Advanced Radar Technologies: Capabilities and Opportunities
This chapter examines technological issues that are central to the concept of a longer-term technological view of a national weather radar system. The results are intended to extend 20–25 years into the future. The approach to the technology assessment centers on a new Radar Data Acquisition (RDA) system consisting of four key elements: the transmitter, the receiver, the antenna, and the processor— i.e., the hardware and software that produce the “base data” from which all relevant weather products are derived (also termed the “Level 2” data stream in NEXRAD). The discussion focuses on the most promising technologies for the NEXRAD replacement weather radar system.
The presentation is organized in four major sections. The first deals with a fundamental requirement of the NEXRAD system—preservation of the system’s integrity through retention of the enabling frequency allocation resource for weather radars. The remaining sections deal with valuable improvements over the current system and describe several related technologies. Topics include means of improving data quality to reduce interpretive uncertainties, quantifying precipitation and improving precipitation classification via polarimetric measurements, use of phased array antennas to reduce volume sampling times and enable agile beam steering, and use of innovative signal processing schemes to improve radar performance.
Two fundamental concerns must be resolved before new advanced radar designs using promising modern technologies may proceed. First, critical technical issues need to be addressed related to the presently high cost of transmit-receive elements inherent in a phased array radar compared to the potential benefit in scanning flexibility and system reliability. It is not clear whether solid-state transmitter amplifiers (either within the individual modules or in a single trans-
mitter configuration) offer enough advantages over high-voltage tube amplifiers to justify their inherent waveform constraints. Second, a political and economic debate continues between the communications industries (wireless and satellite radio) and the federal government regarding frequency allocation issues. It is clear today that future generations of mobile communication will apply pressure for expanded use of the existing S-band (10-cm) spectrum now used by groundbased weather and aviation radars. These potential revised spectrum allocations could limit or preclude radar operation in any of our existing “weather bands,” in particular the low-attenuation S-band.
Retaining the allocation of weather radar frequencies, particularly at S-band wavelengths with their relatively small attenuation in heavy rainfall, is critically important to maintaining the operational integrity of the NEXRAD radar system as well as preserving its capabilities for any future replacement system. Furthermore, an important advancement in the replacement system should be realization of more rapid volumetric sampling rate. This will most likely require use of radar waveforms that employ wider bandwidth than the current NEXRAD system. These frequency allocations are continuously being scrutinized for possible real-location to other sectors of society to meet increasing spectrum demands for communications and entertainment applications. The issue is international in scope, since many spectral-use applications extend beyond national borders. Regarding the latter, the International Telecommunications Union sponsors the World Radio Committee (WRC) meeting in Geneva every two years to discuss and decide on spectrum allocation issues. Thus, in addition to preserving these national allocations, U.S. representatives to the WRC meeting must constantly remain alert in order to preserve the global weather radar spectrum allocations (McGinnis, 2001). The importance and value of weather radar to society has been aptly demonstrated with the NEXRAD system, especially through its utility for providing critical warning of severe weather, monitoring precipitation, and helping ensure safe and efficient operation of the National Airspace System (NAS).
The present weather radar users—e.g., the NWS, the FAA and the DoD— must make convincing arguments to spectrum allocation authorities to preserve the S-band region for weather activities necessary to aviation safety, preservation of life and property, and even national security, especially in light of the current world terrorist threat. These arguments can most readily be based on the ability of the longer-wavelength S-band radar systems to penetrate heavy precipitation and allow the proper interpretation of hydrometeor scattering without the complications that arise when the hydrometeor sizes are large relative to the radar wavelength.
Policy makers and members of the operational community should actively participate in the arena of frequency allocation negotiation. The impact, including the economic and societal costs, of restrictions on operating frequency, bandwidth, and power should be assessed for current and future weather radar systems.
DATA-QUALITY ENHANCEMENTS—POLARIMETRY AND AGILE BEAMS
Collecting and processing base data (the radar reflectivity, the radial velocity, and the spectrum width parameters) and the derived diagnostic (meteorological) products provide the bulk of the operational experience with NEXRAD. These experiences reveal data-quality problems. These problems are being addressed in open-systems architecture activities. The problems should continue to be addressed in the future system. The impacts of data deficiencies on specific products are described at length in a previous NRC report addressing NEXRAD coverage (NRC, 1995) and in Serafin and Wilson (2000).
Data corruption usually results from such factors as range folding, normal and anomalous propagation ground clutter, velocity aliasing, radio frequency (RF) interference, improper maintenance procedures, and nonatmospheric reflectors such as birds or chaff. Depending on the situation, the impact of these artifacts on generating an accurate meteorological product varies between minimal and severe. Product degradation can take the form of an enlarged data void when contaminated data are detected and censored, or it can take the form of erroneous products when biased data are passed on to meteorological algorithms.
Experience has shown that the integration of data-quality analysis prior to data assimilation is an effective way for detecting and masking erroneous data, thereby preventing the introduction of faulty information into the product algorithms. An automated data-quality analysis system should be an integral component of the next generation radar system. The primary component should be automatic detection of known artifacts and flagging of that data for special treatment prior to generation of any products using the radar base data. Certainly, these data-quality issues must be addressed within the data assimilation scheme if not sooner. Even more important for proper data assimilation is the knowledge of error statistics of each data source. Not only must the instrumentation error be known, but also the representativeness error of the specific measurements must be estimated for effective assimilation by a numerical model.
The quality of real-time data should receive prominent consideration in the design and development of a next generation weather surveillance radar system. Real-
time data-quality assessment should be automated and used in deriving error statistics, and alerting users to system performance degradation.
Although several studies are underway to address these important data-quality issues, two technological developments can provide a major improvement—polarimetric observations and electronic, agile beam scanning.
Tests on polarimetric radar tests are already being performed as part of a potential WSR-88D upgrade in the next decade. Polarization diversity observations bring some unique characteristics that are important for addressing data-quality issues. First, without any additional effort, polarimetric measurements automatically suppress the second trip echoes by about 15–20 dB depending on the type of hydrometeors. Second, the precipitation back scatter at horizontal (H) and vertical (V) polarizations exhibits a high degree of coherency (>0.98 in rain) that can be used to detect and filter contamination from noise as well as from nonmeteorological echoes such as surface clutter, chaff, birds and insects. Third, the differential polarization parameters, such as differential reflectivity and specific differential propagation phase, are immune to absolute calibration errors. Furthermore, self-consistency constraints of the covariance matrix measurements in rain impose bounds on errors in absolute reflectivity measurement (Scarchilli et al., 1995).
Dual-polarized radar systems can be configured in different ways depending on the measurement goals and choice of orthogonal polarization states. Fully polarimetric radar measures the complete covariance matrix of precipitation in the resolution volume (Bringi and Chandrasekar, 2001). Radars can be operated with polarization agility where the transmit polarization is changed on a pulse-to-pulse basis and two orthogonal polarizations are received, providing polarization diversity on reception. Alternatively, radars can transmit and receive the same polarization states, or utilize a hybrid mode in which they are different. The hybrid mode of operation where both horizontal and vertical polarization states are simultaneously transmitted but separately received is the mode being considered for the current WSR-88D upgrade. One drawback of the hybrid mode is that it inhibits the cross-polarization measurements that are extremely useful for water/ice discrimination and bright-band detection, though similar information can be obtained through the other polarimetric measurements.
Precipitation Classification and Quantification
Current NEXRAD-based products for the estimation of precipitation rate are seriously deficient. The deficiencies have been identified as a limiting factor for the value of the NEXRAD system in support of hydrologic products, including
flash flood warnings. It is anticipated that the addition of a polarimetric capability to the NEXRAD will address, in part, these limitations. Dual-polarization measurements allow improved accuracy in the rainfall determination, more effective hail detection, and an effective means for characterizing hydrometeors throughout a storm volume. The improved accuracy in the determination of rainfall arises from the inclusion of differential reflectivity (Zdr), which is an effective estimator of drop size, and specific differential phase shift (Kdp) as additional radar parameters to supplement reflectivity factor (Seliga and Bringi, 1976, 1978; Sachidananda and Zrnic, 1986; Bringi and Chandrasekar, 2001). The polarimetric radar will also permit the development of precipitation particle type products, an important addition to the diagnostic product suite (Vivekanandan et al., 1999; Liu and Chandrasekar, 2000). Plate 1 shows an example of vertical sections of reflectivity and hydrometeor type obtained from polarimetric radar. Independent of the impact on precipitation estimates, polarization diversity capability will contribute significantly to improving data quality (Chandrasekar et al., 2002). The next generation radar system should preserve the polarimetric capability anticipated for the WSR-88D, and special attention should be given to the development of products to quantify precipitation rate and to determine precipitation type.
The present inability of the WSR-88D to identify the bright band, and the inadvertent introduction of bright-band reflectivity bias into precipitation and storm products, has been identified as a recurring data-quality problem in existing NEXRAD products. The introduction of a polarimetric capability will allow for the development of high-quality bright-band information that can be used to compensate biased rainfall estimates. Polarimetric measurements offer a significant contribution to improved radar data quality, and the polarimetric capability is important to preserve within the framework of future radar systems.
Agile Beam Techniques
Other new capabilities of future radar will likely include much faster volume scans, or Volume Control Pattern (VCP) updates. Data quality can be improved using the precise steering of phased arrays to minimize clutter echoes. Rapid beam steering combined with flexible waveforms will offer new opportunities to minimize second trip echoes and to prevent velocity aliasing. All of these new capabilities will require much higher levels of digital processing power and storage than today’s processors provide. However, these new digital processing and data storage requirements are not likely to limit the performance of the sensor systems described here. In addition, increased use of optical and wireless communications between the individual sensors and networks is envisioned.
Advanced radar technology can reduce, but not completely eliminate, data corruption due to ground clutter. Electronically scanned phased array radar has the flexibility to suppress ground clutter entering the main beam by steering the main beam immediately above clutter objects appearing on the horizon (moun-
tains and buildings) and by skipping over known interference sources (such as the rising and setting sun). Further, electronic scanning allows complete removal of beam spreading and degradation caused by beam motion during the data acquisition interval, thereby yielding improved clutter rejection of both normal and anomalous propagation-induced clutter. The inherent beam agility allows rapid steering in other directions when objectionable interference beyond the primary echo is being received. Various data impairments associated with terrain blockage, beam divergence with range, overshooting weather caused by the earth’s curvature, and atmosphere-induced beam propagation effects will continue to be present, but the overall data-quality improvement should be substantial.
PHASED ARRAY RADARS
Both the FAA Terminal Area Surveillance System (TASS) study (Rogers et al., 1997) and the European COST-75 action (Collier, 2001) concluded that multiple-face phased array radar having no moving parts would not be economically feasible using the active antenna array module technologies currently envisioned. Several thousand transmit/receive (T/R) modules per face and at least four faces for full volumetric coverage requires on the order of 10,000 elements per radar. The total component cost may be greater than $10 million per radar, which seems an unlikely expenditure in the near future. The cost for a single-frequency dual-polarization T/R module is optimistically a few hundred dollars. Adding a second wavelength sensing capability would likely raise the cost to several hundred dollars per module. Benefit-cost studies will continue, but it appears that a major new T/R module design or array configuration is the only viable way to construct a network of fixed, multifaced phased array radars for weather services.
An alternative architecture utilizing a rotating single array appears entirely satisfactory for the next generation weather radar. A simple rotating slotted waveguide array has been proposed for a one-dimensional elevation steering concept of a single agile beam (Smith, 1974; Keeler and Frush, 1983; Josefsson, 1991; Holloway and Keeler, 1993). In its simplest incarnation, the rotating array covers the azimuth region by slow mechanical scanning while the elevation region is covered by an array tilted back about 20 degrees and using standard rapid 1D electronic scan steering techniques. Plate 2 shows a conceptual schematic of such a system. The single beam may be rapidly scanned electronically in elevation using frequency or phase steering that requires on the order of only 100 phase shifters instead of several thousand. In combination with a high-resolution pulse waveform, range averaging of independent high-resolution samples permits accurate base data estimates in a short dwell time. In this manner, a large volume can be scanned in time intervals on the order of a minute with spatial resolution
comparable to the WSR-88D. This approach requires increasing the transmit bandwidth, which will likely be complicated by future spectrum allocation issues. Typically, the phased array can electronically scan up to 20 degrees off a direction normal to the antenna plane without severely altering the beam shape. As the beam steers to higher elevations to cover the “cone of silence” above the radar, two degradations occur: (1) the beam size in that dimension is increased and (2) a larger number of elements are required (at closer spacing) to suppress grating lobes. It may be advantageous to execute a second mechanical sweep with the antenna tilted to a higher elevation and accept a corresponding loss in VCP update rate.
The FAA TASS program considered various other architectures, including back-to-back 1D scanning arrays, to increase the volume scan rate. A somewhat more complex and costly architecture utilizing 1D scanning in elevation but, in addition, having limited azimuth scan-back capability may be required for certain near-surface measurement applications requiring long dwell times, such as ground clutter cancellation or clear air detection of weak atmospheric signatures. Another architecture uses a small phased array feed, possibly only a linear array that illuminates a reflector that forms the main beam or beams. Digital beam forming (DBF) is feasible with some types of array feed systems, but beam shaping and sidelobe control may be difficult.
Multiple-beam phased array systems have been proposed (Skolnik, 2001; Hansen 1988). For example, multiple transmit beams, each having relatively high sidelobes using a high-power transmit antenna, could be coupled with a precision, low-power receive phased array capable of high overall isolation between the individual beams. Frequency steering techniques or digital beam forming technology would allow the same rapid VCP coverage while using longer dwell times on each beam. The primary disadvantage of multiple-beam architectures is the increased design effort necessary to assure high isolation between the simultaneous beams. Furthermore, the transmitter must generate a correspondingly higher average power (and a greater number of multifrequency waveforms in the frequency steered case) to maintain sensitivity. The same trade-off applies for single-beam techniques using pulse compression—the average transmit power must be increased to maintain sensitivity for a given range resolution.
Design tradeoffs in these alternative architectures are somewhat different. In the pulse compression single-beam system, the range sidelobes of the compressed waveform should be minimized to maintain the contrast in strong radial reflectivity gradients. In digital beam forming and frequency steered multiple-beam configurations using simultaneous transmit beams, advanced receiver beam shape design is needed to maintain isolation of nearby beams. Precision beam shaping may be achievable using low-power pin-diode phase shifters that are not required to accommodate the high-power transmit pulse.
Phased array architectures may be based on different array technologies. The active antenna system uses individual T/R modules to integrate the transmitter, the receiver and the phase shifter with the antenna radiating elements. Examples include the Seimens-Plessey MESAR system and the Digital Terminal Area Surveillance System/Microburst Prediction Radar (DTASS/MPR) system (Protopapa et al., 1994, Katz et al., 1997). Active antenna technology has not been cost effective in the past, but future monolithic microwave integrated circuit (MMIC) designs and high volume may allow economical production for the next weather radar system.
Alternatively, a single transmitter may feed an array of radiators, such as phased array radars built by Lockheed-Martin Corporation (LMC)—the corporate feed SPY-1 radar (Maese et al., 2001) and LMC’s newer space-fed Multiple Object Tracking Radar (MOTR) system. Recently, the University of Oklahoma and the National Severe Storms Laboratory in Norman, Oklahoma have made arrangements with the Office of Naval Research (ONR) and Lockheed-Martin to develop a mechanically scanned phased array using a Navy spare SPY-1 array face (Forsyth et al., 2002). This antenna will be coupled with a WSR-88D transmitter, a modern digital receiver and signal processing system, and (eventually) a pulse compression waveform. This phased array test bed, cumbersome by present standards, will be operated to test and evaluate phased array scanning techniques that will likely employ modern, cost-effective array technologies for a future implementation.
Slotted waveguide antenna technology offers a viable technology for phased array radar steerable in one dimension. Furthermore, the COST-75 action identified the Thomson-CSF reflect-array (Beguin and Plante, 1998) as a promising new technology for phase steering a single beam. When coupled with a high-range resolution waveform, these systems allow much faster volume update rates. Furthermore, antenna beam forming has analogies in spectrum analysis that may provide new insights for high-resolution antenna beams (Palmer et al., 1999).
The likely need to have polarization diversity integrated with the phased array places special constraints. TASS studies (Rogers et al., 1997) have shown that polarimetric transmission and reception is possible using the active antenna architecture whereby each radiator is capable of dual polarization. Studies at Ericsson (Josefsson, 1991) have shown that a slotted ridge waveguide can generate dual-polarization beams. However, these same studies indicate that the polarization becomes distorted when beams are steered away from the principal axes (only in azimuth or elevation from bore sight). Compensation may be possible since these are fixed, predictable patterns. The data-quality benefit will be attained even if the polarization purity deteriorates slightly in phased array antenna implementations when the beam departs from bore sight. These effects must be carefully considered in the design of any new radar system.
Because technology continues to evolve, it is important that there be further investigation and analysis of the risks, costs, and benefits that drive an advanced radar architecture decision. This analysis could lead to a conclusion that the costs and risks of the phased array technology are acceptable for the next generation system. In that event, research will be needed into the development of the appropriate prototype radar technology and processing algorithms for phased array systems. This research is necessary risk mitigation, so that appropriate information is available concerning options and benefits of a network of phased array radars and associated sensors.
The technical characteristics, design, and costs of phased array radar systems that would provide the needed rapid scanning, while preserving important capabilities such as polarization diversity, should be established.
Pulse Compression and Scan Rate
The volume update rate is a critical factor limiting the effectiveness of many meteorological products. Pulse compression utilizes a long, low peak power waveform coded in phase, frequency, or amplitude to effectively compress the full energy of the extended pulse to a much shorter resolution interval. Pulse compression techniques (or short, high peak power pulses) may be employed to acquire independent samples in range that may be averaged to obtain statistically accurate information in a short dwell time. For many aviation and military applications, an extended waveform using a form of pulse compression is entirely acceptable. However, for distributed weather scatterers, the integrated sidelobe contamination presents a large obstacle in the strong reflectivity gradients characteristic of convective weather. Various waveform designs have been proposed to reduce this contamination using nonlinear FM and other techniques of pulse shaping to reduce the Doppler sensitivity of the compression technique (Keeler and Hwang, 1995; Mudukutore et al., 1998; Bucci and Urkowitz, 1993). As other techniques of obtaining independent samples to reduce volume updates times are developed, these obstacles will be overcome. The wind profiler processing technique of range imaging (Palmer et al., 1999) provides independent samples in range similar to pulse compression. Inverse (whitening) filter processing techniques to increase the number of independent range samples (Keeler and Griffiths, 1978; Torres and Zrnic, 2001) may also find application to radar waveforms.
In contrast to the distributed transmitters and receivers characteristic of active array technology, present microwave radars use klystrons, magnetrons, and
traveling wave tubes to generate high-power pulses prone to single-point failures. Much has been written on solid-state transmitters regarding reliability and waveform flexibility. Even high average power solid-state transmitters are known to have high reliability and extremely low phase noise characteristics, primarily because they operate from low-voltage power supplies. They may be designed with failure-tolerant modes so that an entire second back-up transmitter is not required in operational systems as is the case for transmitters based on klystrons and traveling wave tubes (TWTs). However, solid-state transmitters lack the high peak power pulsing capability that may ultimately be required of a new generation weather radar system. Pulse compression of long, low peak power waveforms is a natural resolution; however, range sidelobe contamination in high-reflectivity gradients may limit the quality of the measurements. A definitive feasibility demonstration is needed for pulse compression techniques for weather radars before low peak power solid-state transmitters can be seriously considered for the next generation radars.
A second shorter wavelength combined with an S-band system has been implemented within the research community for attenuation-based rainfall estimation and Mie scattering-based hail detection. The European community dropped this option after the COST-75 study cited potential practical problems, the major one being beam matching (Meischner, 1999). However, this technology is the one that is planned for the next space-borne radar system to be carried on the Global Precipitation Mission (GPM). In view of these conflicting pursuits, it is not clear whether dual-wavelength radar offers any additional benefit for future advanced weather radar.
Multiple Antenna Radar (Interferometric Processing)
Phased array radars offer another advantage over standard prime-focus reflector antenna systems. By separating the full array into two possibly overlapping subarrays in the receive mode and simultaneously processing the two beams, it is possible to retrieve the tangential wind component in the direction connecting the center of the two subarrays. This multiple antenna processing technique is well known in the vertical pointing wind profiler field (Briggs, 1980) and conceptually should be applicable to transverse wind measurements in an electronically scanned antenna operation using stationary beams. This processing application may require a slight azimuth scan-back to freeze the beam motion during the dwell time. By using digital beam forming techniques, all three components of the vector wind can be simultaneously measured at any radar measurement volume in space. In a vertical pointing mode, the profile of vector winds directly above the radar could be measured.
Global Positioning System (GPS)
The Global Positioning System (GPS) is a relatively new technology that continues to be exploited in atmospheric remote sensing. Several GPS-based retrievals are now available, and the suite of measurements will expand. In the context of radar networks, GPS has been used to provide a wide area coverage time and frequency synchronization between spatially separated components of various radar systems, such as the bistatic research radar systems (Wurman et al., 1993). The future network of radars will likely be synchronized by GPS. When networks of a variety of sensors must be synchronized over large distances and times, GPS offers an expeditious and inexpensive solution.
Signal processing is an integral part of any weather radar. The technology and science of signal processing is fairly advanced and is not expected to be a limiting factor in the design of next generation radar. The technical advances and affordability in commercial signal processors, field programmable gate arrays (FPGA), and general-purpose computers are likely to far outpace the progress in other areas such as in transmitter and antenna design. In conjunction with signal processors, the receiver technology is also progressing at a rapid pace. Signal digitization is moving further up the receiver chain, and it is conceivable that future digitization may be performed after the first Low Noise Amplifier (LNA) and anti-aliasing filter. All these advances will permit advanced signal processing algorithms, such as adaptive filtering and spectral processing (Keeler and Passarelli, 1990). Some of these currently are being tested with research radars (Seminario et al., 2001). In summary, technological advances in signal processor and receiver technologies are expected to meet or exceed needs of the next generation weather radar.
Adaptive Waveform Selection and Scanning
Fully adaptive 2D beam steering capability allows extreme flexibility in developing waveforms not only to enhance data quality, but also to acquire rapid update data and to allocate transmit power where it is most needed. Because of the variety of applications and users, it is important to minimize the volume scan time and to use flexible scan strategies that adapt to varying weather situations. For example, much of the physical nature of convective storms resides in their vertical structure. Accordingly, vertical cross sections may be taken by adaptive scan and signal processing along arbitrary paths. Adaptive scan is also important in accelerating the scan cycle so that time is not wasted looking at empty space or low priority targets. Such a capability is critical in rapidly evolving situations such as tornado genesis and microburst formation, where a minute or two may mean the difference between life and death.
Agile beam electronically scanned phased array radars are capable of adaptive beam forming, adaptive scanning, and adaptive waveform selection to dwell on the most important regions of the atmosphere at any particular time. For example, a low-level surveillance scan might detect a region of convergence indicative of a potential thunderstorm near an airport. The radar could select a waveform for high clear air sensitivity, a beam shape that suppresses sidelobe clutter contamination from the ground and from more intense precipitation regions, and a scan pattern that monitors that region on a frequent revisit interval. The radar data system or products of the integrated observing complex can be used to determine the radar scans and waveforms on a constantly changing basis, depending on the evolution of the weather events in the region. Thus, the radar is not purely a data source; it actively feeds back information that it measures to enhance ongoing measurements. Furthermore, adaptive processing techniques that continuously optimize the sensitivity of the radar and suppress interference will be more frequently applied as these techniques are demonstrated in research systems.
Adaptive waveform selection, which may even be applied to present systems, and agile beam scanning strategies, which require an electronically scanned phased array system, should be explored to optimize performance in diverse weather.
Tomographic processing has not been extensively applied to surface radars; however, it has extensive potential with path-integrated measurements (i.e., attenuation and differential propagation phase) to recover detailed atmospheric structures in the measurement plane (Dobaie and Chandrasekar, 1995; Srivastava and Tian, 1996; Testud and Amayenc, 1989). Tomographic processing may also yield path-integrated measurements from a single transmitter and multiple receivers, yielding a different paradigm for remote sensing of precipitation, and may find unique applications in airborne and space borne radar platforms. Nevertheless, tomographic processing is likely to remain in an auxiliary mode since it does not provide everything the current standard radar provides for all its applications.
Synthetic Aperture Processing
Synthetic aperture radar (SAR) systems have found widespread usage in space-borne applications, and the interest in using SAR systems for observing precipitation will grow continuously as additional airborne and space-borne missions are deployed. The SAR systems have the advantage of enhancing the spatial
resolution with a smaller physical antenna (Fitch, 1988; Atlas and Moore, 1987). However, there are several special demands of weather radars that make SAR applications challenging. For example, most of the SAR operation is done at large incidence angles away from nadir, and ground clutter contamination is a serious problem under these conditions; motion of hydrometeors also forces a reduction in performance. Major innovation is needed prior to utilizing SAR technology extensively for space-borne observation of precipitation.
Conventional radar data processing is tacitly directed at precipitation, and the other echoes are “clutter” that must be suppressed. Yet, radar information is considerably richer. For example, anomalous propagation ground echoes reveal changes in the vertical profiles of temperature and moisture in the lower troposphere. Specialized processing is required in order to obtain and use these ancillary sources for meteorological information.
An example of such specialized processing is the measurement of the near-surface refractive index of air using fixed ground targets (Fabry et al., 1997). As the refractive index of air in the propagation path changes, the time of travel of radar waves between the antenna and the fixed ground target varies slightly and causes a change in the measured ground target echo’s absolute phase. That variation is large enough to be accurately measured by carefully selecting and processing a large number of specific targets. If an independent set of temperature and pressure measurements are available, the refractivity can then be converted to spatial and temporal variations of moisture, or water vapor fields, in the surface boundary layer.