radar simulator. 33rd Conf. on Radar Meteorology, Cairns, Australia, Amer. Meteor. Soc.
Doviak, R. J., and G. Zhang, 2006: Crossbeam wind measurements with phased-array Doppler weather radar. ERAD06, Barcelona, Spain.
Reinoso-Rondinel, R., T.-Y. Yu, and R. D. Palmer, 2007: Investigation of Doppler spectra from a tornadic supercell thunderstorm: Are they Gaussian?, 33rd Conf. on Radar Meteorology, Cairns, Australia, Amer. Meteor. Soc., P8.B6.
Zhai, Y., M. Yeary, and J.-C. Noyer, 2006: Target tracking in a sensor network based on particle filtering and energy-aware design. IEEE-IMTC, Sorrento, Italy, 1988-1992.
Zhang, G., and R. J. Doviak, 2007: A theory for phased array weather radar to measure crossbeam wind shear and turbulence? 23rd Conf. on IIPS, San Antonio, TX, Amer. Meteor. Soc.
Zhang, G., R. J. Doviak, and X. Chen, 2007: Spaced-antenna interferometry to locate discrete objects and sub-volume inhomogeneities of reflectivity? 33rd Conf. on Radar Meteorology, Cairns, Australia, Amer. Meteor. Soc.
Comparison of the theoretical and measured sum (top) and azimuth difference (bottom) receive patterns. Theoretical patterns assume a designed sidelobe level of -28 dB, N = 5 (N determines the number of close-in side lobes that have about a -28 dB level), and take the sub-array discretization into account.