Scatterometers are designed to measure the normalized radar backscatter (σo) of the ocean surface. Wind-only Geophysical Model Functions (GMF) are used to relate the σoto the near surface Wind Vector (WV), wind direction and speed. Rain alters the relationship between WV and σocaptured in the wind-only GMF. Making a GMF that encapsulates the rain effects on σolowers the error between retrieved WV and actual measured WV.The Ocean Scatterometers (OSCAT-1 and -2) are a Ku-band dual beam scatterometers that operated from 2009-2021. To create a rain GMF OSCAT is collocated with Tropical Rain Measurement Mission (TRMM) rain measurements and European Centre for Medium-Range Weather Forecast (ECMWF) model winds. The rain GMF is created using a simple phenomenological rain model that relates σoin the presence of rain to surface wind-induced roughness. The Simultaneous Wind and Rain (SWR) algorithm uses the conventional wind-only GMF injunction with the rain GMF to estimate WV from σo.

In this paper we provide an update on the development status and performance estimates for the proposed Global L-band active/passive Observatory for Water cycle Studies (GLOWS) L-band active/passive SMAP continuity mission. GLOWS is currently funded by the NASA Instrument Incubator Program.

Long-term studies of climate change can benefit from a retrospective analysis of past satellite missions. To facilitate such studies, in this paper we describe two new radar backscatter (σo) products derived from the Seasat-A Scatterometer System (SASS) and the NASA Scatterometer (NSCAT). Operating at Ku-band (14.6 and 13.995 GHz, respectively), the two sensors provided global σomeasurements in 1978 and 1996-1997 and are unique benchmarks for studying long-term climate change, particularly over the polar ice sheets. The products are compatibly gridded to be consistent with other existing climate data record products of σoand microwave brightness temperature (TB) that span multiple decades.

We use enhanced-resolution L-band radar backscatter and brightness temperature image time series generated from observations collected by NASA’s Soil Moisture Active Passive (SMAP) mission to map the extent and physical characteristics of an expansive perennial firn aquifer recently identified in the Wilkins Ice Shelf, Antarctic Peninsula. Empirical algorithms to map extent are derived from a continuous logistic model. Distinct physical characteristics of the firn overlying the perennial firn aquifer are distinguished using images of the slope in radar backscatter versus incidence angle. An inverted two-layer L-band brightness temperature model parameterized by the optical thickness is used to map seasonal firn saturation. Our results demonstrate the potential for using low-frequency L-band satellite microwave missions to map large englacial meltwater reservoirs in the firn of Antarctic ice shelves that may trigger hydrofracture and ice shelf instability year-round.

Current and past scatterometers provide useful measurements of vector winds, vegetation, and ice that span multiple decades, with measurements collected at Ku-, C-, and most recently, L-band. To support long-term climate studies cross-calibration of the sensors is important. The multiple frequency bands complicate intersensor calibration. Using a simple scattering model, we evaluate the use of wind glazed regions in Antarctica as a transfer standard to cross-calibrate backscatter from the various sensors in the scatterometer dataset. We characterize the consistency of the data from the wind glaze region and propose a combined surface and volume scattering model for determining the calibration offset. We evaluate the sensitivity of the model in its current state with respect to those parameters.

SMOS and SMAP radiometers have demonstrated the ability to monitor soil moisture and sea surface salinity and continue to provide high quality radiometric measurements to this day in extended mission operations. It is important to maintain data continuity for these science measurements. The proposed instrument concept (Global L-band active/passive Observatory for Water cycle Studies - GLOWS) will enable low-cost L-band data continuity (that includes both L-band radar and radiometer measurements). The objective of this project is to develop key instrument technology to enable L-band observations using an Earth Venture class satellite. Specifically, a new deployable reflectarray lens antenna is being developed that will enable a smaller EELV Secondary Payload Adapter (ESPA) Grande-class satellite mission to continue the L-band observations at SMAP and SMOS resolution and accuracy at substantially lower cost, size, and weight.

The MEaSUREs Calibrated Enhanced-Resolution Passive Microwave Daily EASE-Grid 2.0 Brightness Temperature (CETB) Earth Science Data Record and the SMAP Twice-Daily rSIR-Enhanced EASE-Grid 2.0 Brightness Temperature (SETB) Data Set provide an extensive multi-instrument, multi-decadal, time series of global enhanced-resolution microwave radiometer image products. The Earth-based CETB and SETB archives are generated using swath-based multi-frequency micro-wave brightness temperature (TB) observations collected by the Nimbus-7 Scanning Multichannel Microwave Radiometer (SMMR), the Special Sensor Microwave/Imager (SSM/I) and Special Sensor Microwave Imager/Sounder (SSMIS) series, the Advanced Microwave Scanning Radiometer–Earth Observing System (AMSR-E), and the Soil Moisture Active Passive (SMAP) microwave radiometer. In this paper, we describe new global SMAP enhanced-resolution scatterometer and synthetic aperture radar (SAR) image products that augment the CETB and SETB archives. Earth-based, morning and evening images are generated using swath-based L-band (1.26 GHz) radar backscatter (σ°) observations and previous developed image reconstruction techniques. The available image products include cylindrical and azimuthal projection scatterometer and SAR σ° images with 1- and 3-day imaging intervals, and azimuthal projection scatterometer σ° images with an 8-day imaging interval that have been further corrected for incidence and azimuth angle variation over the Greenland and Antarctic ice sheets. These σ° image time series are compatibly-gridded with CETB and SETB TBimage time series to support a wide variety of geophysical applications.

Radiometric and radar data from NASA’s Soil Moisture Active Passive (SMAP) satellite are presented in a study of the Ross Ice Shelf, Antarctica. L-band brightness temperature (TB) patterns compare favorably to the outflow patterns from East and West Antarctica. CoolerTBis associated with the broad outflow from West Antarctic ice streams and the outflow from narrow outlet glaciers that drain East Antarctica. Aside from outlet glacier discharges, ice from East Antarctica is thinner and radiometrically warmer than that from West Antarctica.TBis stable across the ice shelf over the 6-year period of observations (1-2 K standard deviation). Over shorter times, surface melt events cause pre- and post-meltTBto vary by as much as 5 K in vertical polarization. Radar measurements highlight areas where backscatter is strong from melt related ice lenses and ice layers, consistent with a corresponding decrease inTB. TheTBpolarization ratio on the ice shelf is approximately 1.13 and decreases from the West Antarctic grounding line towards the East Antarctic outlet glaciers. The backscatter polarization ratio increases by several dB from West to East Antarctica, indicating a decreasing influence of volume scatter toward East Antarctica. The decreasingTBpolarization ratio indicates a diminishing role of firn layering on the depth-integrated emission. There is a negative correlation between warming brightness temperature and thinning. An explanation for this observation is that the relatively warm ice shelf has a relatively shallow (500 m or less) penetration depth leading to a warm physical temperature bias.

L-band observations have proven useful for estimating soil moisture and ocean salinity variables to study the land surface and ocean. The European Space Agency (ESA) Soil Moisture and Ocean Salinity (SMOS) mission was the first (2009-present) spaceborne L-band radiometer. This was followed by two L-band missions flown by the National Aeronautics and Space Administration (NASA) to measuresea surface salinity (Aquarius 2011–2015) and soil moisture (SMAP 2015-present). It is critical to continue the time series of L-band observations that these missions have begun. To address this need we propose a new low-cost instrument concept known as the Global L-band active/passive Observatory for Water cycle Studies (GLOWS) that will include an L- band radiometer and radar to provide data continuity. The new mission concept includes a deployable reflectarray lens antenna with a compact feed that can be flown on an Earth Venture class satellite in a EELV Secondary Payload Adapter (ESPA) Grande-class mission. GLOWS will continue the science observations of SMAP and SMOS at the same resolution and accuracy at substantially lower cost, size, and weight.

Image reconstruction from discrete samples is fundamental in remote sensing. The reconstruction problem is more complicated when the sample locations are irregularly spaced and employ different aperture functions. Further, not all 2-D sampling configurations permit full image reconstruction of band-limited signals. In this article, 2-D signal reconstruction from irregular samples with variable apertures is considered using theory and examples. Exact reconstruction requires that the signal be band-limited and the sampling matrix be invertable. The results are sensitive to the sample locations and noise in the measurements. Illustrative examples are provided using simulation and actual data from the L-band Soil Moisture Active Passive (SMAP) radiometer. The general approach can be employed with other sensors.

This paper introduces RapidScat 2.5-km ultrahigh-resolution (UHR) wind estimation and validates it in near-coastal regions. RapidScat UHR wind estimation provides finer resolution ocean wind vector fields than conventional 12.5-km level 2B (L2B) wind products at a cost of higher noise. In addition, this paper applies direction interval retrieval techniques and develops other wind processing improvements to enhance the performance of RapidScat UHR wind estimation. The new algorithm is validated with L2B wind estimates, numerical weather prediction wind products, and buoy measurements. The wind processing improvements produce more spatially consistent UHR winds that compare well with the wind products mentioned above.

Wind scatterometer measurements are collected over an irregular grid, and processing is required to generate backscatter images on an Earth-centered grid. The most common algorithms used for this are “drop in the bucket” (DIB) and variations of the scatterometer image reconstruction (SIR) algorithm. These algorithms are also used for radiometer brightness temperature imaging. The Backus-Gilbert (BG) algorithm has been used for radiometer imaging but has not been applied to scatterometer backscatter imaging. In this paper, the application of BG to scatterometer backscatter imaging is explored and its performance is compared to DIB and SIR. Like SIR, optimally tuned BG is capable of producing higher resolution images than DIB, though its noise performance is slightly inferior to SIR's. While BG and SIR produce similar results for radiometer data, the higher relative noise level of scatterometer data increases the differences between the SIR and BG algorithm performance, and limits the performance of BG relative to SIR in scatterometer imaging. Comparison of the SIR and BG algorithms in scatterometer imaging offers important insights into the inversion/reconstruction problem.

The NASA-sponsored Calibrated Passive Micro- wave Daily Equal-Area Scalable Earth Grid 2.0 Brightness Temperature (CETB) Earth System Data Record Project team has generated a multisensor, multidecadal time series of high-resolution radiometer products designed to support climate studies. This project uses image reconstruction techniques to generate conventional and enhanced-resolution daily brightness temperature images on a standard set of map projections. Sensors included in CETB are the Aqua Advanced Microwave Scanning Radiometer-Earth Observing System (AMSR-E), Scanning Multichannel Microwave Radiometer, and all Special Sensor Microwave/Imager and Special Sensor Microwave Imager/Sounder radiometers. These span frequencies between 6 and 89 GHz. This paper considers the issues of adding the L-band (1.6 GHz) Soil Moisture Active Passive (SMAP) radiometer measurements to the CETB climate record, with emphasis on optimizing the reconstruction to provide the highest possible spatial resolution at the lowest noise level. SMAP radiometer reconstruction on SMAP-standard grids is also considered. Simulation is used to optimize the reconstruction, and the results confirmed using actual data. A comparison of the performance of the Backus-Gilbert approach and the radiometer form of the Scatterometer Image Reconstruction algorithm is provided. These are compared to the conventional drop-in-the-bucket gridded imaging.

Since 1978, a series of Earth-observing, satellite-borne passive microwave sensors has produced a rich record of microwave brightness temperatures. Passive microwave sensors can see through most clouds and collect measurements both day and night, which is especially useful in high latitudes during polar night. Passive microwave measurements are used to derive significant and meaningful climate records of many parameters, including the dramatic decline in Arctic sea ice. Earlier revisions of this significant climate record have been produced as flat, binary, gridded arrays with minimal or no file-level metadata and no machine-readable geolocation. Developed for many applications in polar regions, data were projected to polar azimuthal and global cylindrial projections that users found difficult to handle in standard mapping software packages. Funded by the NASA MEaSUREs program, we are using state-of-the-art image reconstruction techniques to produce a Calibrated Enhanced-Resolution Passive Microwave Equal-Area Scalable Earth Grid 2.0 Brightness Temperature (CETB) Earth System Data Record (ESDR) that leverages the improved EASE-Grid 2.0 projection definitions and netCDF-CF metadata conventions to improve usability of the data products. We describe our approach to defining file-level metadata that is intelligible to standard software packages, including open source netCDF Operators (NCO) and Geospatial Data Abstraction Library (gdal), and the commercial ESRI ArcMap geospatial mapping tool.

A reliable method of automatically estimating Antarctic iceberg contours and sizes from satellite data is desirable to help better understand patterns in iceberg formation and behavior. Starting from scatterometer images, this paper develops a method of using the relatively constant backscatter values across the surface of an iceberg to derive a contour of its shape. Contours are then used to find an angle of rotation between images taken on successive days. This method produces size estimates that are within 10% of the area given by the National Ice Center (NIC).

The loss of the active sensor channels limits the spatial resolution capability of the Soil Moisture Active Passive (SMAP) mission in measuring soil moisture. Image reconstruction techniques can provide an enhanced resolution product. Using such processing techniques and code developed for the NASA MEaSUREs Calibrated Passive Microwave Daily Equal-Area Scalable Earth (EASE) Grid 2.0 Brightness Temperature (CETB) Earth System Data Record (ESDR) project, we are generating enhanced-resolution SMAP TB products on both 3 km SMAP project grids and 3.125 km CETB-compatible grids. The intrinsic tradeoff between resolution and noise is explored using multiple resolution enhancement algorithms, and the spatial response functions for each are presented.

The Ku-band Oceansat-2 Scatterometer (OSCAT) is very similar to the Quick Scatterometer (QuikSCAT), which operated from 1999 to 2009. OSCAT continues the Ku-band scatterometer data record through 2014 with an overlap of 19 days with QuikSCAT's mission in 2009. This letter discusses a particular climate application of the time series for sea ice extent observation. In this letter, a QuikSCAT sea ice extent algorithm is modified for OSCAT. Gaps in OSCAT data are accounted for and filled in to support sea ice extent mapping. The OSCAT sea ice extent data are validated with QuikSCAT and Special Sensor Microwave/Imager sea ice extent data.

Satelliteborne C-band scatterometer measurements of the radar backscatter coefficient $(\sigma^{0})$ of the Earth can be used to estimate soil moisture levels over land. Such estimates are currently produced at 25- and 50-km resolution using the Advanced Scatterometer (ASCAT) sensor and a change detection algorithm originally developed at the Vienna University of Technology (TU-Wien). Using the ASCAT spatial response function (SRF), high-resolution (approximately 15-20 km per pixel) images of $\sigma^{0}$ can be produced, enabling the creation of a high-resolution soil moisture product using a modified version of the TU-Wien algorithm. The high-resolution soil moisture images are compared to images produced with the Water Retrieval Package 5.5 algorithm, which is also based on the TU-Wien algorithm, and to in situ measurements from the National Oceanic and Atmospheric Administration U.S. Climate Reference Network (NOAA CRN). The WARP 5.5 and high-resolution image products generally show good agreement with each other; the high-resolution estimates appear to resolve soil moisture features at a finer scale and demonstrate a tendency toward greater moisture values in some areas. When compared to volumetric soil moisture measurements from NOAA CRN stations for 2010 and 2011, the WARP 5.5 and high-resolution soil moisture estimates perform similarly, with both having a root-mean-square difference from the in situ data of approximately 0.06 m3/m3in one study area and 0.09 m3/m3in another.

For most land and ice surfaces, the measured radar backscatter at the spatial resolution of a wind scatterometer is insensitive to the azimuth angle. However, for regions of East Antarctica, the backscatter strongly depends on the azimuth angle. This relationship between backscatter and azimuth angle is often modeled with a Fourier series. Although previous work has separately examined the data from QuikSCAT, a Ku-band scatterometer, and the Advanced Scatterometer (ASCAT), a C-band scatterometer, this letter compares the two on the same high-resolution grid. We find that, although QuikSCAT has superior azimuth angle coverage (due to its measurement geometry) compared to ASCAT, both are suitable to estimate the radar backscatter azimuth angle modulation of East Antarctica. The ASCAT data exhibit a much larger azimuth modulation than the QuikSCAT data. This is attributed to the different wavelengths of the microwave signal: At C-band (5.7 cm), East Antarctica has features that show radar backscatter to be more dependent on azimuth angle than at Ku-band (2.2 cm). This letter also examines the ASCAT and QuikSCAT azimuth modulation over previously identified regions of wind glaze. Although azimuth modulation is expected to be minimal over wind glaze, we find the wind-glaze regions to contain more structure than previously suggested.

Arctic sea ice can be classified as first-year (FY) or multiyear (MY) based on data collected by satellite microwave scatterometers. The Oceansat-2 Ku-band Scatterometer (OSCAT) was operational from 2009 to 2014 and is here used to classify ice as FY or MY during these years. Due to similarities in backscatter measurements from sea ice and open water, a NASA Team ice concentration product derived from passive microwave brightness temperatures is used to restrict the classification area to within the sea ice extent. Classification of FY and MY ice is completed with OSCAT by applying a temporally adjusted threshold on backscatter values. The classification method is also applied to the Quick Scatterometer (QuikSCAT) data set, and ice age classifications are processed using QuikSCAT for 1999-2009. The combined QuikSCAT and OSCAT classifications represent a 15-year record, which extends from 1999 to 2014. The classifications show a decrease in MY ice, while the total area of the ice cover remains consistent throughout winter seasons over the time series.

The Advanced SCATterometer (ASCAT) measurement spatial response function (SRF) relates the weighted contribution of every location within the measurement footprint to the measured normalized radar cross section σ°. The SRF results from a combination of the antenna response and the onboard processing and is computed during ground processing by modeling in detail the measurement geometry, as this is required for an accurate σ° estimation. However, the computed SRF is not disseminated as part of the L1B data. For some applications of the L1B data, the SRF is additionally required. For these applications, an approximate description of the SRF is often sufficiently accurate, estimated from information contained in the L1B data, rather than from a full calculation based on the measurement geometry. This paper describes a parameterized model of the ASCAT SRF for each measurement. First, an SRF reference estimate that incorporates details on the ASCAT design and onboard measurement processing is created. A parameterized model is fit to the reference estimate. The parameterized SRF is computationally less demanding than the reference estimate and as such more useful for near-real-time processing. The two estimates are validated with the computed SRF used in ground processing and with the transponder data from calibration campaigns. Finally, to validate the SRF in a simple application, the land fraction (a measure of land contamination in near-coastal ocean measurements) is computed and compared to actual data for a sample region.

The standard ocean wind product from the Advanced Scatterometer (ASCAT) is retrieved on a 12.5-km grid. Ultrahigh-resolution (UHR) processing enables ASCAT wind retrieval on a high-resolution 1.25-km grid. Ideally, such a high-resolution grid allows for improved analysis of winds with high spatial variability, such as those in near-coastal regions and storms. This paper provides an analysis and validation of ASCAT UHR wind estimates to evaluate its spatial resolution and accuracy. This is done via a comparison with two other sources: buoy-measured winds in coastal regions and winds estimated from synthetic aperture radar (SAR) data over the open ocean. Near-coastal ocean measurements may be contaminated by nearby land, introducing a wind speed bias in the retrieved winds. To enable near-coastal UHR wind retrieval, we use a land contribution ratio (LCR) approach to discard ASCAT measurements with high land contamination before UHR processing and wind retrieval. Through a comparison with near-coastal buoy winds, we find that the LCR approach is appropriate for precisely controlling the tradeoff between land contamination and spatial coverage near land. We find that the resolution of the UHR data is improved over the 25-km wind product to 10 km, and likely down to 4 km in some cases. In comparing SAR and UHR winds, we find that both products have common fine-scale features and have derivative fields that match well and that the UHR product matches better the expected spectral properties of ocean winds.

A wind scatterometer measures the normalized radar cross section σ° of the Earth's surface in order to estimate the ocean near-surface wind. Each measurement is the weighted integral of σ° over an area, so samples of σ° are filtered by a measurement spatial response function (SRF). Enhanced-resolution σ° data may be produced from σ° measurements and their associated SRF using image reconstruction algorithms. These enhancedresolution data products are useful for land and ice studies in addition to high-resolution ocean wind retrieval. The spatial resolution and noise may be improved by combining data for multiple passes, trading off temporal resolution for spatial resolution. In this paper, we consider the production of enhanced-resolution σ° image reconstruction from the Advanced Scatterometer (ASCAT) on the MetOp satellites for land and ice regions. Two previously introduced image reconstruction algorithms, the weighted AVErage (AVE) and Scatterometer Image Reconstruction (SIR), are applied to ASCAT data, as well as a conventional gridding method. We select optimum values of the reconstruction parameters for ASCAT. The imaging algorithms are compared and evaluated, as quantified by the spatial resolution, pixel mean and variance, and pixel correlation of the produced images. The effective spatial resolution of the AVE and SIR reconstructed images is on the order of 15 to 20 km, an improvement over the nominal spatial resolution of 25 and 50 km for the spatially averaged swath-oriented ASCAT L1B products.

This paper considers some of the issues of radiometer brightness image formation and reconstruction for use in the NASA-sponsored Calibrated Passive Microwave Daily Equal-Area Scalable Earth Grid 2.0 Brightness Temperature Earth System Data Record project, which generates a multisensor multidecadal time series of high-resolution radiometer products designed to support climate studies. Two primary reconstruction algorithms are considered: the Backus-Gilbert approach and the radiometer form of the scatterometer image reconstruction (SIR) algorithm. These are compared with the conventional drop-in-the-bucket (DIB) gridded image formation approach. Tradeoff study results for the various algorithm options are presented to select optimum values for the grid resolution, the number of SIR iterations, and the BG gamma parameter. We find that although both approaches are effective in improving the spatial resolution of the surface brightness temperature estimates compared to DIB, SIR requires significantly less computation. The sensitivity of the reconstruction to the accuracy of the measurement spatial response function (MRF) is explored. The partial reconstruction of the methods can tolerate errors in the description of the sensor measurement response function, which simplifies the processing of historic sensor data for which the MRF is not known as well as modern sensors. Simulation tradeoff results are confirmed using actual data.

Sampling plays a critical role in remote sensing and signal analysis. In conventional sampling theory, the signal is sampled at a uniform rate at a minimum of twice the signal bandwidth. Sampling with an aperture function requires a fixed-aperture function, which can be removed by deconvolution after signal reconstruction. However, in some cases, the signal samples are available only at irregular positions, and different samples use different aperture functions. In this paper, the theory of finite-length signal reconstruction with irregular samples and variable apertures in one and two dimensions is considered. In the 1-D case, a band-limited discrete signal can be exactly reconstructed from a finite number of arbitrarily spaced samples with few restrictions on the aperture functions. Exact reconstruction in the 2-D case requires the sampling matrix be invertable, and is not always possible. Variable aperture functions, while complicating the process, can enable reconstruction for a broader range of sample locations. Practical issues are discussed, and numerical examples are provided. Variable aperture reconstruction has application in a variety of remote sensing problems. In this paper, reconstruction from 2-D irregular sampling with variable apertures is illustrated using Special Sensor Microwave/Imager radiometer observations.

Launched in September2014, RapidScat is currently operating on the International Space Station (ISS). RapidScat estimates ocean vector winds via the measurement of the normalized radar coefficient (σ0) of the ocean's surface. Measurements are also collected over land. The ISS orbit permits, for the first time, the observation of the diurnal variation in Ku-band σ0at midto high-incidence angles. To complement calibration efforts over the ocean, in this paper the calibration and validation of the σ0measurements are performed using natural land targets, namely the Amazon and Congo rainforests. The diurnal σ0cycle of the targets is estimated with respect to incidence angle, azimuth angle, and season using measurements from previous sensors. Understanding this diurnal backscatter response enables the comparison of RapidScat measurements with measurements from the QuikSCAT, NASA Scatterometer, SeaWinds, and Oceansat-II scatterometers. RapidScat σ0measurements are found to be consistent but biased low compared to those of QuikSCAT by up to 0.3 dB. The effectiveness of slice balancing is evaluated and found to be dependent on the pitch of the ISS. Extreme pitches of the ISS are also found to introduce azimuth dependencies in egg measurements. By accounting for seasonal and diurnal cycles, we find that the rainforests are well suited for scatterometer sensor cross-calibration, even for disjoint years.

RapidScat, which is a Ku-band scatterometer mounted on the International Space Station, observes the Earth's surface in a non-sun-synchronous orbit allowing for different local time-of-day (LTOD) observations as the orbit progresses. The unique orbit and different LTOD observations provide surface observations that are composited to describe the diurnal variability of Ku-band normalized backscatter (σ0) measurements over land globally. Previous sun-synchronous scatterometers providing twice-daily surface observations have been used to demonstrate some diurnal changes in σ0in several regions globally, but instrument cross-calibration concerns prevent identifying diurnal changes by combining σ0observations from multiple sensors. As a result, the full extent of diurnal changes to σ0has not been determined until now. In this paper, RapidScat is used to identify diurnal changes to σ0globally. Vegetation type is discussed with respect to the diurnal changes in σ0regionally. The global diurnal changes to σ0are discussed with emphasis on the Amazon, Congo, and Upper Danube river regions. Diurnal cycles are described that could not previously be identified with sun-synchronous instruments. Global means and the magnitude of the diurnal cycle are discussed. With the diurnal changes identified and quantified, RapidScat can be used for future cross-platform calibrations using land targets.

The field of wireless communications has benefited from multiple-input and multiple-output (MIMO) techniques. As researchers seek to apply MIMO (multistatic) techniques to radar and specifically to synthetic aperture radar (SAR), a key factor in determining MIMO application and performance is the level of correlation of signals from different receiver/transmitter pairs. The level of correlation determines whether a MIMO array falls into the category of a collocated array or a distributed array. The type of array dramatically affects which MIMO techniques may be performed and what advantages MIMO offers from conventional techniques. This paper presents models for calculating geometric correlation of multistatic SAR pixels using a ground-plane image formation. The models' results are compared to previous correlation models found in literature. A key result is that correlation depends on pixel resolution and not the number of individual scatterers. This paper concludes that most MIMO arrays operating on a single platform operate in the collocated regime.

This paper presents a generalized treatment of image formation for a linear-frequency-modulated continuous wave (LFM-CW) synthetic aperture radar (SAR) signal, which is a key technology in making very small SAR systems viable. The signal model is derived, which includes the continuous platform motion. The effect of this motion on the SAR signal is discussed, and an efficient compensation method is developed. Processing algorithms are developed including precise and approximate backprojection methods and a generalized frequency scaling algorithm that accounts for an arbitrary number of terms of a Taylor expansion approximation of the SAR signal in the Doppler frequency domain. Together, these algorithms allow for the processing of LFM-CW SAR data for a wide variety of system parameters, even in scenarios where traditional algorithms and signal approximations break down.

The traditional radar stop-and-hop assumption, that a radar is stationary during the transmission of a single pulse, can be violated with LFM-CW SAR. Because the stop-and-hop assumption is used as a basis for the most commonly used SAR processing methods, this can lead to significant degradation of the final image, particularly in low-altitude ultra-wide-band SAR. In this paper we extend previous work compensating for motion during a pulse to ultra-wide-band strip-map LFM-CW SAR.

Originally designed for wind velocity estimation over the ocean, scatterometers have been applied to weather forecasting, global climatological studies, and monitoring of large-scale human interaction on the planet. Launched in September of 2009, the Oceansat-2 Ku-band scatterometer (OSCAT) is an excellent candidate for continuing the data time series begun with QuikSCAT, which was the SeaWinds scatterometer flown on the QuikSCAT mission 1999-2009. Some processing algorithms require knowledge of the spatial response function (SRF) of the scatterometer. With limited knowledge of OSCAT implementation, and thus its SRF, a procedure is developed for estimating the SRF. The estimation procedure uses scatterometer measurements over islands to invert the radar equation. A mathematical model is developed that reduces the solution to rank-reduced least-squares estimation. The geographic sampling region is discussed, and a simulation is performed to verify the efficacy of the method, while also providing guidelines for island choice and the number of singular values used in rank reduction. The utility of OSCAT SRF estimates is demonstrated through the construction of an enhanced-resolution radar backscatter image over the Amazon rainforest.

In the above-named article [ibid., vol. 51, no. 5, pp 1924-1934, April 2014] there are errors in Table I. The corrected table is published here.

Scatterometers provide frequent estimates of near-surface wind vectors over the Earth's oceans. However, in the polar oceans, the presence of sea ice in or near the measurement footprint can adversely affect scatterometer measurements, resulting in inaccurate wind estimates. Currently, such ice contamination is mitigated by discarding measurements within 50 km of the detected sea ice. This approach is imperfect and causes loss of coverage. We present a new algorithm that detects ice-contaminated measurements based on a metric called the ice contribution ratio (ICR), which measures the spatial ice contribution for each measurement. Determined by simulation, we threshold the ICR depending on local wind, ice backscatter, and cross-track location. Using ICR processing, wind is retrieved almost 40 km closer to sea ice than has been previously possible, while ensuring wind accuracy. Using ICR processing, retrieved wind distributions more closely resemble uncontaminated distributions than winds retrieved using previous methods. The algorithm is applied to QuikSCAT in this paper, but could be applied to other scatterometers such as the Oceansat-2 scatterometer.

Polar sea ice is an important input to global climate models and is considered to be a sensitive indicator of climate change. While originally designed only for wind estimation, radar backscatter measurements collected by wind scatterometers have proven useful for estimating the extent of sea ice. During the Quick Scatterometer (QuikSCAT) mission, SeaWinds data were used to operationally map the sea ice extent. The resulting sea ice maps were used to mask near-surface winds to support SeaWinds' primary mission of measuring near-surface winds over the ocean. This paper describes the operational SeaWinds sea ice extent mapping algorithm, provides validation comparisons, and presents results from the ten-year data product. Starting with enhanced resolution horizontal polarization and vertical polarization backscatter images, the algorithm employs an iterative maximum-likelihood classifier with fixed thresholds to segment sea ice and open ocean pixels. Residual classification errors are reduced through binary image processing techniques and sea ice growth/retreat constraint methods. The algorithm results are compared with sea ice concentrations derived from Special Sensor Microwave/Imager data and with RADARSAT synthetic aperture radar imagery. The results suggest differences in the sensitivities of active and passive products given their channel sets and specific algorithms. Derived sea ice extents over the full decade-long QuikSCAT mission data set are analyzed to show important trends in sea ice extent for the Antarctic and Arctic regions.

Wind stress is the most important ocean forcing for driving tropical surface currents. Stress can be estimated from scatterometer-reported wind measurements at 10 m that have been extrapolated to the surface, assuming a neutrally stable atmosphere and no surface current. Scatterometer calibration is designed to account for the assumption of neutral stability; however, the assumption of a particular sea state and negligible current often introduces an error in wind stress estimations. Since the fundamental scatterometer measurement is of the surface radar backscatter (sigma-0) which is related to surface roughness and, thus, stress, we develop a method to estimate wind stress directly from the scatterometer measurements of sigma-0 and their associated azimuth angle and incidence angle using a neural network approach. We compare the results with in situ estimations and observe that the wind stress estimations from this approach are more accurate compared with those obtained from the conventional estimations using 10-m-height wind measurements.

Postlaunch calibration of satellite-borne scatterometers using backscatter data from natural land targets helps to maintain scatterometer accuracy. Due to its temporal stability, the dry-snow zone of the Greenland ice sheet has been proposed in previous studies as a calibration target. Using QuikSCAT data, this letter examines the backscatter properties of the dry-snow zone that are relevant to Ku-band scatterometer calibration including temporal and spatial variabilities, and azimuth-angle and polarization and incidence-angle dependences. The backscatter is found to seasonally vary throughout the dry-snow zone by 0.4 dB on average. Small interannual variations (less than 1.5 dB over a nine year period) in the backscatter are also present in some regions. Azimuth modulation is generally less than 0.4 dB in magnitude and is not significant in some regions of the dry-snow zone. Melting and refreezing appear to cause the quasi-polarization ratio to temporarily decrease. Spatially consistent and relatively temporally stable regions that are well within the interior of the dry-snow zone are best suited as calibration sites.

The electromagnetic spectrum is a valued shared resource. Its scientific use allows us to learn about our universe, measure and monitor our planet, and communicate scientific data. The use of the spectrum is managed by national, regional, and global regulatory frameworks. There are increasing demands for new or extended allocations because of vast technological advances in the past few years. Understanding spectrum management is important in the successful planning and execution of missions and instruments, as well as in determining the potential source of radio frequency interference in existing data and instruments, and in working to ameliorate its impact. This paper provides a summary of this framework for radio scientists and engineers.

QuikSCAT was designed for ocean wind retrieval. However, its wind estimation performance is limited in rainy conditions. Several estimation techniques have been proposed: wind-only (WO), simultaneous wind and rain (SWR), and rain-only, which are appropriate for different levels of rain contamination. To exploit the strengths of each estimation method at mitigating rain contamination, a Bayes estimator selection (BES) technique has been developed for 25-km wind products to select from among the several estimation techniques for each wind vector cell. This paper adapts the BES concept [1] to QuikSCAT ultra-high resolution (UHR) 2.5-km, products and extends BES to include prior selection and noise reduction. Prior selection and noise reduction exploit general spatial characteristics of wind and rain fields to improve the accuracy of estimator selections. Together these techniques enable improved estimator selection performance so that the probability of selecting the estimate with minimum squared error approaches optimal levels. Optimal estimator selection reduces variability of wind estimates during rainy conditions and provides rain estimates when possible without using additional sources of information. Overall, UHR wind estimation performance with the new technique has improved bias and root mean-squared error, -0.16 m/s and 2.15 m/s, respectively, which are lower than either of the UHR WO and UHR SWR estimates.

In this study, we analyze 83 synthetic aperture radar (SAR) images including 73 from RADARSAT-1 and 10 from ENVISAT that contain tropical cyclone eye information. We also obtain ancillary tropical cyclone intensity information from NOAA National Hurricane Center and Japan Meteorological Agency. Based on this information, we generate tropical cyclone morphology statistics. We found that majority of the hurricanes are in wavenumber 1 and 2 category. Marine atmospheric boundary layer rolls can also be extracted from SAR image.

The 2010 Deepwater Horizon oil spill in the Gulf of Mexico covered a sufficiently large area to be observed by the European Space Agency Advanced Scatterometer (ASCAT) on MetOp-A. In this paper, ASCAT data and numerically computed winds from the European Centre for Medium-Range Weather Forecasts (ECMWF) are used to map the spatial extent of oil on the ocean surface over the duration of the spill event. Surface oil alters the ocean radar scattering properties, resulting in a difference between the measured backscatter and the backscatter that would be measured if oil were not present. Wind scatterometers infer the near-surface wind speed and direction using the wind geophysical model function (GMF) in conjunction with measured radar backscatter. The oil-altered backscatter error propagates through the wind retrieval process to create a difference in ASCAT-inferred winds and actual winds. Numerically computed vector winds from ECMWF are compared against ASCAT-inferred vector winds. The GMF is applied to ECMWF winds to create a predicted backscatter value to compare against ASCAT-measured backscatter. Large differences in wind or backscatter indicate areas of the ocean surface affected by oil. An objective function is developed to choose an appropriate threshold level to flag the oil-contaminated regions. Data from other sensors corroborate the ASCAT oil extent mapping.

Long-term trends in Arctic sea ice are of particular interest in studies of global temperature, climate change, and industrial application. This paper analyzes intra-annual and interannual trends in Ku-band backscatter over first-year (FY) and multiyear (MY) sea ice to develop a new sea-ice-type classification method. Histograms of backscatter are derived from high-resolution backscatter images created using the scatterometer image reconstruction (SIR) algorithm applied to measurements obtained by the SeaWinds instrument aboard QuikSCAT. The backscatter of FY and MY sea ice are clearly identifiable and are observed to vary seasonally. Using an average of the annual backscatter trends obtained from QuikSCAT, a classification of MY ice is obtained, which uses a time-dependent threshold value. Validation of the classification method is done using regional ice charts from the Canadian Ice Service. Differences in ice classification are found to be less than 6$\%$for the winters of 2006–2007 and 2007–2008, and the end of 2008. Anomalies in the distribution of sea ice backscatter from year to year suggest a reduction in MY ice cover between 2003 and 2009 and an approximately equivalent increase in FY ice cover.

In the summer 2009 NASA Characterization of Arctic Sea Ice Experiment (CASIE09), the microASAR, a small LFM-CW SAR, was operated on the NASA Sierra unmanned aerial system (UAS). An overview of the microASAR and its role in CASIE09 are described in [1, 2]. While the limitations in the motion measurements stored with the microASAR data during the CASIE09 mission originally precluded full motion compensation, motion data collected for other CASIE sensors can be employed to improve the SAR image focus and calibration. This paper describes the methodology developed to time-align this motion data and applies this data along with other algorithm improvements to the processing of the microASAR data. This paper also presents a concise description of the backprojection image processing used.

Although originally designed solely for wind retrieval, the QuikSCAT scatterometer has proved to be a useful tool for rain estimation as well. Resolution enhancement algorithms designed for QuikSCAT allow for ultra-high-resolution (UHR) (2.5 km) simultaneous wind and rain (SWR) retrieval. The principle advantage of UHR SWR estimation is that compared to conventional resolution, the higher resolution allows for identification of much smaller rain events and their effects on the wind field. To enable SWR retrieval, we adjust the geophysical model function to account for rain effects such as attenuation and increased backscatter due to increased surface roughness. Two possible rain models are proposed, a phenomenological rain model and an effective rain model. Both models are compared by evaluating data fit and rain estimation performance. Comparisons of a co-located data set show that QuikSCAT UHR SWR integrated rain rates are comparable to those from tropical rain measuring mission precipitation radar (TRMM PR) but have higher variance. Buoy comparisons reveal improved wind estimates in the presence of rain. The theoretic estimator bounds are compared to both the simulated estimator variance and the actual estimator variance. The estimator bounds indicate that despite high-noise levels, wind and rain information is still retrievable at UHR, although certain directions have degraded estimator bounds. Both rain models are compared to truth data and are shown to have comparable performance for most rain rates. Comparison with buoy measurements shows that in the presence of rain, QuikSCAT UHR SWR wind estimates have less bias and variability than wind-only estimates. Although QuikSCAT UHR SWR rain estimates are noisier than TRMM PR rain rates, they provide a useful rain flag for QuikSCAT winds.

While originally designed only for wind measurement, the QuikSCAT scatterometer is capable of making wind and rain estimates over the ocean. Three separate estimators are used, a wind-only estimator, a rain-only estimator, and a simultaneous wind-rain estimator. No one of the estimators is suitable under all wind and rain conditions. We therefore propose a Bayesian estimator selection technique whereby the appropriate estimator can be selected from the estimates themselves. This paper introduces the Bayes estimator selection technique and discusses its application to QuikSCAT wind and rain estimation for conventional (25-km) resolution products. Results indicate that using Bayes estimator selection can improve both the bias and mean-squared error of wind estimates in both raining and nonraining conditions, as well as provide an improved rain flag.

This paper approaches scatterometer image reconstruction as the inversion of a discrete noisy aperture-filtered sampling operation. Aperture-filtered sampling is presented and contrasted with conventional and irregular sampling. Discrete reconstruction from noise-free aperture-filtered samples is investigated and contrasted with conventional continuous reconstruction approaches. The discrete approach enables analytical treatment of the reconstruction grid resolution and the effective resolution imposed by the sampling and reconstruction operations. The noisy case is also explored. A reconstruction estimator based on maximum a posteriori (MAP) estimation is proposed to recover the conventional samples from noisy scatterometer measurements. This approach enables the scatterometer noise distribution to be appropriately accounted for in the reconstruction operation. The MAP and conventional reconstruction approaches are applied to the SeaWinds scatterometer and the Advanced Wind Scatterometer, and the effective resolution of the different methods is analyzed. The MAP approach produces results consistent with the well-established scatterometer image reconstruction (SIR) algorithm. The MAP approach significantly enhances the resolution at the expense of increased noise. Although a detailed noise-versus-resolution tradeoff analysis is beyond the scope of this paper, the new framework allows for a more general treatment than the ad hoc tuning parameters of the SIR algorithm.

This paper approaches wind field estimation from scatterometer measurements as the inversion of a noisy nonlinear sampling operation. The forward sampling model is presented and made discrete for practical purposes. Generally, the wind estimation problem is ill-posed at high resolution, which means that there are more parameters to estimate than measurements. A Bayesian approach based on maximum a posteriori (MAP) estimation is proposed to regularize the problem. This allows the simultaneous estimation of the regular samples of the high-resolution wind vector field directly from the noisy aperture-filtered backscatter σ° measurements. The MAP reconstruction approach is applied to the SeaWinds scatterometer, the examples are presented, and the results are compared to standard products. The MAP reconstruction method produces results that are consistent with standard products while preserving the higher spatial resolution information. The MAP estimates result in a similar resolution to the standard ultrahigh-resolution processing method but with a lower bias and a lower variability in the estimates.

Scatterometers have been launched primarily to measure ocean winds. The value of scatterometer data is increased by application of the SIR (Scatterometer Image Reconstruction) algorithm. The SIR algorithm enhances the effective resolution of the scatterometer data to support its use for other studies. SIR has been used successfully on several scatterometers, including QuikSCAT. In this paper, we describe how the SIR algorithm is adapted to ASCAT data. Using SZF data from ASCAT leads to the best resolution enhancement. SIR requires an estimate of the spatial extent for each measurement. We detail our method to estimate an approximate spatial response function for each ASCAT measurement. Finally, SIR parameters are tuned for use with ASCAT.

During the summer of 2009, the Characterization of Arctic Sea Ice Experiment 2009 (CASIE-09) operated a small, unmanned aircraft system (UAS) over the Arctic Ocean for a number of long-distance flights from Svalbard Island. In addition to other instruments, the UAS carried a small C-band synthetic aperture radar (SAR) known as MicroASAR to image sea ice roughness at 1 m resolution. This paper briefly describes the SAR, its role in CASIE-09, and presents sample SAR image results.

The QuikSCAT scatterometer infers wind vectors over the ocean using measurements of the surface backscatter. During rain events the QuikSCAT observations are subject to rain contamination. Three separate estimators have been developed: wind-only, simultaneous wind and rain, and rain-only, which account for rain contamination in varying degrees. This paper introduces a Bayes estimator selection technique to adaptively choose a best estimator from among the three types of estimators at each measurement location. Bayes estimator selection is introduced from a general perspective after which it is applied specifically to QuikSCAT wind and rain estimation. Bayes estimator selection is demonstrated in a case study to illustrate improvements in wind and rain estimation which can be obtained.

The ASCAT scatterometer measures the backscatter from the ocean surface with which it infers the near-surface wind vector. When rain is present in the observation area the windinduced backscatter is modified by the rain. This paper uses co-located observations from TRMM PR to model the effects of rain on the ASCAT observed backscatter. Two model types are considered, a phenomenological rain model and a lumped effect rain model which are comparable for most rain rates. For low rain events the ASCAT observed backscatter due to rain is not substantial, however for moderate to high rains the rain-induced backscatter from the ocean surface can be significant, and for extreme rain rates the atmospheric scattering and attenuation are dominant.

From 1999 to 2009, the SeaWinds scatterometer has been used to detect and track large Antarctic icebergs on a daily basis. Here, we develop an automated estimation algorithm to supplement iceberg position reports with estimates of the iceberg's major axis length, minor axis length, and angle of orientation. A maximum-likelihood objective function that relates measured backscatter to model-based simulated backscatter is developed. The utility of the estimation approach is analyzed in simulation and via a case study of iceberg A22a. Subsequent results agree with and supplement reports compiled by the United States National Ice Center.

This paper considers sampling and reconstruction theory with application to scatterometer image reconstruction. Backscatter imaging is approached as the inversion of a noisy aperture-filtered sampling operation. A reconstruction estimator based on maximum a posteriori probability (MAP) estimation is proposed to recover the conventional samples from noisy aperture-filtered samples. Examples from the SeaWinds scatterometer and the Advanced Wind Scatterometer (ASCAT) are presented.

This paper presents a method whereby the range resolution of multi-frequency-band SAR systems can be enhanced. If multiple signals are coherent and cover disjoint frequency bands, they can be combined into a single signal which can be processed using slightly modified SAR processing algorithms, resulting in an image with a range resolution enhanced by the sum of the constituent bandwidths.

The QuikSCAT enhanced (2.225-km) backscattering product is investigated for sensitivity to changes in soil moisture and its potential for spatial disaggregation of Advanced Microwave Scanning Radiometer (AMSR-E) soil moisture. Specifically, an active-passive methodology based on temporal change detection is tested using data from the 2006 National Airborne Field Experiment data set. This campaign was carried out from October 29 to November 20, 2006 in a 60 km times 40 km area of the Murrumbidgee catchment, southeast Australia. Temporal change detection analysis and accuracy in terms of spatial pattern distribution throughout the domain were assessed using a passive microwave airborne product derived from the Polarimetric L-band Multibeam Radiometer at 1-km spatial resolution. QuikSCAT-AMSR-E intercomparisons indicated higher correlations when using C-band observations. The greatest sensitivity to soil moisture was observed when using V-polarized backscatter measurement. While backscattering data showed adequate temporal sensitivity to changes in soil moisture due to precipitation events, the spatial agreement was complicated by the presence of irrigation and standing water (rice fields). This resulted in low Cramer's Phi values (less than 0.06), which were used as a measure of spatial correspondence in terms of change in soil moisture and backscatter. In addition, the high QuikSCAT sensor frequency and existence of noise in the observed data contributed to the observed discrepancies.

An experimental 2.5-km ultrahigh-resolution (UHR) wind product provided by NASA's QuikSCAT scatterometer offers the potential for new access to coastal surface wind dynamics at the mesoscale level and below. To give future users the best indication of the value of these data, the UHR wind retrievals must be fully validated in nearshore areas. Comparison with meteorological buoys and standard QuikSCAT products allows detailed investigation of UHR winds. Speed and direction residuals are calculated between all scatterometer products and collocated buoys. An ambiguity selection routine improves wind direction agreement between the UHR winds and the other products. Magnitude residuals follow the patterns of the standard QuikSCAT winds, with a 1-2 m/s positive bias in light winds (below 4 m/s) and high winds (above 16 m/s) and standard deviations consistently below 3 m/s. After application of a land contamination removal algorithm, the UHR product provides extended coverage near the coast. An example of a specific wind event illustrates the potential benefits of improved resolution measurements for examining ocean-atmosphere dynamics.

The SeaWinds scatterometer was originally designed to measure wind vectors over the ocean by exploiting the relationship between wind-induced surface roughening and the normalized radar backscatter cross section. Rain can degrade scatterometer wind estimation; however, the simultaneous wind/rain (SWR) algorithm was developed to enable SeaWinds to simultaneously retrieve wind and rain rate data. This algorithm is based on colocating data from the Precipitation Radar on the Tropical Rainfall Measuring Mission and SeaWinds on QuikSCAT. This paper develops a new wind and rain radar backscatter model for SWR using colocated data from the Advanced Microwave Scanning Radiometer (AMSR) and SeaWinds aboard the Advanced Earth Observing Satellite II. This paper accounts for rain height in the model in order to calculate surface rain rate from the integrated rain rate. The performance of SWR using the new wind/rain model is measured by comparison of wind vectors and rain rates to the previous SWR algorithm, AMSR rain rates, and National Center for Environmental Prediction numerical weather prediction winds. The new SWR algorithm produces more accurate rain estimates and improved winds, and detects rain with a low false alarm rate.

The QuikSCAT scatterometer is used to accurately retrieve winds over the ocean at both high (2.5 km) and low (25 km) resolutions. In near-coastal regions, land contamination of measurements results in inaccurate wind estimates using current techniques. Here, we show that identifying land-contaminated measurements allows wind retrieval to be accurately achieved in near-coastal regions using QuikSCAT data at up to 2.5-km resolution using the AVE algorithm. To identify and remove land contamination, two metrics are compared, namely, the minimum distance to land and the land contribution ratio (LCR). The LCR is used as a metric to identify and remove land-contaminated backscatter sigmaomeasurements before wind retrieval by discarding measurements with LCR values above a threshold. LCR thresholds used to remove land-contaminated measurements are determined using Monte Carlo simulations and set during processing using a lookup table based on the local wind speed, land brightness, and the cross-track swath location. Wind retrieval from sigmaofields generated using the LCR is more accurate closer to the coast than previously achieved using both low- and high-resolution processing.

The range-Doppler algorithm and the chirp-scaling algorithm (CSA) process synthetic aperture radar (SAR) data with approximations to ideal SAR processing. These approximations are invalid for data from systems with wide beamwidths, large bandwidths, and/or low center frequencies. While simple and efficient, these frequency-domain methods are thus limited by the SAR parameters. This paper explores these limits and proposes a generalized chirp-scaling approach for extending the utility of frequency-domain processing. We demonstrate how different order approximations of the SAR signal in the 2-D frequency domain affect image focusing for varying SAR parameters. From these results, a guideline is set forth, which suggests the required order of approximation terms for proper focusing. A proposed generalized frequency-domain processing approach is derived. This method is an efficient arbitrary-order CSA that processes the data using the appropriate number of approximation terms. The new method is demonstrated using simulated data.

The Artemis microASAR is a flexible, robust SAR system built on the successful legacy of the BYU μSAR [1] [2] and other BYU SAR systems [3]. It is an LFM-CW SAR system designed for low-power operation on small, manned aircraft or UAVs. This paper describes the high-level methodology used in designing the microASAR system and contains a description of the hardware specifications. Performance projections are also calculated and presented.

Advanced Microwave Scanning Radiometer (AMSR) brightness temperature (Tb) data are typically organized into daily average images characterized by compromised temporal resolution. An alternate approach considers the location-dependent measurement time of each sample and organizes data by local time-of-day (LTOD), producing bi-daily images with high temporal locality and reduced noise. These LTOD images show greater improvement and fewer artifacts than daily images when spatial resolution enhancement algorithms are applied. LTOD ice edges are on average 8 km closer to MODIS ice edges than daily ice edges.

This paper presents a study distilling radar backscatter measurements made by three Ku-band scatterometers of the Antarctic ice sheet over a nearly thirty-year study period. Backscatter signature modeling techniques are used to compensate for observation geometry and operating frequency differences among the sensors. Findings of this study include large regions of both positive and negative change in average backscatter, particularly in regions of West Antarctica and along much of the coast.

RADARSAT-1 ScanSAR SWA images of Hurricane Katrina are used to retrieve surface wind vectors over the ocean. Collocated H*wind wind directions are used as the wind direction while the wind speed is derived from SAR sigmadeg by inversion of a C-band HH-polarization Geophysical Model Function (GMF) that is derived from the VV-polarization GMF, CMOD5, using a polarization ratio model. Because existing polarization models do not fit the ScanSAR SWA data well, a recalibration model is proposed to "recalibrate" the ScanSAR SWA images. Validated with collocated H*wind wind speed estimates, the mean difference between SAR-retrieved and H*wind speed is small and the root mean square (RMS) error is below 4 m/s. Rain effects on the ScanSAR measurements are analyzed for an incidence angle range of 22deg to 23.6deg using collocated ground-based Doppler weather radar (NEXRAD) rain measurements.

Although originally designed solely for wind retrieval, the QuikSCAT scatterometer has also proved to be a useful tool for rain retrieval. Resolution enhancement algorithms designed for QuikSCAT allow for ultra-high-resolution (UHR) (2.5 km) simultaneous wind and rain (SWR) retrieval. To enable SWR retrieval, we adjust the geophysical model function (GMF) to account for rain effects such as attenuation or increased backscatter due to increased surface roughness. Comparisons of a co-located data set show that QuikSCAT UHR SWR rain rates are comparable to those from Tropical Rainfall Measuring Mission Precipitation Radar (TRMM PR) but have higher variance. The noise level of the QuikSCAT rain estimates can be reduced by forming the reduced resolution rain rate estimates. As expected, rain estimates are significantly worse in regions where wind dominates the backscatter.

Researchers at Brigham Young University have created an experimental 2.5 km ultra-high resolution (UHR) wind product from the QuikSCAT scatterometer. This product adds to the standard 25 km and new 12.5 km resolution data provided by that satellite, and offers the potential for new access to coastal surface wind dynamics at the sub-mesoscale level. With its nineteen meteorological buoys, the Gulf of Maine provides an excellent test site for evaluating the UHR wind retrievals. Comparison with these buoys, mesoscale meteorological model winds, and standard QuikSCAT products throughout the month of October 2006 allows detailed investigation of UHR wind speed and direction. Even with a land contamination mask, the UHR product provides extended coverage near the coast. An additional ambiguity reselection routine improves wind direction agreement between the UHR winds and the other products. With this refinement, the ultra-high resolution winds show great promise in the coastal region.

The 25 km resolution standard wind products (L2B) are available operationally in near-real time from SeaWinds on QuikSCAT. This relatively low resolution can be enhanced to yield a 2.5 km ultra-high resolution (UHR) product that can be used to identify hurricane eye centers more accurately. A comparison is made between the analyst's choice of eye location based on UHR images and interpolated best-track position. In this analysis, the UHR images are divided into two categories based on the analyst's confidence level of finding the eye center location. In each category, statistical error quantities are computed. UHR images within the high-confidence category can provide, for a given year and basin, mean error distance as small as 15 km with a 9 km standard deviation. The use of these categories may facilitate the realization of QuikSCAT's effectiveness in helping to identify and track hurricane eye centers.

QuikSCAT tracking of Antarctic icebergs is discussed, and iceberg movement trends are illustrated. Iceberg melting factors are explored and a backscatter time-series is presented. Interactions between ice-berg and sea ice are examined using QuikSCAT data. To improve QuikSCAT sea ice mapping capability, a threshold algorithm to detect polynyas is developed. The 2006-2007 Antarctic ablation season is analyzed and correlations between polynya formation and sea ice melting are explored. General sea ice melting patterns are discussed, and statistics are derived from QuikSCAT's life mission and compared with SSM/I measurements. This study finds that while average Antarctic sea ice coverage is increasing, minimum sea ice extent is decreasing.

A Bayesian method for estimating wind and rain in hurricanes from SeaWinds at ultra-high resolution is developed. We use a hurricane model to generate prior distributions for the wind speed, wind direction, and rain rate. The rain prior is derived from data from the Tropical Rainfall Measuring Mission Precipitation Radar (TRMM-PR). The new method reduces the variability of the standard simultaneous wind and rain estimates while preserving meso-scale detail.

Common methods for processing SAR data, such as the Range-Doppler Algorithm (RDA) and the Chirp Scaling Algorithm (CSA), make certain approximations about the SAR signal. As the transmit frequency drops, the bandwidth grows, and the beamwidth increases, the approximations miss important factors required to precisely process the data. This paper shows that the approximations correspond to keeping lower order terms of the expansion of the SAR transfer function. We demonstrate the limits for focusing low frequency SAR data with these approximations. Previous methods for correcting the approximation errors are shown and a new method for including an arbitrary number of terms in processing the data is discussed. The concepts presented are verified using simulated SAR data.

Methods for internal calibration of a certain class of microwave polarimetric radiometers are presented by Piepmeier. In that work, the calibration parameters are estimated algebraically. We demonstrate that Bayesian estimation decreases the root-mean-square error of the estimates by a factor of two. This improvement is obtained by using knowledge of the noise structure of the measurements and by utilizing all of the information provided by the measurements. Drawbacks are the increased complexity of the method and an increase in computation. We also extend the method to estimate several hardware component parameters of interest in system calibration.

WindSat has systematically collected the first global fully polarimetric passive microwave data over both land and ocean. As the first spaceborne polarimetric microwave radiometer, it was designed to measure ocean surface wind speed and direction by including the third and fourth Stokes parameters, which are mostly related to the asymmetric structures of the ocean surface roughness. Although designed for wind vector retrieval, WindSat data are also collected over land and ice, and this new data has revealed, for the first time, significant land signals in the third and fourth Stokes parameter channels, particularly over Greenland and the Antarctic ice sheets. The third and fourth Stokes parameters show well-defined large azimuth modulations that appear to be correlated with geophysical variations, particularly snow structure, melting, and metamorphism, and have distinct seasonal variation. The polarimetric signatures are relatively weak in the summer and are strongest around spring. This corresponds well with the formation and erosion of the sastrugi in the dry snow zone and snowmelt in the soaked zone. In this paper, we present the full polarimetric signatures obtained from WindSat over Greenland, and use a simple empirical observation model to quantify the azimuthal variations of the signatures in space and time.

Using collocated ERS scatterometer (ESCAT), Tropical Rainfall Measuring Mission (TRMM) Precipitation Radar (PR), and European Centre for Medium-Range Weather Forecasts (ECMWF) data, the effects of rain on ESCAT wind-only retrieval are evaluated. Additional scattering from rain causes estimated wind speeds to appear higher than expected. Selected directions of the rain-corrupted wind vectors are biased toward along-track directions under conditions of heavy rain, which is regardless of the true wind direction. Rain becomes more significant for data acquired at a high incidence angle. To compensate for rain-induced backscatter, a simultaneous wind/rain retrieval (SWRR) method, which simultaneously retrieves wind velocity and surface rain rate from ESCAT measurements with an incidence angle >40deg, is developed by using a C-band wind/rain backscatter model. The performance of SWRR under typical wind/rain conditions is evaluated through simulation and validation with collocated TRMM PR and ECMWF data sets. SWRR is shown to significantly improve wind velocity estimates, and the SWRR estimated surface rain rate has relatively high accuracy in moderate to heavy rain cases. Although SWRR-retrieved rain is somewhat biased, it can be corrected. We note that for ESCAT, about 1.5% of all the collocated measurements is affected by significant rain.

The growing importance of polarimetric radiometers has led to the need for a detailed theory for Stokes antenna temperatures. In this paper, we provide a full Stokes vector formulation of an antenna temperature that accounts for the entire antenna pattern, which includes polarization mixing in the main-beam and sidelobe effects. To derive the Stokes antenna temperatures, we follow the conventional methods in the Earth remote sensing literature while relying on a coherency algebra approach from radio astronomy. Connections and parallels to the conventional approaches are noted along the way. We also introduce generalizations of beam efficiency and cross polarization for use with polarimetric radiometers. These provide important metrics in the design of future systems.

Although the SeaWinds scatterometer was not specifically designed to observe tropical cyclones, new high-resolution wind products resolve much of the horizontal structure of these storms. However, these higher resolution products (2.5 km) are inherently noisier than the standard 25-km near-surface wind products. These noise levels combined with rain contamination complicate high-resolution wind estimation—particularly in tropical cyclones. Fortunately, tropical cyclones have structures that can be exploited by using a wind field model. This paper develops a new procedure for hurricane wind field estimation from the SeaWinds instrument at ultrahigh resolution. A simplified hurricane model is developed to provide prior information to be used in maximuma posterioriprobability estimation of ocean winds. Using the hurricane model ameliorates the effects of rain and noise and provides useful hurricane parameters such as the eye center location. The model also improves ambiguity selection. The new method reduces the variability of the wind speed and direction estimates, although high wind speeds still tend to be underestimated. The method also greatly improves wind direction estimates in hurricanes—even in rain-contaminated portions of the storm.

Small low-cost high-resolution synthetic aperture radar (SAR) systems are made possible by using a linear frequency-modulated continuous-wave (LFM-CW) signal. SAR processing assumes that the sensor is moving in a straight line at a constant speed, but in actuality, an unmanned aerial vehicle (UAV) or airplane will often significantly deviate from this ideal. This nonideal motion can seriously degrade the SAR image quality. In a continuous-wave system, this motion happens during the radar pulse, which means that existing motion compensation techniques that approximate the position as constant over a pulse are limited for LFM-CW SAR. Small aircraft and UAVs are particularly susceptible to atmospheric turbulence, making the need for motion compensation even greater for SARs operating on these platforms. In this paper, the LFM-CW SAR signal model is presented, and processing algorithms are discussed. The effects of nonideal motion on the SAR signal are derived, and new methods for motion correction are developed, which correct for motion during the pulse. These new motion correction algorithms are verified with simulated data and with actual data collected using the Brigham Young University muSAR system.

Using collocated ESCAT, TRMM PR, and ECMWF data, the effects of rain on the ESCAT wind-only retrieval has been evaluated. For high incidence angle measurements, the additional scattering of rain causes estimated wind speeds to appear higher than expected. It is also noted that the selected directions of the rain-corrupted wind vectors generally point along swath in heavy rain, regardless of the true wind. A simultaneous wind/rain retrieval method (SWRR) is developed using a simple wind/rain backscatter model. Validation shows that SWRR method significantly improves the wind vector estimates at high incidence angles in heavy rain cases. It also provides an estimate of the surface rain rate.

Although SeaWinds was not originally designed to observe tropical cyclones, new higher resolution products resolve much of the horizontal structure of these storms. However, these higher resolution products (reported at 2.5 km) are inherently noisier than the standard 25 km products and the high rain rates often associated with hurricanes corrupt the wind estimates. Fortunately, these storms have structure which can be exploited using a model. This paper develops a new procedure for hurricane wind field estimation from the SeaWinds instrument at ultra high resolution. We develop a simplified hurricane model to provide prior information to be used in maximum a posteriori probability (MAP) wind estimation. Using the hurricane model ameliorates the effects of rain and noise on the scatterometer measurements and directly provides useful hurricane parameters such as the eye center location and intensity.

Small, low-cost, high-resolution SAR systems, such as the Brigham Young University (BYU) muSAR, are made possible by using a linear frequency modulated continuous wave (LFM-CW) signal. SAR processing assumes that the sensor is moving in a straight line at a constant speed, but in actuality a UAV or airplane will deviate, often significantly, from this ideal. This non-ideal motion can seriously degrade the SAR image quality. In a continuous wave system this motion happens during the radar pulse which means that existing motion compensation techniques, which approximate the position as constant over a pulse, are limited for an LFM-CW SAR. In this paper, the LFM-CW SAR signal model is presented and processing algorithms are discussed. The effects of non-ideal motion during the SAR signal are derived and new methods for motion correction are developed which correct for motion during the pulse. These new motion correction algorithms are verified with simulated data and with actual data collected using the BYU muSAR system.

Yueh proposed a method of using the third Stokes parameter TUto correct brightness temperatures such as Tvand Thfor polarization rotation. This paper presents an extended error analysis of the estimation of Tv, Th, and TQequiv Tv- Thby Yueh's method. In order to carry out the analysis, we first develop a forward model of polarization rotation that accounts for the random nature of thermal radiation, receiver noise, and (to first order) calibration. Analytic formulas are then derived for the bias, standard deviation (STD), and root-mean-square error (RMSE) of estimated TQ, Tv, and Th, as functions of scene and radiometer parameters. These formulas are validated through independent calculation via Monte Carlo simulation. Examination of the formulas reveals that: 1) natural TUfrom planetary surface radiation, of the magnitude expected on Earth at L-band, has a negligible effect on correction for polarization rotation; 2) RMSE is a function of rotation angle Omega, but the value of Omega that minimizes RMSE is not known prior to instrument fabrication; and 3) if residual calibration errors can be sufficiently reduced via postlaunch calibration, then Yueh's method reduces the error incurred by polarization rotation to negligibility.

With the confirmed evidence of rain surface perturbation in recent studies, the rain effects on C-band scatterometer measurements are reevaluated. By using colocated Tropical Rainfall Measuring Mission Precipitation Radar, ESCAT on European Remote Sensing Satellites, and European Centre for Medium-Range Weather Forecasts data, we evaluate the sensitivity of C-band sigmadeg to rain. We develop a low-order wind/rain backscatter model with inputs of surface rain rate, incidence angle, wind speed, wind direction, and azimuth angle. We demonstrate that the wind/rain backscatter model is accurate enough for describing the total backscatter in raining areas with relatively low variance. We also show that the rain surface perturbation is a dominating factor of the rain-induced backscatter. Using three distinct regimes, we show under what conditions the wind, rain, and both wind and rain can be retrieved from the measurements. We find that the effect of rain has a more significant impact on the measurements at high incidence angles than at low incidence angles

Radar backscatter (sigmadeg) from ergs is modulated with view direction [incidence (thetas) and azimuth (phi) angles], where the modulation characteristics reflect the surface geometry. sigmadeg also varies spatially and reflects the spatial inhomogeneity of the sand surface. We use sigmadeg measurements at different thetas and phi angles from the NASA, European Remote Sensing satellite, and SeaWinds scatterometers to understand the relationship between wind and erg bedforms. A model incorporating the sigmadeg phi-modulation and spatial inhomogeneity is proposed. Surface slope variations are related to the sigmadeg spatial inhomogeneity. We compare the backscatter model results with numerically predicted wind direction data provided by the European Centre for Medium-Range Weather Forecasts (ECMWF) over the erg surfaces. We use the maxima of the phi-modulation at thetas=33deg to infer the orientation of the dominant slip-sides on the sand surface. These orientations are consistent with the ECMWF wind directions spatially and temporally

Abstract — Yueh [1] proposed a method of using the third Stokes parameter, TU, to correct brightness temperatures, such as Tv and Th, for polarization rotation. This paper presents an extended error analysis of the retrieval of TQ ≡ Tv − Th by Yueh’s method. Analytical formulas are derived for the bias, standard deviation, and mean-squared error (MSE) of retrieved TQ, as functions of scene and radiometer parameters. These formulas are validated through independent calculation via Monte Carlo simulation. The formulas predict several interesting effects: (a) MSE is minimized by rotating the radiometer by 45 ◦ with respect to the natural polarization basis defined by the Earth’s surface, (b) TU from planetary surface radiation (of the magnitude expected on Earth) has a negligible effect on correction for polarization rotation, and (c) three-channel polarimetric radiometry (with the radiometer rotated by 45 ◦ ) has lower MSE than conventional two-channel radiometry that suffers no polarization rotation. I.

Students at Brigham Young University have developed a new, low-cost synthetic aperture radar (SAR) system, the BYU muSAR. The simple design, based on a linear frequency modulated continuous wave signal (LFM-CW), reduces the size and power compared to a conventional pulsed SAR system. This enables the BYU muSAR to fly on a small UAV, further reducing the cost of operation and extending the use of SAR into new areas. Design parameters and specifications for the BYU muSAR are presented in this paper, together with results from experimental data collection and test flights.

Algorithms for estimating sea-ice extent from remotely sensed microwave sensor data can benefit from knowledge of the "a priori" distribution of the daily expansion and contraction of the sea-ice pack. To estimate the probability distribution of daily Antarctic sea-ice extent change, two independent sea-ice datasets are analyzed: sea-ice extent derived from the QuikSCAT scatterometer and ice concentration estimates from the Special Sensor Microwave/Imager. The daily sea-ice advance and retreat is tracked over a four-year period. The distribution of the daily sea-ice advance/retreat from each sensor is similar and is approximately double-exponential. Daily ice-pack statistics are presented.

Ku-band dual-polarization radar backscatter measurements from the SeaWinds-on-QuikSCAT scatterometer are used to determine periods of surface freeze and melt in the Antarctic ice shelves. The normalized horizontal-polarization radar backscatter (sigmao) and backscatter polarization ratio are used in maximum-likelihood estimation of the ice state. This method is used to infer the daily ice-surface conditions for 25 study locations located on the Ronne, Ross, Larsen, Amery, Shackleton, and other ice shelves. The temporal and spatial variations of the radar response are observed for various neighborhood sizes surrounding each given location during the study period. Criteria for determining the dates of melt onset and freeze-up for each Austral summer are presented. Validation of the ice-state and melt-onset date estimates is performed by analyzing the corresponding brightness temperature (Tb) measurements from Special Sensor Microwave/Imager (SSM/I) radiometers. QuikSCAT sigmaomeasurements from 1999 to 2003 are analyzed and found to be effective in determining periods of melt in Antarctic ice sheets at high temporal and spatial resolutions. These estimates can be used in studies of the climatic effects of the seasonal and interannual melting of the Antarctic ice sheets

The SeaWinds-on-QuikSCAT (QuikSCAT) and SeaWinds-on-ADEOS-2 (SeaWinds) scatterometers measure the normalized radar backscatter (/spl sigma//sup o/) of the earth's surface. These identical radar sensors are on different spaceborne platforms in similar orbits. QuikSCAT and SeaWinds data are used to infer near-surface wind vectors, polar sea-ice extent, polar-ice melt events, etc. In order to verify the relative calibration of these sensors, a simple cross calibration based on land backscatter measurements is performed. A first-order polynomial model is used to remove the incidence angle dependence of /spl sigma//sup o/ for selected regions of the Amazon rainforest and the Sahara Desert. It is shown that the two sensors are well-calibrated to each other and require no bias corrections. Additionally, evidence of a diurnal cycle in the Amazon rainforest backscatter is given.

The Ku-band SeaWinds scatterometer estimates near-surface ocean wind vectors by relating measured backscatter to a geophysical model function for the near-surface vector wind. The conventional wind retrieval algorithm does not explicitly account for SeaWinds' sensitivity to rain, resulting in rain-caused wind retrieval error. A new retrieval method, termed "simultaneous wind/rain retrieval," that estimates both wind and rain from rain-contaminated measurements has been previously proposed and validated with Tropical Rain Measuring Mission data. Here, the accuracy of rains retrieved by the new method is validated through comparison with the Next Generation Weather Radar (NEXRAD) in coastal storm events. The rains detected by both sensors are comparable, though SeaWinds-estimated rains exhibit greater variability. The performance of simultaneous wind/rain retrieval in flagging excessively rain-contaminated winds is discussed and compared to existing methods. A new rain-only retrieval algorithm for use in rain-backscatter-dominated areas is proposed and tested. A simple noise model for SeaWinds rain estimates is developed, and Monte Carlo simulation is employed to verify the model. The model shows that SeaWinds rain estimates have a standard deviation of 2.5 mm/h, which is higher than the NEXRAD measurements. Thresholding SeaWinds rain estimates at 2 mm/h yields a better rain flag than current rain flag algorithms.

A sea ice mapping algorithm for SeaWinds is developed that incorporates statistical and spatial a priori information in a modified maximum a posteriori (MAP) framework. Spatial a priori data are incorporated in the loss terms of a Bayes risk formulation. Conditional distributions and priors for sea ice and ocean statistics are represented as empirical histograms that are forced to conform to a set of expected histograms via principal component filtering. Tuning parameters for the algorithm allow adjustments in the algorithm's performance. Results of the algorithm exhibit high correlation with the Remund-Long sea ice mapping algorithm for SeaWinds and the Special Sensor Microwave/Imager National Aeronautics and Space Administration Team 30% ice edge, and are verified with RADARSAT-1 ScanSAR imagery. The resulting sea ice maps exhibit high edge detail, preserve polynyas and ice bodies disjoint from the primary ice sheet, and thus are suitable for use with wind retrieval and sea ice studies. Principles employed in the algorithm may be of interest in other classification studies.

Characterization of the microwave signature of the Greenland snow surface enables delineation of the different snow facies and is a tool for tracking the effects of climate change. A new empirical observation model is introduced that uses a limited number of parameters to characterize the snow surface based on the dependence of radar backscatter on incidence angle, azimuth angle, spatial gradient, and temporal rate of change. The individual model parameters are discussed in depth with examples using data from the NASA Scatterometer (NSCAT) and from the C-band European Remote Sensing (ERS) satellite Advanced Microwave Instrument in scatterometer mode. The contribution of each model term to the overall accuracy of the model is evaluated. The relative contributions of the modeled dependencies vary by region. Two studies illustrating applications of the model are included. First, interannual changes over the Greenland ice sheet are investigated using nine years of ERS data. Surface changes are observed as anomalies in the /spl sigma//spl deg/ model parameters. Second, intraannual variations of the surface are investigated. These changes are observed in the average backscatter normalized to a given observation geometry. The results indicate that the model can be used to obtain a more complete understanding of multiyear change and to obtain low-variance high temporal resolution observations of intraannual changes. The model may be applied for increased accuracy in scatterometer, synthetic aperture radar (SAR), and wide-angle SAR studies.

Microwave remote sensing detection of snow melt and ablation generally focuses on the detection of liquid moisture in the snow-pack. For ablation estimation, it is important to determine if wet snow is in the process of melting or freezing. The different stages of the melt cycle are observed in the diurnal variation of T/sub b/ measurements from the Special Sensor Microwave Imager (SSM/I) over Greenland. SSM/I channel ratios exhibit patterns indicating that they are sensitive to melt and freeze stages of the daily melt cycle. The horizontal to vertical polarization ratio is sensitive to surface wetness associated with melting. The 19-37-GHz frequency ratio is sensitive to a frozen surface layer over wet snow which is associated with the freeze stage of the melt cycle. These observations are supported by conceptual models presented here and in in situ measurements from other investigators.

Recent developments in reconstruction and resolution enhancement for microwave instruments suggest a possible tradeoff between computation, resolution, and downlink data rate based on postcollection reconstruction/resolution enhancement processing. The Hydrospheric State (HYDROS) mission is designed to measure global soil moisture and freeze/thaw state in support of weather and climate prediction, water, energy, and carbon cycle studies, and natural hazards monitoring. It will use an active and passive L-band microwave system that optimizes measurement accuracy, spatial resolution, and coverage. The active channels use synthetic aperture radar-type processing to achieve fine spatial resolution, requiring a relatively high downlink data rate and ground processor complexity. To support real-time applications and processing, an optional postcollection reconstruction and resolution enhancement method is investigated. With this option, much lower rate real-aperture radar data are used along with ground-based postprocessing algorithms to enhance the resolution of the observations to achieve the desired 10-km resolution. Several approaches are investigated in this paper. It is determined that a reconstruction/resolution enhancement technique combining both forward- and aft-looking measurements enables estimation of 10-km resolution or better backscatter values at acceptable accuracy. Key tradeoffs to achieve this goal are considered.

The Sahara desert includes large expanses of sand dunes called ergs. These dunes are formed and constantly reshaped by prevailing winds. Previous study shows that Saharan ergs exhibit significant radar backscatter (/spl sigma//spl deg/) modulation with azimuth angle (f). We use /spl sigma//spl deg/ measurements observed at various incidence angles and f from the NASA Scatterometer (NSCAT), the SeaWinds scatterometer, the ERS scatterometer (ESCAT), and the Tropical Rainfall Measuring Mission's Precipitation Radar to model the /spl sigma//spl deg/ response from sand dunes. Observations reveal a characteristic relationship between the backscatter modulation and the dune type, i.e., the number and orientation of the dune slopes. Sand dunes are modeled as a composite of tilted rough facets, which are characterized by a probability distribution of tilt with a mean value, and small ripples on the facet surface. The small ripples are modeled as cosinusoidal surface waves that contribute to the return signal at Bragg angles only. Longitudinal and transverse dunes are modeled with rough facets having Gaussian tilt distributions. The model results in a /spl sigma//spl deg/ response similar to NSCAT and ESCAT observations over areas of known dune types in the Sahara. The response is high at look angles equal to the mean tilts of the rough facets and is lower elsewhere. This analysis provides a unique insight into scattering by large-scale sand bedforms.

Sand seas (ergs) of the Sahara are the most dynamic parts of the desert. Aeolian erosion, transportation, and deposition continue to reshape the surface of the ergs. The large-scale features (dunes) of these bedforms reflect the characteristics of the sand and the long-term wind. Radiometric emissions from the ergs have strong dependence on the surface geometry. We model the erg surface as composed of tilted rough facets. Each facet is characterized by a tilt distribution dependent upon the surface roughness of the facet. The radiometric temperature (T/sub b/) of ergs is then the weighted sum of the T/sub b/ from all the facets. We use dual-polarization T/sub b/ measurements at 19 and 37 GHz from the Special Sensor Microwave Imager aboard the Defense Meteorological Satellite Program and the Tropical Rainfall Measuring Mission Microwave Imager to analyze the radiometric response of erg surfaces and compare them to the model results. The azimuth angle (/spl phi/) modulation of T/sub b/ is caused by the surface geometrical characteristics. It is found that longitudinal and transverse dune fields are differentiable based on their polarization difference (/spl Delta/T/sub b/) /spl phi/-modulation, which reflects type and orientation of dune facets. /spl Delta/T/sub b/ measurements at 19 and 37 GHz provide consistent results. The magnitude of /spl Delta/T/sub b/ at 37 GHz is lower than at 19 GHz due to higher attenuation. The analysis of /spl Delta/T/sub b/ over dry sand provides a unique insight into radiometric emission over ergs.

Observing near-surface ocean winds is important to the understanding of global weather patterns. Wind speed information can be retrieved from /spl sigma//spl deg/ measurements made by Jason-1 and SeaWinds on QuikSCAT. Though co-located measurements from the two satellites generally agree, significant differences between the co-located measurements are occasionally observed, particularly associated with rain events. Co-located data from TRMM PR 2A25 and ECMWF is used to study the effects of rain on Jason-1. Key observations are discussed.

Backscatter measurements from the Sea Winds on QuikSCAT scatterometer are used to determine periods of surface freeze and melt on Antarctic ice-shelves. A maximum likelihood method is used to infer the daily ice-surface conditions for various study points located on the Ronne, Ross, Larsen, Fimbul, Amery, and Shackleton ice-shelves. Criteria for determining the dates of melt-onset and freeze-up for each Austral summer are presented. Validation of the ice-state and melt-onset date estimates is performed by analyzing the corresponding brightness temperature (T/sub b/) measurements from special sensor microwave/imager (SSM/I) radiometers. QuikSCAT /spl sigma//spl deg/ measurements from 1999 through 2003 are analyzed and found to be very useful for determining periods of melt in Antarctic ice-sheets and provide high temporal and spatial resolution ice-state estimates.

The Sahara desert includes large expanses of sand dunes called ergs. These dunes are formed and constantly reshaped by prevailing winds. Previous study shows that Saharan ergs exhibit significant radar backscatter (sigmadeg) modulation with azimuth angle (Phi). We use sigmadeg measurements observed at various incidence angles (thetas) and Phi from the NASA scatterometer (NSCAT), the Seawinds scatterometer aboard QuikSCAT (QS-CAT), the ERS scatterometer (ESCAT) and the Tropical Rain Monitoring Mission's Precipitation Radar (TRMM-PR) to model the sigmadeg response from sand dunes. Sand dunes are modeled as a composite of tilted rough facets and small ripples. The dune fields are modeled as composed of many simple dunes. The sigmadeg measured by the scatterometer from (thetas, Phi) look direction is the sum of the returns from all the rough facets in the footprint. The model is applied to linear and transverse dunes with rough facets and Gaussian tilt distributions. The model results in a sigmadeg response similar to the NSCAT and ESCAT observations over areas of known dune types in the Sahara. This analysis gives a unique insight into scattering by large scale sand bedforms

The SeaWinds scatterometers onboard the QuikSCAT and the Advanced Earth Observing Satellite 2 measure ocean winds on a global scale via the relationship between the normalized radar backscattering cross section of the ocean and the vector wind. The current wind retrieval method ignores scattering and attenuation of ocean rain, which alter backscatter measurements and corrupt retrieved winds. Using a simple rain backscatter and attenuation model, two methods of improving wind estimation in the presence of rain are evaluated. First, if no suitable prior knowledge of the rain rate is available, a maximum-likelihood estimation technique is used to simultaneously retrieve the wind velocity and rain rate. Second, when a suitable outside estimate of the rain rate is available, wind retrieval is performed by correcting the wind geophysical model function for the known rain via the rain backscatter model. The new retrieval techniques are evaluated via simulation and validation with data from the National Centers for Environmental Prediction and the Tropical Rainfall Measuring Mission Precipitation Radar. The simultaneous wind/rain estimation method yields most accurate winds in the "sweet spot" of SeaWinds' swath. On the outer-beam edges of the swath, simultaneous wind/rain estimation is not usable. Wind speeds from simultaneous wind/rain retrieval are nearly unbiased for all rain rates and wind speeds, while conventionally retrieved wind speeds become increasingly biased with rain rate. A synoptic example demonstrates that the new method is capable of reducing the rain-induced wind vector error while producing a consistent (yet noisy) estimate of the rain rate.

While SeaWinds was designed to measure ocean winds, it can also measure rain over the ocean. SeaWinds on QuikSCAT active measurements of integrated columnar rain rate obtained via simultaneous wind/rain retrieval are evaluated via Monte Carlo simulation and the Crame/spl acute/r-Rao lower bound on estimate accuracy. Although sufficiently accurate in many conditions, the simultaneous wind/rain retrieval method used with SeaWinds on QuikSCAT data is ill-conditioned for certain wind directions and measurement geometries, sometimes yielding spurious rain rates in zero-rain conditions. To assess the validity of SeaWinds-derived rain rates, a simple empirically based rain thresholding scheme is presented, derived from simulated data. Thresholded QuikSCAT rain rates are compared to Tropical Rainfall Measuring Mission Microwave Imager monthly-averaged data, demonstrating good correlation for monthly-averaged data.

The results from two field experiments in the Mediterranean Sea are used to study the wind speed dependence of brightness temperature at L-band. During the EuroSTARRS airborne experiment, an L-band radiometer made measurements across a large wind speed gradient, enabling us to study this dependence at high wind speed. We compare our results with a two-scale emissivity model using several representations of the sea state spectrum. While the results are encouraging, unfortunately the accuracy of the measurements does not permit us to distinguish between the so-called twice Durden and Vesecky spectrum and the Elfouhaily spectrum above 7 m/spl middot/s/sup -1/. The effect of foam is certainly small. During the WISE 2001 field experiment carried on an oil rig, we studied this dependence at low wind speed, finding an abrupt decrease of the wind speed effect on the brightness temperature below 3 m/spl middot/s/sup -1/.

Sea-ice edge detection is an essential task at the different national ice services to secure navigation in ice-covered seas. Comparison between the Remund and Long ice mask image from enhanced-resolution QuikScat/SeaWinds (QS) products and the analyzed ice edge from high-resolution RADARSAT synthetic aperture radar has shown that the automatically determined QS ice mask underestimates the Arctic ice extent. QS data was statistically analyzed by colocating the data with ice charts around Greenland and with the National Aeronautics and Space Administration Team's Special Sensor Microwave/Imager (SSM/I) ice concentration algorithm over the whole Arctic region. All variables, i.e., the backscatter in vertical and horizontal polarization, the active polarization ratio (APR) and the daily standard deviation, are sensitive to ice types and are strongly correlated with ice concentration when the relationship is expressed in exponential form. Our study showed that the APR is especially suitable for ice-ocean separation, and a threshold of -0.02 was determined. An ice edge algorithm based on this APR threshold was developed using the other variables with conservative season-dependent thresholds to eliminate additional ocean noise. Also, the history of the ice cover is considered in order to detect single ice fields that are separated from the main Arctic pack ice. Validation with RADARSAT 1 and with the Advanced Very High Resolution Radiometer showed that the new algorithm successfully detects very low ice concentrations of about 10% during the entire year. The validity of the detected ice edge for near-real-time issues is also discussed in relation to the ice motion in the Marginal Ice Zone and the integration time necessary to produce the enhanced-resolution images. The new algorithm improves the automatic global ice edge resolution by a factor of two when compared to SSM/I products and could be used in both model initialization and data assimilation.

Capable scatterometer designs have been developed that seek improved resolution through the use of higher pulse sampling rates. The high sampling rates bring the traditional scatterometry assumption of measurement independence into question. This paper uses fundamental scattering theory to derive general expressions for correlation and covariance between scatterometer measurements and provides practical analysis using current and future instruments as examples. The paper derives expressions for the measurement variance parameter K/sub p/ when measurement correlation due to Rayleigh fading effects is present and relates K/sub p/ to the statistics of multiple pulse measurements. A function of the transmit signal modulation and receive processing, the measurement correlation is zero for nonoverlapping measurements but can become significant for overlapping measurements at high pulse sampling rates. The paper discusses the effects of correlation on the accuracy of scatterometer measurements and evaluates tradeoffs between spatial overlap, levels of additive noise, and measurement precision.

Microwave remote sensing provides an excellent means for mapping polar ice extent. In this study, a new algorithm for polar sea ice mapping is developed for use with the SeaWinds instrument. The approach utilizes a priori information within the framework of Bayes detection to produce sea ice extent maps. Statistical models for sea ice and ocean are represented in histograms which are filtered using a principal component (PC) based filtering technique. Spatial a priori information is incorporated through the loss terms associated with Bayes risk. Sea ice extent maps produced by the algorithm correlate well with the Remund-Long algorithm.

Abstract — Increasing the accuracy of melt estimation using microwave instruments by focusing on sub-day time scales is addressed. Three Special Sensor Microwave Imager (SSM/I) instruments operating simultaneously over the summer of 2002 are combined to estimate brightness temperature (Tb) over daily cycles. Normalized radar cross-section (σ o) measurements from SeaWinds on QuikSCAT are employed to provide similar estimates of σ o daily variation. The estimates are compared with automatic weather station temperatures. The combination of these measurements provides an increased understanding of the diurnal melt cycle over Greenland and estimation of the melt profile.

Azimuth modulation of the normalized radar cross- section in satellite data sets over Greenland is investigated. Data sets from the NASA Scatterometer (NSCAT) and from the European Remote Sensing Advanced Microwave Instrument (ERS) are employed. Azimuth dependence is clearly observed. The largest azimuth dependence occurs in the C-band ERS data with peak-to-peak azimuth modulations up to 3.0 dB. The Ku- band NSCAT data exhibits slightly smaller modulations of upto 2.0 dB. Azimuth modulation is largest in the lower dry snow zone for ERS and in the dry to percolation transition zone for NSCAT. The incidence angle dependence of the azimuth modulation is parameterized over the ice sheet. In general, the azimuth modulation is found to either decrease with increasing incidence angles, or be relatively independent of incidence angle. Regions of large incidence angle dependence for the azimuth modulation include the western dry snow zone for ERS and the northeast dry snow to percolation transition zone for NSCAT. The second order azimuth modulation orientation is highly correlated with wind direction. A new simple surface model is introduced to relate azimuth modulation to surface properties. Using this model, the size and orientation of surface sastrugi are estimated. I. INTRODUCTION The Greenland ice sheet is a critical area of study in esti- mating effects of global climate change. With only a limited number of in situ measurements due to the considerable effort associated with on site studies, remote sensing is an essential tool for studying the dynamics of this region. While satellite based normalized radar cross-section (σ o ) data have been used in a variety of successful studies over Greenland, past studies have generally ignored microwave measurement dependency on azimuth angle. This work is an extension of previous work found in (1). Data from the C-band European Remote Sensing Advanced Microwave Instrument in wind scatterometer mode (hereafter ERS) and the Ku-band NASA Scatterometer (NSCAT) are employed. Both sensors cover a large range of incidence angles. For NSCAT, only the vertical polarization (v-pol) measurements have sufficient azimuth coverage to be used in the study. The ERS measurements are also v-pol. The data shown herein is from the time interval Julian Day (JD) 330 to 360, 1996. This is during the winter when the Greenland surface is relatively constant. Although scarcely studied in Greenland, azimuth modula- tion is common in microwave remote sensing studies over other areas of the Earth. The primary application of azimuth modulation is in using scatterometers to measure vector wind speeds over the ocean surface. The azimuth modulation over Greenland is attributed to a complex snow surface structure created by wind driven snow deposition, aeolian transport, and the formation of wind slabs and hoar layers. The scale of the surface roughness varies from dunes on the km scale to meter scale erosional features known as sastrugi. First, the location and properties of the azimuth modulation are discussed. In particular, the incidence angle dependence of the azimuth modulation is addressed and the locality of the maximum modulation with respect to the different ice facies is addressed. Then the orientation of the azimuth modulation is presented to compare with Greenland wind models. Finally, a simple surface model is presented which relates the azimuth modulation to the geophysical properties of the snow surface.

Sea-ice motion in Antarctica is studied applying methods from computer vision and scale-space theory to a sequence of images obtained from scatterometer data. The proposed method can obtain a dense motion vector field for any specific observation scale. Spatial and temporal scales are used to focus on relevant geophysical features and events. A preprocessing stage involving spatial and temporal filtering at selected scales reduces noise and artifacts produced in the image generation phases, allowing reliable feature extraction and tracking at relevant scales. Optical flow (OF) methods provide a dense estimation of the motion field. The limitations and advantages of this approach are discussed. Optical flow sea-ice motion data are in agreement with sea winds data obtained independently and with very different methodologies. OF results are compared to data from wavelet methods.

The SeaWinds series of scatterometer instruments were designed to measure vector winds over the ocean at a nominal resolution of 25 km. However, the radar backscatter measurements from which the wind is inferred are made at finer resolution than this. Further, the spatial sampling of the backscatter measurements can support reconstruction of finer scale backscatter values which can then be used to estimate the near-surface wind vector field at high resolution (pixel spacing of 2.5 km). At this resolution, the data can be used to study coastal and mesoscale wind features such as tropical cyclones and convective storms. In this paper we use simulation and actual data to validate the accuracy of the retrieved winds and analyze the effective resolution of the wind estimates. An empirical technique to compensate for the estimate bias ism developed and applied and an improved method for ambiguity selection is developed. Comparison of the derived vector winds with a variety of other wind sensors is very encouraging.

During ONR's Shoaling Waves Experiment (SHOWEX) off the coast of North Carolina in November and December 1999, measurements of wind speed and direction as well as air and water temperatures were made using a variety of techniques. This paper shows a comparison of the measurements taken on December 3, 1999.

Saharo-Arabian deserts includes large expanses of sand dunes called ergs. These dunes are formed and constantly reshaped by prevailing winds. Previous study shows that Saharan ergs exhibit significant backscatter (σº) azimuth angle (φ) modulation. The phases of the modeled σo versus φ harmonics over the ergs are correlated to the orientation of the dune fields. Most of the vertical variation in ergs has low frequencies due to dunes (>100m) while high frequency variation is due to centimeter scale surface ripples. Although large scale features may be aperiodic, the small scale features are periodic and rapidly respond to changes in the prevailing wind which change their orientation and wavelength. In this paper, we use σo measurements from ERS scatterometer (ESCAT) and NASA scatterometer (NSCAT) to determine the surface profile statistics of the sand dunes. The total backscattering coefficient σº(T) from the sand as a function of incidence angle (θ) and φ is modeled as a sum is the contributions from surface scattering, subsurface volume scattering and subsurface bedrock scattering. The σo variations with the θ for different φ directions are observed.

SeaWinds on QuikSCAT, a spaceborne Ku-band scatterometer, estimates ocean winds via the relationship between the normalized radar backscatter and the vector wind. Scatterometer wind retrieval generates several possible wind vector solutions or ambiguities at each resolution cell, requiring a separate ambiguity selection step to give a unique solution. In processing SeaWinds on QuikSCAT data, the ambiguity selection is "nudged" or initialized using numerical weather prediction winds. We describe a sophisticated new ambiguity selection approach developed at Brigham Young University (BYU) that does not require nudging. The BYU method utilizes a low-order data-driven Karhunen-Loeve wind field model to promote self-consistency. Ambiguity selected winds from the BYU method and standard SeaWinds processing are compared over a set of 102 revs. A manual examination of the data suggests that the nonnudging BYU method selects a more self-consistent wind field in the absence of cyclonic storms. Over a set of cyclonic storm regions, BYU performs better in 9% of the cases and worse in 20% of the cases. Overall, the BYU algorithm selects 93% of the same ambiguities as the standard dataset. This comparison helps validate both nonnudging and nudging techniques and indicates that SeaWinds ambiguity selection can be generally accomplished without nudging.

Polar sea ice characteristics provide important inputs to models of several geophysical processes. Microwave scatterometers are ideal for monitoring these regions due to their sensitivity to ice properties and insensitivity to atmospheric distortions. Many forward electromagnetic scattering models have been proposed to predict the normalized radar cross section (/spl sigma//spl deg/) from sea ice characteristics. These models are based on very small scale ice features and generally assume that the region of interest is spatially homogeneous. Unfortunately, spaceborne scatterometer footprints are very large (5-50 km) and usually contain very heterogeneous mixtures of sea ice surface parameters. In this paper, we use scatterometer data in a large-scale inverse modeling experiment. Given the limited data resolution, we adopt a simple geometric optics forward-scattering model to analyze surface and volume scattering contributions to observed Ku-band signatures. A model inversion technique based on recursive optimization of an objective function is developed. The result is a least squares estimate of three surface parameters: the power reflection coefficient at nadir, the rms surface slope, and the volume scattering albedo. Simulations demonstrate the performance of the method in the presence of noise. The inverse model is implemented using Ku-band image reconstructed data collected by the National Aeronautics and Space Administration scatterometer. The results are used to analyze and interpret /spl sigma//spl deg/ phenomena occurring in the Antarctic and the Arctic.

Abstract—Conically scanning pencil-beam scatterometer systems, such as the recently launched SeaWinds radar, constitute an important class of instruments for spaceborne climate observation. In addition to ocean winds, scatterometer data are being applied to a wide range of land and cryospheric applications. A key issue for future scatterometer missions is improved spatial resolution. Pencil-beam scatterometers to date have been real-aperture systems where only range discrimination is used, resulting in a relatively coarse resolution of approximately 25 km. In this paper, the addition of Doppler discrimination techniques is proposed to meet the need for higher resolution. Here, the unique issues associated with the simultaneous application of range and Doppler processing to a conically scanning radar are addressed, and expressions for the theoretical measurement performance of such a system are derived. Important differences with side-looking imaging radars, which also may employ Doppler techniques, are highlighted. Conceptual design examples based on scatterometer missions of current interest are provided to illustrate this new high-resolution scatterometer approach. It is shown that spatial resolution of pencil-beam scatterometer systems can be improved by an order of magnitude by utilizing combined range/Doppler discrimination techniques, while maintaining the wide-swath and constant incidence angle neaaaaeded for many geophysical measurements. Index Terms—Ocean winds, radar, backscatter. I.

Development of scatterometer designs and applications requires extensive data simulation. The advancing capabilities of instruments motivates our proposal for an improved simulation model for noisy scatterometer measurements. Previous simulation models do not separately account for the two forms of random variation - signal fading and additive noise - which affect scatterometer measurements. The proposed model is able to generate data that are statistically equivalent (in a mean and variance sense) to actual instrument measurements by accounting for both variations, while maintaining ease of implementation. The model is particularly adept at handling design tradeoffs related to signal-to-noise ratios by appropriately separating fading and additive noise. Unlike previous models, the new model also can account for correlation between measurements, an issue that has recently become important.

Tabular Antarctic icebergs are regularly formed by the separation of massive sections of ice from ice shelves and glaciers. The National Ice Center (NIC) uses a variety of satellite sensors to track large Antarctic icebergs and reports iceberg position,. According to the NIC database, the number of large Antarctic icebergs has been increasing in recent years. A long term analysis of Antarctic iceberg activity based on scatterometer and radiometer data is presented. Our analysis suggests this increase is largely due to improved resources and technological advancements for iceberg tracking. Recent calving events may represent natural variability in iceberg activity. This study identifies some of the advantages and limitations of tracking icebergs using scatterometer data.

The scatterometer SeaWinds on QuikSCAT measures ocean winds via the relationship between the wind and the normalized radar backscatter cross-section (/spl sigma//spl deg/) from the ocean surface. Scattering and attenuation from falling rain droplets along with ocean surface perturbations due to rain change the backscatter signature of the waves induced by near-surface winds. A Simple model incorporates the effects of rain on ocean /spl sigma//spl deg/. Colocated data from the precipitation radar (PR) aboard the Tropical Rainfall Measuring Mission (TRMM) satellite is used to simulate the effects of rain as seen by SeaWinds. PR-derived backscatter, atmospheric rain attenuation, and rain rates are averaged over the SeaWinds footprint. The enhancement in backscatter from rain striking the ocean surface is estimated as a function of rain rate using a least-squares technique. QuikSCAT /spl sigma//spl deg/ values are simulated from the PR-derived parameters and numerical weather prediction wind data using the simple backscatter model. The simple model estimates 90% of the observed rain-contaminated QuikSCAT /spl sigma//spl deg/ values to within 3 dB.

The normalized radar backscatter (/spl sigma//sup o/) in scatterometer measurements over water is theorized to go to zero below a threshold wind speed due to insufficient friction between the wind and water to create capillary waves from which the radar signal scatters. Evidence of the threshold wind speed and a hysteresis effect have been observed in airship and wave tank data. There is additional evidence for a threshold wind speed in tower-mounted scatterometer data in an uncontrolled marine environment. In situ wind measurements and corresponding /spl sigma//sup o/ values obtained from YSCAT, an ultra-wideband scatterometer deployed on the Canada Centre for Inland Waters research tower at Lake Ontario, are used to detect and estimate the threshold wind speed. There is evidence for a detectable threshold in approximately half of the observations. The observed threshold wind speeds correspond well to theoretical threshold wind speeds.

The motion of sea-ice in Antarctica is studied using QuikSCAT scatterometer imagery using methods from computer vision and image processing such as intensity edges and optical flow (OF). Features are computed as differential invariants based on spatial and temporal derivatives at various scales. The first estimates of the motion vector field obtained through optical flow are used as the starting point for a regularization scheme that imposes constraints that bring the estimate closer to feature tracking results and observed motion. Constraints from fluid dynamics are brought in by separating the motion field into its divergence free and rotational free components with another field obtained through a convex combination of these components. The advantages of this approach are that it produces a dense motion field that can be globally processed and locally adjusted to fit data and a model. This study complements the analysis of sea ice motion by application of wavelet theory.

While originally designed for measuring vector winds over the ocean at a nominal resolution of 25 km, data from SeaWinds-on-QuikSCAT can be used to retrieve winds over the ocean at higher resolution with the aid of resolution enhancement reconstruction algorithms. This paper analyzes a technique for retrieving enhanced resolution winds at high resolution from SeaWinds data. Techniques for minimizing the required computation are described. As can be expected the wind estimates are noisier than standard resolution products. Some of the limitations of the approach and resulting products are considered. While validation of this approach to high resolution wind retrieval continues, the results thus far are very encouraging.

While originally designed for ocean wind measurement, radar scatterometers have proven to be very effective in monitoring land cover and ice conditions and in climate studies. The NASA Scatterometer Climate Record Pathfinder (SCP) project has generated a unique climate data record from the series of historic and ongoing scatterometer missions. Image data are archived from the Sea-Winds on QuikSCAT instrument, the scatterometer mode of the ERS-1/-2 AMI, the NASA Scatterometer (NSCAT), and the Seasat Scatterometer (SASS). In this paper we describe several SCP data sets and illustrate their application to studies of climate change over Greenland.

The Amazon basin represents a vast geographical zone containing large proportion of global biomass. We use the SeaWinds scatterometer (QSCAT), ERS-1/-2 scatterometer (ESCAT), NASA scatterometer (NSCAT) Seasat scatterometer (SASS), Tropical Rain Measuring Mission Precipitation Radar (TRMM-PR) and Special Sensor Microwave/Imager (SSM/I) data to study the multi-spectral microwave response of Amazon vegetation. Incidence angle signatures of combined backscatter measurements (/spl sigma//sup o/) from the scatterometers and precipitation radar indicate a good inter-calibration of the sensors. The multi-frequency signatures of both /spl sigma//sup o/ and radiometric temperature measurements (T/sub b/) from SSM/I are also studied. Temporal variability of the Amazon basin is studied using C-band ERS data and a Ku-band time series formed by SASS, NSCAT and QSCAT data. ESCAT data reveals a possible mismatch in the calibration of scatterometers between ERS-1 and ERS-2. Although the central Amazon forest represents an area of very stable radar backscatter measurements, portions of the southern region exhibit backscatter changes over the past two decades.

Saharo-Arabian deserts includes large expanses of sand dunes called ergs. These dunes are formed and constantly reshaped by prevailing winds. We use backscatter (/spl sigma//sup 0/) measurements observed at various azimuth angles from SeaWinds scatterometer (QSCAT) and ERS scatterometer (ESCAT) to determine the /spl sigma//sup 0/ azimuthal modulation over the sand dunes. A second order harmonic equation is used to model the /spl sigma//sup 0/ measurements as a function of azimuth angle and provides similar results for both sensors. Images of the magnitudes of the model harmonics show many small-scale linear features corresponding to the large dune fields in the ergs. The images of the phases of the two harmonics are coherent with the large scale topography of the area. Most ergs exhibit isophase related to the general orientation of large- and small-scale features of the sand surface.

Six months of data from the YSCAT94 experiment conducted at the CCIW WAVES research platform on Lake Ontario, Canada, are analyzed to evaluate a compound probability model. YSCAT was an ultrawideband small footprint (/spl ap/1 m) microwave scatterometer that operated at frequencies of 2-18 GHz, incidence angles from 0/spl deg/ to 60/spl deg/, both h-pol and v-pol, and which tracked the wind using simultaneous weather measurements. The probability distribution function of the measured instantaneous backscattered amplitude (p(a)) is compared to theoretical distributions developed from-the composite model and a simple wave spectrum. Model parameters of the resulting Rayleigh/generalized lognormal distribution probability density function (pdf) (C, a/sub 1/, and a/sub 2/) are derived directly from the data and are found to demonstrate relationships with wind speed, incidence angle, and radar frequency.

Arctic sea ice motion for the period from October 1999 to March 2000 derived from QuikSCAT and ocean buoy observations.Special Sensor Microwave/Imager (SSM/I) data using the wavelet analysis method agrees well with ocean buoy observations. Results from QuikSCAT and SSM/I are compatible when compared with buoy observations and complement each other. Sea ice drift merged from daily results from QuikSCAT, SSM/I, and buoy data gives more complete coverage of sea ice motion. Based on observations of six months of sea ice motion maps, the sea ice motion maps in the Arctic derived from QuikSCAT data appear to have smoother (less noisy) patterns than those from NSCAT, especially in boundary areas, possibly due to constant radar scanning incidence angle. For late summer, QuikSCAT data can provide good sea ice motion information in the Arctic as early as the beginning of September. For early summer, QuikSCAT can provide at least partial sea ice motion information until mid-June. In the Antarctic, a case study shows that sea ice motion derived from QuikSCAT data is consistent with pressure field contours.

While high resolution, regularly gridded observations are generally preferred in remote sensing, actual observations are often not evenly sampled and have lower-than-desired resolution. Hence, there is an interest in resolution enhancement and image reconstruction. This paper discusses a general theory and techniques for image reconstruction and creating enhanced resolution images from irregularly sampled data. Using irregular sampling theory, we consider how the frequency content in aperture function-attenuated sidelobes can be recovered from oversampled data using reconstruction techniques, thus taking advantage of the high frequency content of measurements made with nonideal aperture filters. We show that with minor modification, the algebraic reconstruction technique (ART) is functionally equivalent to Grochenig's (1992) irregular sampling reconstruction algorithm. Using simple Monte Carlo simulations, we compare and contrast the performance of additive ART, multiplicative ART, and the scatterometer image reconstruction (SIR) (a derivative of multiplicative ART) algorithms with and without noise. The reconstruction theory and techniques have applications with a variety of sensors and can enable enhanced resolution image production from many nonimaging sensors. The technique is illustrated with ERS-2 and SeaWinds scatterometer data.

This paper presents a new non-iterative, closed-form approximation to the maximum entropy (ME) image restoration method. A fast frequency domain implementation of this closed form approach is developed for the case of circular convolutional blur. This result dramatically reduces the computational demands compared to conventional iterative ME algorithms such as MART. Some limitations and advantages of ME restoration are investigated, including its dismal performance for high resolution restoration of decimated or randomly sampled blurred observations.

While designed for ocean observation, scatterometer and radiometer data have proven very useful in a variety of cryosphere studies. Over large regions of Antarctica, ice sheet and bedrock topography and the snow deposition, drift, and erosional environment combine to produce roughness on various scales. Roughness ranges from broad, basin-scale ice-sheet topography at /spl sim/100 km wavelengths to large, spatially coherent dune fields at /spl sim/10 km wavelength to erosional features on the meter scale known as sastrugi. These roughness scales influence the microwave backscattering and emission properties of the surface, combining to introduce azimuth-angle dependencies in the satellite observation data. In this paper, the authors explore the use of NASA scatterometer (NSCAT) data, European remote sensing (ERS) advanced microwave instrument (AMI) scatterometer mode data, and special sensor microwave/imager (SSM/I) data to study surface roughness effects in Antarctica. All three sensors provide strong evidence of azimuth modulation, which is correlated with the surface slope environment and results in a katabatic wind flow regime. Due to its broad azimuth coverage, NSCAT data appears to be the best suited for azimuth-angle observations. A simple empirical model for the azimuth variation in the radar backscatter is developed, and an algorithm for computing the parameters of the model from NSCAT data at a fine scale is presented. Results indicate relationships exist between the azimuthal variation of the data and the orientation of the surface slope and small-scale roughness relative to the sensor-look direction.

Characterizing the variability in sea ice in the polar regions is fundamental to an understanding of global climate and the geophysical processes governing climate changes. Sea ice can be grouped into a number of general classes with different characteristics. Multisensor data from NSCAT, ERS-2, and SSM/I are reconstructed into enhanced resolution imagery for use in ice-type classification. The resulting twelve-dimensional data set is linearly transformed through principal component analysis to reduce data dimensionality and noise levels. An iterative statistical data segmentation algorithm is developed using maximum likelihood (ML) and maximum a posteriori (MAP) techniques. For a given ice type, the conditional probability distributions of observed vectors are assumed to be Gaussian. The cluster centroids, covariance matrices, and a priori distributions are estimated from the classification of a previous temporal image set. An initial classification is produced using centroid training data and a weighted nearest-neighbor classifier. Though validation is limited, the algorithm results in an ice classification that is judged to be superior to a conventional k-means approach.

The SeaWinds scatterometer was launched on the NASA QuikSCAT spacecraft in June 1999 and is planned for the Japanese ADEOS-II mission in 2000. In addition to generating a global Ku-band backscatter data set useful for a variety of climate studies, these flights will provide ocean-surface wind estimates for use in operational weather forecasting. SeaWinds employs a compact "pencil-beam" design rather than the "fan-beam" approach previously used with SASS on Seasat, NSCAT on ADEOS-I, and the AMI scatterometer on ERS-1, 2. As originally envisioned and reported, the resolution of the SeaWinds backscatter measurements were to be antenna-beamwidth limited. In order to satisfy an emerging demand for higher resolution backscatter data, however, the SeaWinds signal-processing design has been significantly modified. Here, the various options considered for improving the resolution of the SeaWinds measurements are discussed, and the selected hardware modification (the addition of deramp processing for range discrimination) is described. The radar equation specific to a rotating pencil-beam scatterometer with digital range filtering is developed, and the new challenges associated with calibrating the resulting improved resolution measurements are discussed. A formulation for assessing the variance of the measurements due to fading and thermal noise is presented. Finally, the utility of improved resolution SeaWinds measurements for land and ice studies is demonstrated by simulated enhanced-resolution imaging of a synthetic Earth backscatter scene.

Though designed to measure vector winds over the ocean, new imaging techniques facilitate the use of NASA scatterometer data (NSCAT) in cryosphere studies. NSCAT provides data of unprecedented coverage, resolution, and quality which, when coupled with the scatterometer image reconstruction with filtering (SIRF) algorithm, enables images of /spl sigma//sup o/ at resolutions approaching 8 km over stationary targets. Such images are useful in ice mapping and classification, and multidecadal studies are possible by comparison with Seasat Scatterometer (SASS) data. The utility of NSCAT data in polar ice studies is illustrated through a review of two cryosphere applications of NSCAT data: (1) sea-ice mapping and tracking and (2) ice-sheet change in Greenland and Antarctica. The wide swath and frequent revisit, coupled with incidence and azimuth angle diversity makes NSCAT data very effective in mapping the extent of sea-ice. In Greenland, snow and ice "facies" are clearly delineated on the basis of the seasonally dependent radar backscattering cross section, due to sensitivity of radar backscatter to diagenetic changes occurring at and beneath the surface. Comparison of NSCAT and SASS data enables study of change in Greenland between 1978 and 1996.

Microwave radiometers, while traditionally utilized in atmospheric and oceanic studies, can also be used in land surface applications. However, the problem of undesirable atmospheric effects caused by clouds and precipitation must be addressed. In this paper, temporal composite surface brightness images are generated from special sensor microwave/imager (SSM/I) data with the aid of new algorithms to eliminate small-scale distortion caused by clouds or precipitation. Mean, second-highest value, modified maximum average (MMA), and hybrid compositing algorithms are compared. The effectiveness of each algorithm is illustrated through simulation and real data distribution analysis. The results show that the second-highest value algorithm is biased high. MMA provides a more accurate brightness temperature estimate in areas of atmospheric distortion, while the mean is superior in regions with little or no distortion. A hybrid algorithm is developed that is a combination of MMA and mean. It utilizes the strengths of both to create a superior algorithm for regions with varying levels of distortion. Uses of composite images produced by these algorithms include studies of vegetation change, land cover classification, and surface parameter extraction.

A wind scatterometer makes measurements of the normalized radar-backscatter coefficient /spl sigma//spl deg/ of the ocean surface. To retrieve the wind, a geophysical model function (GMF), which relates /spl sigma//spl deg/ to the near-surface wind, is used. The wind vector can be estimated using maximum-likelihood techniques from several /spl sigma//spl deg/ measurements made at different azimuth angles. The probability density of the measured /spl sigma//spl deg/ is assumed to be Gaussian with a variance that depends on the true /spl sigma//spl deg/ and therefore, depends on the wind through the GMF. With this model for wind estimation, the Cramer-Rao (C-R) bound is derived for wind estimation, and its implications for wind retrieval are discussed. As part of this discussion, the role of geophysical modeling error is considered and shown to play a significant role in the performance of near-surface wind estimates. The C-R bound is illustrated using parameters from the ERS AMI, NSCAT, and Sea Winds scatterometers.

Scatterometer instruments are active microwave sensors that transmit a series of microwave pulses and measure the returned echo power to determine the normalized radar backscattering cross section (sigma-0) of the ocean surface from which the speed and direction of near-surface ocean winds are derived. The NASA Scatterometer (NSCAT) was launched on board the ADEOS spacecraft in August 1996 and returned ten months of high-quality data before the failure of the ADEOS spacecraft terminated the data stream in June 1997. The purpose of this paper is to provide an overview of the NSCAT instrument and sigma-0 computation and to describe the process and the results of an intensive postlaunch verification, calibration, and validation effort. This process encompassed the functional and performance verification of the flight instrument, the sigma-0 computation algorithms, the science data processing system, and the analysis of the sigma-0 and wind products. The calibration process included the radiometric calibration of NSCAT using both engineering telemetry and science data and the radiometric beam balance of all eight antenna beams using both open ocean and uniform land targets. Finally, brief summaries of the construction of the NSCAT geophysical model function and the verification and validation of the wind products will be presented. The key results of this paper are as follows: The NSCAT instrument was shown to function properly and all functional parameters were within their predicted ranges. The instrument electronics subsystems were very stable and all of the key parameters, such as transmit power, receiver gain, and bandpass filter responses, were shown to be stable to within 0.1 dB. The science data processing system was thoroughly verified and the sigma-0 computation error was shown to be less than 0.1 dB. All eight antenna beams were radiometrically balanced, using natural targets, to an estimated accuracy of about 0.3 dB.

One of the limitations in using Special Sensor Microwave/Imager (SSM/I) data for land and vegetation studies is the relatively low-spatial resolution. To ameliorate this limitation, resolution-enhancement algorithms can be applied to the data. In this paper, the Backus-Gilbert inversion (BGI) technique and the scatterometer image-reconstruction (SIR) algorithm are investigated as possible methods for creating enhanced resolution images from SSM/I data. The two algorithms are compared via both the simulation and the actual SSM/I data. The algorithms offer similar resolution enhancement, though SIR requires significantly less computation. Sample results over two land regions of South America are presented.

In a continuing evaluation of the ERS-1 C-band scatterometer as a tool for studying polar sea ice, the authors evaluate the azimuthal modulation characteristics of Antarctic sea ice. ERS-1 AMI scatterometer mode data sets from several study regions dispersed in the Antarctic seasonal sea ice pack are evaluated for azimuthal modulation. When appropriate, the incidence angle dependence is estimated and removed in a study region before determining whether azimuthal modulation is present in the data. Other comparisons are made using the fore and aft beam measurement difference. The results show that over the ice pack, azimuthal modulation is less than 1 dB at the scale of observation of the ERS-1 C-band scatterometer.

Scatterometers are remote sensing radars designed to measure near-surface winds over the ocean. The difficulties of accommodating traditional fan-beam scatterometers on spacecraft has lead to the development of a scanning pencil-beam instrument known as SeaWinds. SeaWinds will be part of the Japanese Advanced Earth Observing Satellite II (ADEOS-II) to be launched in 1999. To analyze the performance of the SeaWinds design, a new expression for the measurement accuracy of a pencil-beam system is required. In this paper the authors derive a general expression for the backscatter measurement accuracy for a pencil-beam scatterometer which includes the effects of transmit signal modulation with simple power detection. Both separate and simultaneous signal+noise and noise-only measurements are considered. The utility of the new expression for scatterometer design tradeoffs is demonstrated using a simplified geometry. A separate paper, ibid., 1997, describes detailed tradeoffs made to develop the SeaWinds design.

SeaWinds is a spaceborne wind scatterometer to be flown on the second Japanese Advanced Earth Observing Satellite (ADEOS-II) in 1999. An important international element of NASA's Earth Observing System (EOS), SeaWinds is an advanced follow-on to the NASA scatterometer (NSCAT) on the first ADEOS platform. Unlike previous operational spaceborne scatterometer systems. SeaWinds employs a scanning "pencil-beam" antenna rather than a "fan-beam" antenna, making the instrument more compact and yielding greater ocean coverage. The goals of this paper are twofold. First, the overall SeaWinds functional design and backscatter measurement approach are described, and the relative advantages of the pencil-beam technique are outlined. Second, the unique aspects of measurement accuracy optimization and signal processing for the SeaWinds instrument are discussed. Applying the results of a separate companion paper, ibid., 1997, a technique to significantly improve measurement accuracy by modulating the transmit pulse is described. Trade-offs to optimize the transmit modulation bandwidth are presented.

Measurements of the normalized radar cross section (/spl sigma//spl deg/) made by the YSCAT ultrawideband scatterometer during an extended deployment on the Canada Centre for Inland Waters (CCIW) Research Tower located at Lake Ontario are analyzed and compared with anemometer wind measurements to study the sensitivity of /spl sigma//spl deg/ to the wind speed as a function of the Bragg wavelength. This paper concentrates on upwind and downwind azimuth angles in the wind speed range of 4.5-12 m/s. While YSCAT collected measurements of /spl sigma//spl deg/ at a variety of frequencies and incidence angles, this paper focuses on frequencies of 2.0, 3.05, 5.30, 10.02, and 14.0 GHz and incidence angles within the Bragg regime, 30-50/spl deg/. Adopting a power law model to describe the relationship between /spl sigma//spl deg/ and wind speed, both wind speed exponents and upwind/downwind (u/d) ratios of /spl sigma//spl deg/ are found using least squares linear regression. The analysis of the wind speed exponents and u/d ratios show that shorter Bragg wavelengths (/spl Lambda/<4 cm) are the most sensitive to wind speed and direction. Additionally, vertical polarization (V-pol) /spl sigma//spl deg/ is shown to be more sensitive to wind speed than horizontal polarization (H-pol) /spl sigma/, while the H-pol u/d ratio is larger than the V-pol u/d ratio.

Wind scatterometers are radar systems designed specifically to measure the normalized radar backscatter coefficient (/spl sigma//spl deg/) of the ocean's surface in order to determine the near-surface wind vector. Postlaunch calibration of a wind scatterometer can be performed with an extended-area natural target such as the Amazon tropical rain forest. Rain forests exhibit a remarkably high degree of homogeneity in their radar response over a very large area though some spatial and temporal variability exist. The authors present a simple technique for calibrating scatterometer data using tropical rain forests, Using a polynomial model for the rolloff of /spl sigma//spl deg/ with incidence angle, the technique determines gain corrections to ensure consistency between different antennas and processing channels. Corrections for the time varying instrument gain are made consistent with a seasonally fixed rain forest response; however, without ground stations or aircraft flights, it is difficult to uniquely distinguish between seasonal variations in the rain forest and slow variations of the system gain. Applying the corrections, the intrinsic variability of the /spl sigma//spl deg/ of the rain forest is estimated to be /spl plusmn/0.15 dB, which is the limit of the accuracy of calibration using the rain forest. The technique is illustrated with Seasat scatterometer (SASS) data and applied to ERS-1 Active Microwave Instrument scatterometer (Escat) data. Gain corrections of up to several tenths of a decibel are estimated for SASS. Corrections for Escat data are found to be very small, suggesting that Escat data is well calibrated.

The Seasat-A scatterometer (SASS) was designed to measure the near-surface wind field over the ocean by inferring the wind from measurements of the surface radar backscatter. While backscatter measurements were also made over land, they have been primarily used for the calibration of the instrument. This has been due in part to the low resolution of the scatterometer measurements (nominally 50 km). In a separate paper the present authors introduced a new method for generating enhanced resolution radar measurements of the Earth's surface using spaceborne scatterometry. In the present paper, the method is used with SASS data to study vegetation classification over the extended Amazon basin using the resulting medium-scale radar images. The remarkable correlation between the Ku-band radar images and vegetation formations is explored, and the results of several successful experiments to classify the general vegetation classes using the image data are presented. The results demonstrate the utility of medium-scale radar imagery in the study of tropical vegetation and permit utilization of both historic and contemporary scatterometer data for studies of global change. Because the scatterometer provides frequent, wide-area coverage at a variety of incidence angles, it can supplement higher resolution instruments which often have narrow swaths with limited coverage and incidence angle diversity.>

A method for generating enhanced resolution radar images of the Earth's surface using spaceborne scatterometry is presented. The technique is based on an image reconstruction technique that takes advantage of the spatial overlap in scatterometer measurements made at different times to provide enhanced imaging resolution. The reconstruction algorithm is described, and the technique is demonstrated using both simulated and actual Seasat-A Scatterometer (SASS) measurements. The technique can also be used with ERS-1 scatterometer data. The SASS-derived images, which have approximately 4-km resolution, illustrate the resolution enhancement capability of the technique, which permits utilization of both historic and contemporary scatterometer data for medium-scale monitoring of vegetation and polar ice. The tradeoff between imaging noise and resolution inherent in the technique is discussed.>

The minimisation of the objective function is a common requirement in signal processing. The author addresses the problem of finding all of the global minima of a real-valued objective function f which can be expressed as a polynomial over a finite interval. The approach is based on generating a modified Sturm sequence from f-z where z is the value of f at the global minima. The technique provides the number of the global minima of f, the value of z and the locations of the global minima of f. A simple numerical example is given. While the author concentrates on the one-dimensional case, the extension of the method to multidimensional objective functions is also discussed.>

Scatterometry and scatterometer design issues are reviewed. The design of the NASA Scatterometer (NSCAT) to be flown on the Japanese ADEOS mission is presented. Building on Seasat experience, the NSCAT system includes several enhancements, such as three antenna azimuths in each of two swaths, and an onboard digital Doppler processor to allow backscatter measurements to be colocated everywhere within the orbit. These enhancements will greatly increase the quality of the NSCAT wind data. The ground processing of data is discussed, and scatterometers of the next decade are briefly described.>

The problem of identifiability of a wind vector that is estimated from wind scatterometer measurements of the radar backscatter of the ocean's surface is addressed. The traditional wind estimation approach produces multiple estimates of the wind direction. A second processing step, known as dealiasing or ambiguity removal, is used to select a single wind estimate from these multiple solutions. Dealiasing is typically based on various ad hoc considerations. The traditional wind estimation approach results in multiple solutions associated with local minima in an objective function formed from the noisy backscatter measurements. The authors discuss the question of the uniqueness of the wind vector estimates resulting from this intuitive approach.>

A description is given of the baseline NSCAT (the NASA scatterometer) ambiguity removal algorithm and the method used to select the set of optimum parameter values. An extensive simulation of the NSCAT instrument and ground data processor provides a means of testing the resulting tuned algorithm. This simulation generates the ambiguous wind-field vectors expected from the instrument as it orbits over a set of realistic mesoscale wind fields. The ambiguous wind field is then de-aliased using the median-filter-based ambiguity removal algorithm. Performance is measured by comparison of the selected wind fields with the true wind fields. Results have shown that this median-filter-based ambiguity removal algorithm satisfies NSCAT mission requirements, and it therefore has been incorporated into the baseline geophysical data-processing system for NSCAT.>

A direct relationship between a multidimensional time series with finite support and its unwrapped phase is shown. This relationship shows that the unwrapped phase of a multidimensional sequence is unique in the sense that once the phase at the origin is specified the phase everywhere in the frequency domain follows. Additionally, the uniqueness of the unwrapped phase for multidimensional sequences which have a rational Z transform is shown. In either case, the unwrapped phase at a given point is shown to be compatible using a real 1-d finite-length phase unwrapping procedure based on Sturm sequence polynomials.>

A direct relationship between a one-dimensional time series and its unwrapped phase was shown by R. McGowan and R. Kuc (1982). They proposed an algorithm for computing the unwrapped phase by counting the number of sign changes in a Sturm sequence generated from the real and imaginary parts of the DFT (discrete Fourier transform). Their algorithm is limited to relatively short sequences by numerical accuracy. An extension of their algorithm is proposed which, by using all-integer arithmetic, permits exact computation of the number of multiples of pi which must be added to the principal value of the phase to uniquely give the unwrapped phase of a one-dimensional rational-valued finite-length sequence of arbitrary length. This extended algorithm should be of interest when highly accurate phase unwrapping is required.>