Remotely measuring changes in vegetation moisture is important for agricultural and environmental applications. While radiometry is sensitive to moisture, the resolution is low; optical and infrared measurements can be very fine-resolution but cannot directly measure moisture, instead relying on proxies such as brightness at wavelengths sensitive to chlorophyll. InSAR (interferometric synthetic aperture radar) produces images at fine resolution, and its measurement frequencies are sensitive to water. InSAR closure phase, a measurement of the residual phase from a circular combination of three multilooked interferograms, has been shown to relate to changing moisture content within a radar image. Here, we show that InSAR closure phase tracks several metrics of vegetation moisture from in situ measurements. In a forest in central Massachusetts, cumulative InSAR closure phase at C-band, while not correlated with soil moisture or L-band vegetation optical depth, is anticorrelated with the canopy wetness and well- correlated with the xylem dielectric constant.

We quantify and characterize an often-overlooked error source in InSAR and derived time series: aliasing from insufficiently sampled interferograms in space and time. Conditions that lead to unrecoverable aliasing, namely, when true phase gradients greater than π rad per pixel form a closed loop, result in a biased loss in unwrapped phase magnitude that is proportional to the number of high-gradient (> π) loops encircling a pixel. High-gradient interferogram loops are influenced by the radar system wavelength, its spatial resolution, imaging geometry, and the local noise level. Furthermore, spatial filtering may induce further aliasing because it decreases the spatial resolution and, consequently, the physical gradient tolerance. We show here that beyond aliasing in a single interferogram, there follows a similar loss in time series, which we demonstrate with small-baseline subset (SBAS) techniques. We find that aliasing has an intimate relationship with the time between image acquisitions as the displacement field evolves; consequently, aliasing errors often increase with increasing temporal baseline, and we observe a systematic decrease in SBAS solution magnitudes and a spatiotemporal distortion of displacement patterns with increasing maximum temporal baseline. Sentinel-1 observations of three study areas (Kilauea Volcano, the Delaware Basin, and California’s Central Valley) show that some of the best time series solutions with respect to ground-truth include long-temporal baseline interferograms and others require their explicit exclusion, suggesting there is a delicate balance between aliasing and noise reduction that not only varies between study areas, but may even be unique to each pixel.

Soil moisture can vary spatially at the scale of agricultural fields (~ 10 − 100 m), which is generally too fine to resolve using passive radiometric methods. Active radar provides an opportunity for finer resolution measurements; in particular, the interferometric synthetic aperture radar (InSAR) closure phase parameter is sensitive to changing soil moisture. We have developed a model showing that systematic non-zero closure phase can result from scattering from objects at different depths in a medium of time-varying dielectric, such as from changes in soil moisture. The model predicts that interference between surface and subsurface reflections is needed for closure phase to be non-zero. We find that, under certain circumstances, we can estimate soil moisture from closure phase using a data reduction approach that includes a cumulative sum of closure phase over time and removal of a trend. The correlation between cumulative closure phase and soil moisture suggests that the closure phase is related to the change in soil moisture. We examine a large test region in Oklahoma, where the detrended cumulative closure phase from Sentinel-1 data demonstrates some agreement within situsoil moisture measurements. In other areas, the match is weaker, implying a terrain dependence for the quality of fit. Cumulative InSAR closure phase promises to provide a valuable new method to remotely estimate soil moisture.

Differential SAR interferometry for deformation measurement and coherent change detection has revolutionized our understanding of many geophysical processes on the Earth since its first demonstration in the 1980’s. The recently selected NASA Discovery mission VERITAS aims to be the first planetary mission to use differential radar interferometric techniques for measuring surface deformation and for coherent change detection measurements. These measurements provide the most sensitive means for detecting and measuring present day geologic activity on Venus. Differential SAR interferometric measurements at Venus by the VERITAS mission has several interesting and unique features including the atmosphere and orbit considerations that are described in this paper.

InSAR (interferometric synthetic aperture radar) phase closure, the net phase from linking three multilooked interferograms formed from three acquisitions, has been linked to soil moisture [1]. Here, we show one possible way to predict soil moisture from InSAR phase closure. Our data reduction approach includes an integration of the phase closure over time and subtraction of a random walk component to relate the differential phase values to soil moisture level. We find that for a large test region of Oklahoma, the integrated phase closure using Sentinel-1 data tracks the soil moisture observed in the field. In other cases, the match is less than good. If we can determine under what circumstances these InSAR measurements provide a good match to soil moisture, we have a valuable utility to remotely estimate soil moisture at scales useful for agricultural assessments, as conventional radiometric measurements cover hundreds of adjacent fields in each resolution cell.

To produce high-resolution ultrasound images, synthetic aper-ture acquisitions are used, which contain a large amount of data that is difficult to beamform in real time with traditional methods like delay and sum (DAS). Although frequency do-main beamforming methods have been adapted to ultrasound, the existing adaptations require interpolation. The chirp scaling algorithm (CSA), which was originally developed for radar remote sensing applications, avoids interpolation by using the Fourier transform shift property, so modifying CSA for ultrasound could expedite computation while avoiding interpolation error, improving rapid image formation. This work adapts CSA theory for typical monostatic ultrasound acquisitions and tests the algorithm on point target and le-sion simulations to analyze image quality and computational speed. While CSA achieves similar image quality as related frequency domain beamformers (such as the range-Doppler algorithm - RDA), its runtime is lower than both DAS and RDA. With further adaptations, the algorithm could benefit real-time, high-resolution clinical imaging.

The Permafrost Dynamics Observatory (PDO) combines L-band interferometric synthetic aperture radar (InSAR) and P-band polarimetric synthetic aperture radar (PolSAR) to simultaneously estimate the seasonal thaw depth and soil moisture profile of the active layer in permafrost regions. L-band InSAR can measure seasonal subsidence due to thawing of the active layer and P-band PolSAR backscatter is sensitive to subsurface soil moisture. A joint retrieval scheme is developed as both subsidence and soil moisture are essential to accurate active layer thickness (ALT) estimation. The PDO joint retrieval has been applied to airborne L- and P-band SAR data acquired over Arctic-boreal region during the 2017 Arctic-Boreal Vulnerability Experiment (ABoVE) airborne campaign. In this paper, we describe the forward models and joint inversion used in the PDO retrievals and compare the results with in-situ ALT and soil moisture data estimated from ground-penetrating radar (GPR).

It is well-known that the backscatter of high-resolution Synthetic Aperture Radar (SAR) imagery is non-Gaussian in nature. As a result, corresponding heavy-tailed models have been successfully incorporated for the design of improved SAR target detectors. However, Gaussian-based detectors are largely still applied for selection of persistent scatterers (PS) in Interferometric Synthetic Aperture Radar (InSAR) imagery, and implications for the performance of PS techniques have not been well-studied. Here, we extend an existing Gaussian model for PS to incorporate non-Gaussian behavior. We then implement the model for PS detection and compare its performance to its Gaussian counterpart, finding that the non-Gaussian model finds a slightly denser network of PS. Further work will focus on analyzing the characteristics of this disparity, including its relationship with terrain and system parameters such as wavelength and bandwidth, and compare the estimated deformation from the non-Gaussian detector compared to an existing Gaussian-based model. Understanding the limitations of Gaussian models will inform the design of improved PS detectors to produce more complete deformation maps and enable the broader application of InSAR for challenging applications, such as observing small strain rates in natural terrain.

Radar imaging using the cross-correlation of receiver signals possesses several benefits over traditional matched-filter techniques that make it more suitable for imaging fast-moving objects without well-defined flight tracks, such as fast-moving space debris. Correlation-based imaging is implemented with a network of receivers, and requires no knowledge of the imaging pulse or the emitter and receiver locations and is able to compensate for unknown linear and rotational motion with a resolution on the order of the imaging wavelength, comparable to the performance of matched-filter imaging. Here, we compare the performance of a simplified bistatic configuration for correlation-based imaging and the traditional matched-filter method in a representative case study experiment, where a small aircraft is used as a target. We find that the bistatic case is effective even when there are motions that exceed the transmitter wavelength and is more resilient to unaccounted random motion than the monostatic case. However, as expected for a two-receiver configuration, the resolution is lower than the matched-filtering method even in the ideal case. Finally, we comment on extensions as well as implications for multistatic configurations.

Frequency-domain beamforming has become increasingly popular for fast processing of large synthetic aperture data in medical ultrasound. Here, we modify the Range Doppler Algorithm (RDA) to focus ultrasound signals from multistatic acquisitions. RDA, which was first proposed for fast beamforming of monostatic data in radar remote sensing, is suitable for fast processing of large datasets because all operations are done in one dimension at a time, allowing for an efficient and intuitive implementation. We demonstrate through simulation that multistatic RDA achieves similar image quality as traditionally used, multistatic delay-and-sum (DAS), while increasing the reconstruction speed by approximately a factor of three. We also show that the RDA and DAS images from the multistatic acquisition show reduced sidelobe levels compared to their counterparts from the monostatic acquisition. Demonstrated version of the multistatic RDA might be applicable beyond ultrasound medical imaging, such as for processing of synthetic aperture radar (SAR) data from satellite constellations.

Interferometric synthetic aperture radar (InSAR) is a well-known imaging technique used by geophysicists to measure deformation of the Earth's surface over time. Accurate measurements of these displacements are crucial to correctly interpret subsurface processes. We show that when a closed loop of displacement gradients exceeding π radians occurs in a wrapped interferogram, the unwrapped solution will systematically underestimate the total displacement due to aliasing, even in the absence of noise. Common InSAR practices such as spatial filtering and averaging decrease resolution and thus increase the risk of aliasing. For time-varying processes, interferograms formed over longer time spans are more likely to under-sample the true displacement. Synthetic and real time series analyses using small baseline and subset (SBAS) [1] techniques show that the inclusion of aliased interferograms results in systematic and significant errors. These examples suggest that adaptive filters and varying the subset of interferograms used in SBAS throughout the scene will be important improvements to InSAR analyses.

The joint NASA/ISRO SAR (NISAR) satellite mission is anticipated to provide routine L-band coverage of most of the Earth's land surface every 12-days for both ascending and descending orbits. In terms of impact on solid earth science (SES), the primary measurement will be Interferometric SAR (InSAR) observations of ground deformation in two satellite line-of-sight (LOS) directions. Key observation characteristics include acquisitions with small interferometric baselines to maximize interferometric coherence and decrease sensitivity to topography, wide bandwidth allowing for split-band processing to model out the impacts of the ionosphere, and joint L- and S-band observations in selected regions. We describe here the key measurement requirements for solid earth science, as well as our approach to validating these requirements once the mission is underway.

InSAR time series analysis is a technique for observing changes in range or phase of a radar resolution element from a sequence of radar acquisitions. General methods of retrieving temporal signals, such as small baseline subset analysis (SBAS) [1], allow for redundant measurement in order to reduce statistical variations from system noises. Recent work using significant redundancy shows that spatial and temporal aliasing leads to additional errors in measurement [2]. Here we show that it is necessary to optimally balance the amount of redundancy with he amount of aliasing to obtain the highest quality estimates of deformation, and further that a spatially variable level of averaging interferograms is needed for many analyses. Development of a practical algorithm remains before we are operationally able to fully exploit the capabilities of an InSAR system.

Wide-swath radar imaging requires that the time interval to collect each radar pulse echo is large and can often exceed the interpulse period. As it is difficult to both transmit and receive from the same antenna simultaneously, there will be “blind ranges” when the receive and transmit times overlap. This leads to gaps in the radar echo and thus degradation of system performance. Today, most wide-swath systems address this by segmenting the swath in range, such as in the ScanSAR or Terrain Observation with Progressive Scan (TOPS) mode operation, so that each subswath is within the range ambiguity limit. Some groups have started experimenting with the variations in the sweepSAR technology, in which the receive antenna tracks the radar echo across the swath, realizing a complete wide-swath range scan but leading to the echo gaps. Here, we look at minimizing the effect of blind ranges by varying the radar pulse-repetition frequency (PRF) and interpolating across the gaps to preserve azimuth signal continuity. If the pulse times are selected properly, the main effect is to raise the noise floor of the echoes. Geocoded magnitude images, interferograms and correlation images, and deformation time series show strong robustness with respect to dithering, demonstrating that choosing essentially random PRFs allows for accurate generation of SAR and Interferometric Synthetic Aperture Radar (InSAR) data products while retaining wide-swath, fine-resolution coverage.

The accuracy of geophysical parameter estimation made with interferometric synthetic aperture radar (InSAR) time-series techniques can be improved with rapidly increasing available data volumes and with the development of noise covariance matrices applicable to joint analysis of networks of interferograms. In this article, we present a new decorrelation phase covariance model and discuss its role in noise reduction in unwrapped interferometric phase stacks. We demonstrate with an example in which we average unwrapped interferogram phase stacks that span over a transient event how a noise covariance model can aid in noise reduction. Our model suggests that, for rapidly decorrelating surfaces (i.e., surfaces with much shorter correlation time than SAR acquisition intervals), it is preferable to incorporate all available interferograms from long observation windows. For slowly decorrelating surfaces (i.e., surfaces with longer correlation time than SAR acquisition intervals), our model suggests that a small subset of interferometric pairs is sufficient. We validate our model and three existing models of decorrelation phase covariance matrices in both Cascadia, a region with heavy vegetation cover, and Death Valley, a desert region with C-band Sentinel-1 A observations. Our proposed model matches observations with the smallest average discrepancy between theory and observations.

Seasonal subsidence measured by repeat-pass interferometric synthetic aperture radar (InSAR) can be used to infer the active layer thickness (ALT) in permafrost regions. The differential volume of soil water undergoing the phase change over the thaw season is one of the factors impacting the seasonal subsidence and is a function of both soil moisture profile and thaw depth. Without the information about soil moisture, this InSAR approach can have large biases in the ALT estimates when soil moisture profile is below saturation. Soil moisture and ALT can also be estimated from polarimetric synthetic aperture radar (PolSAR) backscatter observations but the sensing depth of the PolSAR approach is limited when deep ALT is present. In this paper, we integrated these two approaches and applied a joint retrieval method to estimate the soil moisture profiles and ALT from the L-band InSAR and P-band PolSAR data acquired over the Arctic-boreal region during the 2017 Arctic-Boreal Vulnerability Experiment (ABoVE) airborne campaign.

With the increasing availability of spaceborne Synthetic Aperture Radar (SAR) imagery, Interferometric Synthetic Aperture Radar (InSAR) has become a prominent technique to study geophysical phenomena. However, its application can be limited in natural terrain due to rapid changes in the surface known as decorrelation. Persistent scatterers (PS) are stable elements that can be exploited in InSAR imagery to allow fine deformation mapping of the Earth's surface even in areas that are highly decorrelated. Despite the widespread use of PS techniques, little research has been dedicated to studying the relationship in between key system parameters and observed PS statistics. Such knowledge is critical in creating effective detection algorithms and informing design choices for satellite missions. Here, we present an analytical expression for the probability density function (PDF) of PS density with respect to system wavelength and bandwidth. This equation could be directly implemented in traditional detection algorithms for PS detection. We offer comparisons with real data, describing how the model could be improved. Further work will extend this expression to the probability of PS incidence and examine performance over different types of terrain.

Phase inconsistencies, or closure phase, in interferometric synthetic aperture radar (InSAR) images are associated with the lack of phase closure for any given triplet of SAR scenes. While nonzero phase closure is fundamentally due to signal decorrelation between SAR scenes within a triplet, there has been particular interest in relating closure phase to temporal variations in surface scattering properties, namely surface and near-subsurface soil moisture state. In this manuscript, we will provide a brief overview of closure phase, and propose a methodology for retrieving soil moisture information from closure phase observations. We will briefly discuss implementation of this methodology, and then discuss the degree to which the assumed interferometric soil moisture model, and the assumed statistics of suface scatterers impacts the sensitivity of closure phase observations to surface soil moisture.

We use Sentinel-1 A/B InSAR to measure deformation rates of -2 to 2.5 cm/yr in the Delaware Basin from fluid injection and extraction relating to oil and gas activity. Long-wavelength atmospheric effects make it challenging to obtain reliable time series, which are necessary to study the time-dependent deformation. We show that high-pass filters isolate the short-wavelength deformation signals from broad-scale atmospheric contributions. Future work will explore automatic ways to choose the most-suitable filter, which may be applied to other tectonically-stable regions.

Wide swath radar imaging implies that the time interval to collect each radar pulse echo is large, and can exceed the interpulse period. As it is difficult to both transmit and receive from the same antenna simultaneously, there will be “blind ranges” when the receive and transmit times overlap. Today most wide swath systems address this by segmenting the swath in range, such as in ScanSAR or TOPS mode operation. Some groups have started experimenting with variations of sweepSAR technology, in which the receive antenna tracks the radar echo across the swath. Here we look at minimizing the effect of blind ranges by varying the radar PRF. If the pulse times are selected randomly, the main effect is to raise the noise floor of the echoes. Geocoded magnitude images, interferograms, and correlation images show strong robustness with respect to dithering, demonstrating that choosing essentially random PRFs allows for accurate generation of SAR and InSAR data products.

Modern spaceborne radar systems today achieve near-global coverage at moderate resolution with weekly or better repeat intervals. Consideration of adequate temporal sampling and the need to average images to reduce noise suggests that reducing radar revisit times can enable heretofore impossible applications. We propose that the aspirational goal of fine-resolution repeat coverage every 6 hours would be a revolutionary advance in radar and InSAR capability and utility, and should be a guidepost for new system design. Here we review the science and user needs for such a system, and show that current technologies, with some evolutionary extensions can enable such a system.

Here we present a physics-based decorrelation phase covariance model and discuss its role in effective decorrelation noise reduction in interferogram stacks. We test our model in both Cascadia - a rapidly decorrelating region, and Death Valley - a slowly decorrelating region, with observations collected by Sentinel-1. We find that in Cascadia, including redundant interferograms in the stack reduces phase variance from 0.28 rad2to 0.04 rad2, while in Death Valley, both redundant and independent interferogram stacking yield phase variances of 0.10 rad2. Both observations are consistent with predictions from our model. Comparing with three existing decorrelation phase covariance models, our proposed model matches observations with the smallest average discrepancy between theory and observations - 0.017 rad2in Cascadia and 0.066 rad2in Death Valley.

In InSAR time series analysis, persistent scatterers (PS) are pixels that remain temporally correlated even in areas of otherwise high decorrelation. Using these stable pixels as reference points, we can analyze deformation signals on the Earth's surface down to the millimeter scale. PS possess different physical and statistical properties compared to the background, and analyzing the statistical behavior of PS can help to elucidate radar backscattering behavior from different terrain. Here, we propose a new statistical characterization of SAR image power using a mixture distribution of PS and non-PS points, i.e. clutter. We find that a lognormal distribution fits both pixel types, validated by a scattering model that is a function of bandwidth. We present here from C-band RADARSAT-2 data, and will compare these with additional C-band Sentinel data and L-band data from ALOS-2.

1. The ionospheric correction can be applied by processing the geocoded single look complex (SLC) product in two subbands using the frequency dispersion correction algorithm. 2. The effective number of looks for correlation estimation is obtained by multiplying the number of spatial looks by the cosine of the incidence angle. 3. A DEM error will result in an extra phase term that is dependent on baseline, but for modern radar sensors this term is very small. 4. The same as you currently do using range-Doppler products.

We propose a novel method for quantifying and correcting phase errors in interferometric synthetic aperture radar (InSAR) data associated with signal decorrelation. This proposed method relates the observed phase nonclosure (referred to as the closure phase) of triplet combinations of any three individual SAR scenes to the decorrelative phase signal in individual interferograms (pairs of SAR scenes). A singular value decomposition (SVD) method is applied to solve the minimum-norm least-squares best fitting estimate of the decorrelation phase for any arbitrary collection of SAR images. This decorrelative phase is then removed from individual interferograms; these corrected interferograms can then be used with existing InSAR time-series analysis algorithms. We demonstrate this method on the Advanced Land Observation Satellite Phased Array type L-band Synthetic Aperture Radar (ALOS PALSAR) scenes of a groundwater pumping subsidence feature in the Central Valley of California and briefly discuss potential future applications of this algorithm to study a variety of environmental and surface physical processes that contribute to InSAR signal decorrelation.

InSAR time series analysis of deformation fields, a goal of many modern radar systems, almost always starts with processed range-Doppler single look complex (SLC) radar images. These images are precisely coregistered and pairwise phase differences form the interferograms. Computing coregistered SLCs directly in a desired geometry saves significant data storage and readily produces data products that are easy to analyze in coordinate systems that are well matched to a desired application. Here we demonstrate InSAR data products created using a backprojection algorithm that skips the range- Doppler, coregistration, and elevation compensation steps. Implementing this in an on-demand system saves not only on the volume of products that must be computed, stored, and delivered, it further greatly simplifies analysis by the end user and opens up InSAR techniques to a wider group of applications.

InSAR time series analysis of deformation fields requires a set of single look complex (SLC) radar images whose pixels accurately represent the InSAR phase. In some surface cover environments, such as vegetated regions where plant growth produces echoes unrelated to the underlying surface, reliable phase estimates are difficult to identify. Persistent scattering (PS) methods sort for the most stable points in an image, so that the deformation estimates are as robust as possible. To date the often-used selection algorithms have not always fully incorporated knowledge of PS point distributions or how they vary by terrain type. Here we show estimates of the probability density functions (pdfs) for several types of surface cover and how the added information can be applied to use MAP rather than MLE estimation to improve selection, and also give insights to the radar scattering processes that give rise to PS pixels.

Interferometric synthetic aperture radar (InSAR) methods provide high-resolution maps of surface deformation applicable to many scientific, engineering, and management studies. Despite its utility, the specialized skills and computer resources required for InSAR analysis remain as barriers for truly widespread use of the technique. Reduction of radar scenes to maps of temporal deformation evolution requires not only detailed metadata describing the exact radar and surface acquisition geometries, but also a software package that can combine these for the specific scenes of interest. Furthermore, the range-Doppler reference frame and radar coordinate system itself are confusing, so that many users find it hard to incorporate even useful products in their customary analyses. Finally, the sheer data volume needed for interferogram time series makes InSAR analysis challenging for many analysis systems. We show here that it is possible to deliver radar data products to users that address all of these difficulties, so that the data acquired by large, modern satellite systems are ready to use in more natural coordinates, without requiring further processing, and in as small volume as possible.

Interferometric Synthetic Aperture Radar (InSAR) methods provide high resolution maps of surface deformation applicable to many scientific, engineering and management studies. Modern spaceborne satellites provide long sequences of observations that we can reduce to many interferograms, which in turn provide the deformation histories of many points on the surface. Here we show how raw radar data, or partially processed single look complex (SLC) images may be precorrected for imaging geometry so that formation of the hundreds of interferograms from an observation sequence is both reliable and efficient. Our approach is to use motion compensation to precisely coregister the images and a common master orbit to compensate for the topographic phase terms so that simple cross multiplication yields the needed interferograms.

Interferometric synthetic aperture radar (InSAR) is a valuable tool for the study of geophysical phenomena such as crustal deformation, ice motion and structure, and vegetation canopy depths, but it is adversely affected by uncharacterized inhomogeneities in ionospheric propagation delay. Ionospheric disturbances distort both InSAR phase and correlation maps. Here, we present a method to compensate ionospheric propagation variations using accurate image coregistration. This significantly improves both the interferometric coherence and phase accuracy. An azimuth gradient in the total electron content (TEC) from a spatially variable ionosphere results in a range-dependent azimuth phase gradient being added to the phase histories of the pixels being imaged. These phase gradients are equivalent to Doppler shifts, and thus they cause azimuth offsets between the actual and imaged positions of the pixels. Measuring these offsets accurately permits estimation of the gradient and correction of the interferograms for much of the phase distortion, resulting in more accurate estimates of coherence. We show an example over Greenland where the TEC variation causes the correlation to drop from about 0.7 to about 0.2 in one region if spatially varying offsets are not accounted for; it also adds an estimated 4.4 radians of interferometric phase over an 80 km InSAR scene. After applying our algorithm, we find that the correlation in regions affected by the ionospheric inhomogeneity becomes comparable to correlation in the rest of the image. In a more challenging example over Iceland, we show that our method improves the correlation from 0.15 to 0.25 in some areas.

Interferometric synthetic aperture radar (InSAR) is an effective tool for measuring temporal changes in the Earth's surface and producing high accuracy, wide coverage images of crustal deformation fields. Decorrelation due to spatial and temporal baseline is a major limiting factor in estimating the deformation signal, but may be ameliorated by using persistent scatterer (PS) techniques. Phase unwrapping and subsequent deformation estimation on the spatially sparse PS network depend largely on the accurate selection of PS pixels and the density of the network. Many additional pixels can be added to the PS list if we are able to identify those in which a dominant scatterer exhibits partial, rather than complete, correlation across all radar scenes. In this work, we discuss and compare statistical methods to model, characterize, and select partially correlated PS pixels.

Phase artifacts in interferometric synthetic aperture radar (InSAR) images frequently degrade the interpretability of the phase and correlation signatures of terrain. Often, these distortions are attributed to spatially variable ionospheric propagation delays at two different SAR acquisition times. We present here L-band InSAR data from Iceland, California, and Hawaii. The California and Hawaii interferograms show no significant ionospheric artifacts, while the Iceland interferogram shows a maximum misregistration of three pixels in the azimuth direction, which leads to severe phase decorrelation artifacts in the InSAR image. We relate the misregistration of complex pixels seen in the interferograms to the gradient of the ionospheric total electron content (TEC) observed by global positioning system (GPS) data and confirm that indeed the phase artifacts in the Iceland interferogram are due to dispersive ionospheric propagation rather than other decorrelation factors such as neutral atmospheric delays. We develop a method to measure the spatial TEC variation at synthetic aperture length scales using dual-frequency GPS carrier phase data. We solve for the GPS data ambiguities using a low-resolution ionosphere reference derived from either available ionospheric observations or the GPS carrier phase data themselves. GPS observations show directly the level of ionospheric variability, and the spatial TEC gradient as observed by GPS predicts the misregistration of complex pixels in interferograms in all three areas. This confirmation of the cause of the image artifacts suggests that they can be routinely corrected from the InSAR data alone, provided that the sensor measures the change in TEC along the radar swath.

The InSAR Scientific Computing Environment (ISCE) is a flexible, extensible software tool designed for the end-to-end processing and analysis of synthetic aperture radar data. ISCE inherits the core of the ROIPAC interferometric tool, but contains improvements at all levels of the radar processing chain, including a modular and extensible architecture, new focusing approach, better geocoding of the data, handling of multi-polarization data, radiometric calibration, and estimation and correction of ionospheric effects. In this paper we describe the characteristics of ISCE with emphasis on the ionospheric modules. To detect ionospheric anomalies, ISCE implements the Faraday rotation method using quad-polarimetric images, and the split-spectrum technique using interferometric single-, dual- and quad-polarimetric images. The ability to generate co-registered time series of quad-polarimetric images makes ISCE also an ideal tool to be used for polarimetric-interferometric radar applications.

Time-series interferometric synthetic aperture radar (InSAR) methods estimate the spatiotemporal evolution of deformation over large areas by incorporating information from multiple SAR interferograms. Persistent scatterer (PS) and small baseline (SB) methods, which identify areas where the surface is least affected by geometric and temporal decorrelation, represent two families of time-series InSAR techniques to study successfully a wide spectrum of ground deformation phenomena worldwide. However, little is known comparatively about the performance of PS and SB techniques applied to the same region. Here, we compare quantitatively and cross validate the time-series InSAR results generated using two representative algorithms-the maximum likelihood PS method and the small baseline subset algorithm-in selected test sites, over the San Francisco Bay Area imaged by European Remote Sensing (ERS) sensors during 1995-2000. We present line of sight (LOS) velocities and deformation time series using both techniques and show that the root mean squared differences of the estimated mean velocities and deformation from each method are about 1 mm/year and 5 mm, respectively. These values are within expected noise levels and a characteristic of the pixel selection parameters for both the time-series techniques. We validate our deformation estimates against creep measurements from alignment arrays along the Hayward Fault and show that our estimates agree to within 0.5 mm/year LOS velocity and 1.5 mm LOS displacement.

In 2005, the planet Mars was illuminated 56 times for 20 min each by an unmodulated 75-cm wavelength circularly polarized wave transmitted from the SRI International 46-m antenna in the Stanford foothills. The direct signal and a Martian surface echo, separated by differential Doppler frequency shifts, were received simultaneously at the Mars Odyssey orbiter. The surface echoes exhibit both fluctuating amplitude and varying spectral width, which are responses to surface reflectivity and roughness variations along the surface track, described by a surface scattering function. We analyze the echo data using quasi-specular scattering theory and exploit high-resolution Mars orbiter laser altimeter topographic maps to model the scattering surface in 3-D at positions along the specular track of the echo, assuming a two-scale classical scattering model. We solve for a surface scattering function via regularized matrix inversion using the measured data and our surface scattering model. Optimizing the matrix inversion regularization over the course of each experiment requires consideration of the changing properties of the surface echoes caused by different viewing geometries. At one extreme the echoes appear as a strong narrowband signal, but on the other extreme the echo signals are weak and have a wide bandwidth. In this paper, we describe a scheme for determining a set of regularization constants that result in physically plausible scattering functions over the full range of echo types.

Satellite synthetic aperture radar interferometry (InSAR) is an invaluable tool for land displacement monitoring. Improved access to time series of satellite data has led to the development of several innovative multitemporal algorithms. Small baseline (SB) is one such time-series InSAR method, based on combining and inverting a set of unwrapped interferograms for surface displacement. Two-dimensional unwrapping of sparse data sets is a challenging task, and unwrapping errors can lead to incorrectly estimated deformation time series. It is well known that L1-norm is more robust than L2-norm cost function minimization if the data set has a large number of outlying points. In this paper, we present an L1-norm-based SB method using an iteratively reweighted least squares algorithm. We show that the displacement phase of both synthetic data, as well as a real data set that covers the San Francisco Bay area, is recovered more accurately than with L2-norm solutions.

We present a new interferometric synthetic aperture radar (InSAR) processing approach that capitalizes on the precise orbit tracking that is available with modern radar satellites. Our method uses an accurate orbit information along with motion-compensation techniques to propagate the radar echoes to positions along a noninertial virtual orbit frame in which the location and focusing equations are particularly simple, so that images are focused without requiring autofocus techniques and are computed efficiently. Motion compensation requires two additional focus correction phase terms that are implemented in the frequency domain. If the images from an interferometric pair or stack are all computed along the same reference orbit, flat-Earth topographic correction is not needed, and image coregistration is simplified, obviating many difficulties that are often encountered in InSAR processing. We process several data sets collected by the ALOS PALSAR instrument and find that the geodetic accuracy of the radar images is 10-20 m, with up to 20 m of additional image distortion needed to align 100 km × 100 km scenes with reference digital elevation models. We validated the accuracy by using both known radar corner reflector locations and by the registration of the interferograms with digital maps. The topography-corrected interferograms are free from all geometric phase terms, and they clearly show the geophysical observables of crustal deformation, atmospheric phase, and ionospheric phase.

Phase unwrapping is usually defined as the reconstruction of a function sampled on a spatial grid given its value modulo 2pi. Phase unwrapping is a key step in image reconstruction in many imaging techniques including interferometric synthetic aperture radar (InSAR). In recent years, many new methods have been developed to exploit the presence of coherent or persistent scattering points for extracting deformation signatures in regions where the conventional InSAR fails. These techniques often yield measurements that are only poorly sampled spatially, yet these sparse data must still be unwrapped if we are to be able to extract useful geophysical information. The conventional well-sampled 2-D phase unwrapping problem based on phase residues is fairly well understood, and many novel techniques involving geometry and network flow concepts have been implemented successfully to date. For sparse data, residues may be computed over the Delaunay triangulation of the data points, but published algorithms meet with limited success when the sparse data are unwrapped. The advantages of modern unwrapping methods applicable to well-sampled data are often lost when sparse data are analyzed. In this letter, we show that a nearest neighbor interpolation scheme allows powerful and existing 2-D solvers to be applied to sparse data. We present results using both simulated and real data sets to illustrate our method.

Phase unwrapping is usually defined as the reconstruction of a function sampled on a spatial grid given its value modulo 2pi. Phase unwrapping is a key step in image reconstruction in many imaging techniques including interferometric synthetic aperture radar (InSAR). In recent years, many new methods have been developed to exploit the presence of coherent or persistent scattering points for extracting deformation signatures in regions where conventional InSAR fails. These techniques often yield measurements that are only poorly sampled spatially, yet these sparse data must still be unwrapped if we are to be able to extract useful geophysical information. The conventional well-sampled 2-D phase unwrapping problem based on phase residues is fairly well understood and many novel techniques involving geometry and network flow concepts have been implemented successfully to date. For sparse data, residues may be computed over the Delaunay triangulation of the data points, but published algorithms meet with limited success when the sparse data are unwrapped. The advantages of modern unwrapping methods applicable to well-sampled data are often lost when sparse data are analyzed. In this letter, we show that a nearest neighbor interpolation scheme allows powerful and existing 2-D solvers to be applied to sparse data. We present results using both simulated and real data sets to illustrate our method.

The Cassini RADAR is a multimode instrument used to map the surface of Titan, the atmosphere of Saturn, the Saturn ring system, and to explore the properties of the icy satellites. Four different active mode bandwidths and a passive radiometer mode provide a wide range of flexibility in taking measurements. The scatterometer mode is used for real aperture imaging of Titan, high-altitude (around 20 000 km) synthetic aperture imaging of Titan and Iapetus, and long range (up to 700 000 km) detection of disk integrated albedos for satellites in the Saturn system. Two SAR modes are used for high- and medium-resolution (300-1000 m) imaging of Titan's surface during close flybys. A high-bandwidth altimeter mode is used for topographic profiling in selected areas with a range resolution of about 35 m. The passive radiometer mode is used to map emission from Titan, from Saturn's atmosphere, from the rings, and from the icy satellites. Repeated scans with differing polarizations using both active and passive data provide data that can usefully constrain models of surface composition and structure. The radar and radiometer receivers show very good stability, and calibration observations have provided an absolute calibration good to about 1.3 dB. Relative uncertainties within a pass and between passes can be even smaller. Data are currently being processed and delivered to the planetary data system at quarterly intervals one year after being acquired.

The National Research Council Earth Science Decadal Survey, Earth Science Applications from Space, recommends that DESDynI (Deformation, Ecosystem Structure, and Dynamics of Ice), an integrated L-band InSAR and multibeam Lidar mission, launch in the 2010-2013 timeframe. The mission will measure surface deformation for solid Earth and cryosphere objectives and vegetation structure for understanding the carbon cycle. InSAR has been used to study surface deformation of the solid Earth and cryosphere and more recently vegetation structure for estimates of biomass and ecosystem function. Lidar directly measures topography and vegetation structure and is used to estimate biomass and detect changes in surface elevation. The goal of DESDynI is to take advantage of the spatial continuity of InSAR and the precision and directness of Lidar. There are several issues related to the design of the DESDynI mission, including combining the two instruments into a single platform, optimizing the coverage and orbit for the two techniques, and carrying out the science modeling to define and maximize the scientific output of the mission.

Interferometric synthetic aperture radar (InSAR) images of geophysical events such as preeruptive volcano deformation or interseismic strain accumulation are often limited by phase distortions from the superimposed atmospheric signature. Additionally, the approximate monthly repeat cycle of many radar satellites cannot accurately capture rapidly time-varying processes. The Scanning Synthetic Aperture Radar (ScanSAR) mode of the ENVISAT/ASAR instrument permits more frequent revisits of a given area, potentially overcoming both of these limitations. In particular, stripmap mode-to-ScanSAR images provide a denser time series of interferograms than is possible with conventional stripmap-to-stripmap mode InSAR. We present images of ENVISAT/ASAR data acquired over Hawaii in which data acquired roughly weekly in ScanSAR mode are combined with ENVISAT/ASAR conventional stripmap mode data to form interferograms at a much denser temporal spacing. The burst nature of ScanSAR data requires a new processing method to form the interferograms. We use traditional matched filtering for the range compression. For the azimuth processing, we compute the stripmap mode data on the ScanSAR sampling grid using a variation, consisting of different reference functions, of Lanari's modified SPECAN algorithm that is itself an adaptation of the chirp$z$-transform to readjust the azimuth pulse spacing. The resulting interferograms faithfully reflect the phase of conventional interferograms, but exhibit fewer looks and coarser resolution than those produced by fully stripmap mode data. For many problems, temporal density of the deformation observations is paramount, and the time series analysis and temporal averaging that were made possible using ScanSAR interferograms far outweigh the loss in looks and resolution.

Snow accumulation in remote regions, such as Greenland and Antarctica, is a key factor for estimating the Earth's ice mass balance.In situdata are sparse; hence, they are useful to derive snow accumulation from remote sensing observations, such as microwave thermal emission and radar brightness. These data are usually interpreted using electromagnetic models in which volume scattering is the dominant mechanism. The main limitation of this approach is that microwave brightness is not well related to backscatter if the ice sheet is layered. Because larger grain size and thicker annual layers both increase radar image brightness, with the first corresponding to lower accumulation rate and the second to higher accumulation rate, models of radar brightness alone cannot accurately reflect accumulation. Consideration of correlation measurements can also resolve this ambiguity. We introduce an interferometric ice scattering model that relates the interferometric synthetic aperture radar correlation and radar brightness to both ice grain size and hoar layer spacing in the dry-snow zone of Greenland. We use this model and the European Remote Sensing satellite radar observations to derive several parameters related to snow accumulation rates in a small area in the dry-snow zone. These parameters show agreement with fourin situcore accumulation rate measurements in this area, whereas models using only radar brightness data do not match the observed variation in accumulation rates

Interferometric synthetic aperture radar (InSAR) correlation, a measure of the similarity of two radar echoes, provides a quantitative measure of surface and subsurface scattering properties and hence surface composition and structure. Correlation is observed by comparing the radar return across several nearby radar image pixels, but estimates of correlation are biased by finite data sample size and any underlying interferometer fringe pattern. We present a method for correcting bias in InSAR correlation measurements resulting in significantly more accurate estimates, so that inverse models of surface properties are more useful. We demonstrate the value of the approach using data collected over Antarctica by the Radarsat spacecraft.

High-resolution digital elevation models (DEMs) are often limited in spatial coverage; they also may possess systematic artifacts when compared to comprehensive low-resolution maps. Here we correct artifacts and interpolate regions of missing data in airborne Topographic Synthetic Aperture Radar (TOPSAR) DEMs using a low-resolution Shuttle Radar Topography Mission (SRTM) DEM. We use a prediction error (PE) filter to interpolate and fill missing data so that the interpolated regions have the same spectral content as the valid regions of the TOPSAR DEM. The SRTM DEM is used as an additional constraint in the interpolation. We use cross-validation methods to obtain the optimal weighting for the PE filter and SRTM DEM constraints.

This paper presents a design for a new train control system, which we call the radar system for train tracking and control (RSTTC). The RSTTC is based on the communication-based train control (CBTC) system, and uses radar technology with a spread-spectrum scheme. The main advantage to using the CBTC system is that it allows (1) the headway between trains to be adjustable and (2) the train traffic to be flexible. Radar is employed to identify both train position and the speed necessary for train control. The spread-spectrum technique helps prevent the system from being disturbed by noise and interference on the communication channel. A combination of code-division multiple access (CDMA) and time-division multiple access (TDMA) is used for tracking and controlling multiple trains on a single communication channel.

Two-dimensional (2-D) phase unwrapping is a key step in the analysis of interferometric synthetic aperture radar (InSAR) data. While challenging even in the best of circumstances, this problem poses unique difficulties when the dimensions of the interferometric input data exceed the limits of one's computational capabilities. In order to deal with such cases, we propose a technique for applying the statistical-cost, network-flow phase-unwrapping algorithm (SNAPHU) of Chen and Zebker (2001) to large datasets. Specifically, we introduce a methodology whereby a large interferogram is partitioned into a set of several smaller tiles that are unwrapped individually and then divided further into independent, irregularly shaped reliable regions. These regions are subsequently assembled into a full unwrapped solution, with the phase offsets between regions computed in a secondary optimization problem whose objective is to maximize the a posteriori probability of the final solution. As this secondary problem assumes the same statistical models as employed in the initial tile-unwrapping stage, the technique results in a solution that approximates the solution that would have been obtained had the full-size interferogram been unwrapped as a single piece. The secondary problem is framed in terms of network-flow ideas, allowing the use of an existing nonlinear solver. Applying the algorithm to a large topographic interferogram acquired over central Alaska, we find that the technique is less prone to unwrapping artifacts than more simple tiling approaches.

Radar interferometry provides a novel way to study the subsurface of glaciers: interferometric correlation. Since the two complex images that comprise the interferogram have slightly different incidence angles at each point on the ground, a decorrelating phase noise, with statistics related to the scattering medium, is present in the interferogram. The amount of "surface" decorrelation is increased by volume scatter. The larger the vertical extent of the scatterers contributing to the radar echo, the greater the decorrelation will be. By modeling this effect, the authors can estimate radio wave penetration depths within the upper layers of the glacier or ice sheet. Observations of the Greenland Ice Sheet using ERS data yield penetration depths (one-way, 1/e point for power) that range from 12 to 35 m. Due to the contribution of volume scatter, the critical interferometer baseline is decreased, and the authors find for the Greenland data, the baseline must be restricted to be less than 300 m. The authors also compare penetration depths measured within the dry snow zone with those found in the percolation zone and coastal areas. They find that as expected, the rocky coastal areas evidence minimal penetration. Interestingly, the penetration depths that the authors measure in the percolation zone, /spl sim/23 m, indicate a large degree of volume scatter, which is contrary to earlier results that found the scattering in the percolation zone dominated by structures in the first few meters. This discrepancy may be due to unmodeled scattering behavior, or the radar return may indeed include significant contributions from scatterers far beneath the surface.

We have evaluated the accuracy of digital elevation models (DEM's) generated by the JPWNASA TOPSAR synthetic aperture radar interferometer instrument by acquiring topographic radar data in the summer of 1992 over the National Training Center, near Ft. Irwin, California, and comparing the measurements to a very accurate digital elevation model derived for this area by the US. Army Topographic Engineering Center (TEC). Fiducial corner reflectors were deployed in the area, and their locations were determined to cm accuracy by the Defense Mapping Agency (DMA). DEM’s generated from the acquired radar data were rotated and translated to precisely overlay the reference DEM, allowing an analysis of the achieved height accuracy. We present here a detailed description of horizontal and vertical errors and their characteristics. The standard deviation measured over a 5.6 x 7 km area was approximately 2 m, the corresponding figures for relatively flat areas were in the 1-2 meter range and for mountainous areas in the 2-3 meter range, consistent with theoretical expectations. We also discuss key factors that presently limit the system performance

Global-scale topographic data are of fundamental importance to many Earth science studies, and obtaining these data is a priority for the Earth science community. Several groups have considered the requirements for such a data set, and a consensus assessment is that many critical studies would be enabled by the availability of a digital global topographic model with accuracies of 2 and 30 m in the vertical and horizontal directions, respectively. Radar interferometric techniques have been used to produce digital elevation models at these accuracies and are technologically feasible as the centerpiece of a spaceborne satellite mission designed to map the world's land masses, which we denote TOPSAT. A radar interferometer is formed by combining the radar echoes received at a pair of antennas displaced across-track, and specialized data processing results in the elevation data. Two alternative implementations, one using a 2 cm-/spl lambda/ radar, and one using a 24 cm-/spl lambda/ radar, are technologically feasible. The former requires an interferometer baseline length of about 15 m to achieve the required accuracy, and this could be built on a single spacecraft with a long extendible boom. The latter necessitates a kilometers long baseline, and would thus be best implemented using two spacecraft flying in formation. Measurement errors are dominated by phase noise, due largely to signal-to-noise ratio considerations, and attitude errors in determining the baseline orientation. For the 2-m accuracy required by TOPSAT, the orientation must be known to 1 arc-second. For the single-spacecraft approach, where attitude would be determined by star tracking systems, this performance is just beyond the several arc-second range of existing instruments. For the dual-spacecraft systems, though, differential global positioning satellite measurements possess sufficient accuracy. Studies indicate that similar performance can be realized with either system.>

An interferometric radar technique for topographic mapping of surfaces promises a high-resolution approach to the generation of digital elevation models. The authors present analyses of data collected by the synthetic aperture radar instrument on-board the ERS-1 satellite on successive orbits. Use of a single satellite in a nearly repeating orbit is attractive for reducing cost and spaceborne hardware complexity; also it permits inference of changes in the surface from the correlation properties of the radar echoes. The data have been reduced to correlation maps and digital elevation models. The correlation maps show that temporal correlation decreases significantly with time, but not necessarily at a constant well-defined rate, likely depending on environmental factors. When correlation among passes remains high, however, it is possible to form digital elevation models. Analyses of noise expected in ERS-1 interferometric data collected over Alaska and the southwestern United States indicate that maps with relative errors less than 5 m rms are possible in some regions. However, orbit uncertainties imply that tie points are required in order to reduce absolute height errors to a similar magnitude. The authors find that about 6 tie points per 40/spl times/40 km scene with 5 m rms or better height accuracy are needed to keep systematic map height errors below 5 m rms. The performance of the ERS-1 radar system for topographic applications, though useful for a variety of regional and local discipline studies, may be improved with respect to temporal decorrelation errors and absolute height acuity by modifying the orbit repeat period and incorporating precise orbit determination techniques. The resulting implementation will meet many, but not all, objectives of a global mapping mission.>

A new processing algorithm for the NASA JPL TOPSAR topographic radar mapper is described. It incorporates extensive motion compensation features as well as accurate three-dimensional target location algorithm. The processor applies an algorithm to resolving the absolute phase ambiguity. This allows rectified height maps to be generated without any use of ground reference points. The processor was tested using data acquired with extreme aircraft motion so that performance could be evaluated under adverse conditions. The topographic maps generated by the radar were compared to digital elevation models (DEMs) derived using conventional optical stereo techniques. In one region, the RMS elevation deviations measured were less than the specified DEM accuracy, and, in the region covered by the more accurate DEM, errors varied from 2.2 m RMS in relatively flat terrain up to 5.0 m in mountainous area. The RMS difference between radar and DEM elevation over the 6.5-km by 22-km area covered by the more accurate DEM was 3.6 m.>

It is shown that the assumptions about the backscatter and the polarimetric radar system for the two approaches described in J.D. Klein's (1992) and J.J. van Zyl's (1990) papers are equivalent. It is demonstrated that, to first order in the radar system crosstalk (i.e. neglecting terms of second order and above), an exact solution to the Stokes matrix format data calibration problem exists. It is shown that van Zyl's approach can give this first-order solution for appropriately symmetrized polarimetric radar data. Then it is shown how, if the data are properly symmetrized, van Zyl's approach can be used to calibrate both scattering matrix and Stokes matrix format data. These conclusions should be generally applicable to polarimetric imaging radar system data.>

It is known that interferometric synthetic-aperture radar (SAR) images can be inverted to perform surface elevation mapping. Among the factors critical to the mapping accuracy are registration of the interfering SAR images and phase unwrapping. A registration algorithm is presented that determines the registration parameters through optimization. A figure of merit is proposed that evaluates the registration result during the optimization. The phase unwrapping problem is approached through a new method involving fringe line detection. The algorithms are tested with two SEASAT SAR images of terrain near Yellowstone National Park. These images were collected on SEASAT orbits 1334 and 1420, which were very close together in space, i.e. less than 100 m.>

The authors have augmented the NASA DC-8 AIRSAR instrument with a pair of C-band antennas displaced across track to form an interferometer sensitive to topographic variations of the Earth's surface. During the 1991 DC-8 flight campaign, data were acquired over several sites in the US and Europe, and topographic maps were produced from several of these flight lines. Analysis of the results indicate that statistical errors are in the 2-4-m range, while systematic effects due to aircraft motion are in the 10-20-m range. The initial results from development of a second-generation processor show that aircraft motion compensation algorithms reduce the systematic variations to 2 m, while the statistical errors are reduced to 2-3 m.>

A radar interferometric technique for topographic mapping of surfaces, implemented utilizing a single synthetic aperture radar (SAR) system in a nearly repeating orbit, is discussed. The authors characterize the various sources contributing to the echo correlation statistics, and isolate the term which most closely describes surficial change. They then examine the application of this approach to topographic mapping of vegetated surfaces which may be expected to possess varying backscatter over time. It is found that there is decorrelation increasing with time but that digital terrain model generation remains feasible. The authors present such a map of a forested area in Oregon which also includes some nearly unvegetated lava flows. Such a technique could provide a global digital terrain map.>

Presents a tutorial review of the broad sweep of topics relating to imaging radar polarimetry, ranging from mathematical foundations to hardware and from implementation approaches to signal processing and calibration. The authors examine current developments in sensor technology and implementation for recording polarimetric measurements, and describe techniques and areas of application for this form of remotely sensed data. Those aspects of ground signal processing and calibration peculiar to the polarimetric signals are addressed. Several of the currently operating instruments and some of the implementations planned for future use are discussed.>

The authors present the results of an experiment using the NASA/JPL DC-8 AIRSAR (aircraft synthetic-aperture radar) over a coniferous forest near Mt. Shasta (California) in 1989. Calibration devices were deployed in clearings and under the forest canopy and passes at 20 degrees , 40 degrees , and 55 degrees incidence angles were made with the AIRSAR. A total of eight images at differing incidence angles have been processed and calibrated. The multipolarization multifrequency data were examined, and it was found that the C-band cross section averaged over like and cross polarizations is the best parameter for distinguishing between two stands with differing forest biomass. The average cross section at P- and L-bands is useful only for smaller incidence angles. Parameters describing the polarization behavior of the scattering were primarily useful in identifying the dominant scattering mechanisms for forest backscatter.>

The authors developed a calibration procedure for imaging radar polarimeters and applied it to a set of images acquired by the NASA DC-8 multifrequency radar system. The technique requires the use of ground reflectors of known cross-section for absolute calibration, that is, solution for sigma /sup 0/; however, the image data themselves can usually provide all information necessary for phase calibration and for antenna crosstalk correction. The accuracy of the approach, as measured by calculating the cross-section residuals of known targets in each calibrated scene, is on the order of +or-1-2 dB at the P- and C-band, but improves to +or-0.5 dB at the L-band. The authors present the results of applying this technique to radar scenes of lava flows of varying roughness, temperate and tropical rain forests, and ocean water surfaces. They also present several example applications which are feasible with calibrated data but which would be difficult to implement with uncalibrated data.>

We have analyzed radar polarimeter signals from a set of trihedral corner reflectors located in the Goldstone Dry Lake in Cafifornia, and observed three types of scattering behavior: i) Bragg-like slightly rough surface scattering that represents the background signal from the dry lake, ii) trihedral corner reflector scattering that returns the incident polarization, and iii) two-bounce corner reflector scattering resulting from a particular alignment of a trihedral reflector. In the latter case, we can measure within about 3° the orientation angle of the apparent dihedral trough, even though the 2-m reflector is much smaller than the 10-m resolution element of our radar. Thus a radar calibration approach using trihedral corner reflectors should be designed such that precise alignment of the reflectors is ensured, as three-bounce and two-bounce geometries lead to very different cross sections and hence very different inferred calibration factors.