Tropical cyclones wind retrievals from satellite observations are typically associated with long revisit intervals, that limits the systems operational use especially in case of rapid intensification rates of the cyclones. Here, we use two observational satellite datasets in order to explore the spatiotemporal range of wind speed stability for possible fusion of data from different sensors. On one hand the CYGNSS platforms ensure relatively short revisit times, while on the other hand imaging radars, such as the Sentinel-1, provide high resolution images at longer revisit times. Results provide some preliminary valuable information on wind data variability in the cyclone moderate wind zone.

Flooding is one of the deadliest and costliest natural disasters. Climate change-induced flooding events are increasing worldwide, disproportionately impacting low-income and developing communities. While early warning systems save lives, satellite-based observation systems are vital for the disaster relief and recovery phases. Current satellite-based operational flood products are largely based on either optical remote-sensing methods, which exhibit a limited ability to detect water through clouds and vegetation, or microwave remote sensing, which provides relatively low spatial and temporal resolution. New small satellite constellations using radar or GNSS reflectometry (GNSS-R) have been shown to enhance our ability to overcome these deficiencies. In this work, we quantify the performance of using GNSS-R measurements from the NASA Cyclone Global Navigation Satellite Systems (CYGNSS) satellite constellation to map surface water in South Sudan and the Sudd wetland in comparison with a set of representative operational products. We make quantitative comparisons of our results with operational flood products based on Visible Infrared Imaging Radiometer Suite (VIIRS) and MODIS and with C-band Sentinel-1 synthetic aperture radar. We find that our method detects 35.4% more surface water than Sentinel-1, while the VIIRS- and MODIS-based products underestimate by 4.8% and 83.7%, respectively. We use several metrics commonly used to evaluate classification performance: precision, true positive rate (TPR), true negative rate (TNR), F2-score, and the Matthews correlation coefficient (MCC) and assess the comparisons in this statistical framework. We discuss the consequences of our findings, including ways CYGNSS data may enhance current flood products and assist decision-makers and emergency managers.

When GNSS signals reflect off of the surfaces of lakes, rivers, wetlands, and other inland water bodies, the surfaces are often sufficiently smooth to produce coherent reflections. The observable produced from coherent reflections made by GNSS Reflectometry (GNSS-R) instruments exhibits particular features with respect to diffusely scattered signals by rough land and wind-driven oceans allowing detection of such smooth bodies. Several different GNSS-R coherence detection approaches have been reported in the literature and developed among the GNSS-R community over the last several years; however, the merits of each approach are difficult to compare because they are often applied to different scenarios and quantified in different ways, independently of each other. This paper provides a unified comparison of a wide variety of different GNSS-R coherence detection approaches, which is the most extensive published to date. The approaches are applied to a common data set from the NASA CYGNSS satellites that includes both the standard Level-1 DDM science product as well as raw baseband signal recordings. Additionally, simulated observables are generated with varying coherent and non-coherent reflection components to exercise algorithms over a wide range of SNRs and relative powers. Objective measures of accuracy are used to quantify the performance of each approach in the context of relative implementation complexity. Conclusions are presented on the pros/cons of the various methods as they relate to various applications such as real-time in-orbit coherence detection or post-processing on the ground.

GNSS-R measurements from inland waters and wetlands show strong coherent reflections. These measurements can be utilized to detect the presence of surface water and to measure its extent; however, there are numerous geophysical phenomena may affect the received signal properties, such as wind induced waves and vegetation attenuation of both the electromagnetic signal and the surface water waves. Under-standing the impact of each of these geophysical effects in the inland water scene is useful for developing and assessing the capabilities of GNSS- R retrieval algorithms for detection and monitoring of dynamically changing inland water scenes. This paper proposes a combination of models to characterize these effects on the coherent GNSS reflection to a first order.

As climate change-induced global flooding increases in both frequency and magnitude, having accurate and timely flood maps becomes essential for humanitarian and future flood mitigation efforts. The Dartmouth Flood Observatory (DFO) aids humanitarian organizations and inundation observation research efforts through its archive of historical flood events extending back through 1985, as well as by providing current daily flood maps derived from a combination of observations, and also precipitation-based modeling products. Both could benefit from the addition of microwave observations that penetrate through clouds, rain, and vegetation, such as GNSS-R data now becoming available on a daily basis. In this work, we discuss a current collaboration to combine CYGNSS data with operational MODIS flood maps and evaluate the expected benefits for an example scenario over the recent anomalous flooding in South Sudan.

While CYGNSS exhibits a high revisit rate, the sparse, quasi-random tracks make it challenging to utilize data at short timescales for inundation mapping. In order to make use of the high revisit rate of CYGNSS, we implement a model-driven approach in which we compare simulated CYGNSS measurements over a binary water mask with actual CYGNSS measurements. We consider a simple water body like a reser-voir that has limited or no vegetation and exhibits dynamic changes in surface water extent within the CYGNSS mission timeframe. We find that our method has the potential to reliably detect changes in surface water extent down to 4 sq. km., corresponding to a 3% increase. Using simple, non-physics-based flood model outputs as inputs to the End-to-End Simulator (E2ES) can lead to the ability to estimate surface water extent over a given area using sparse CYGNSS measurements on sub-weekly timescales.

A novel metric for detecting coherence in global navigation satellite system reflectometry (GNSS-R) signals is presented and evaluated. It applies the Von Neumann information entropy metric for density matrices, a powerful indicator of the degree of mixing between states, coherent and incoherent, of the scene under investigation. The metric is applied to a set of raw IF data acquired by the cyclone global navigation satellite system (CYGNSS) observatories over Lake Okeechobee FL, in order to test the sensitivity of the entropy to different land cover types, including wetlands and open water. Visual comparison of results with Sentinel-1 images provides a first step in the validation of the effectiveness of entropy in detecting the presence of water covered by emergent vegetation. In addition, the entropy-based metric could be implemented on future space-based GNSS-R receivers to adapt the incoherent integration times to the observed scene, thus achieving an improvement in along-track resolution.

The sensitivity of remote sensing instruments for measuring inundation extent can vary widely. Many sensors are suitable for accurate delineation of open water extent, but in vegetated environments the vegetation canopy can obscure the presence of standing water from detection. Detecting inundation extent in these vegetated environments is especially critical for identifying flooding extent where excess surface water extends into the forests surrounding lakes and streams. In addition, cloud cover can impede timely acquisition of imagery by optical sensors. Here, we examine sensitivity of L-band Global Navigation Satellite Systems Reflectometry (GNSS-R) to flooded conditions relative to the well-known signatures of inundation by L-band SAR, and confirm that there is noticeable sensitivity of GNSS reflected signal to inundated areas, including wetlands covered by vegetation, captured by the strong response of the specular reflection by the underlying water surface.

Quantitative observations of dynamic changes in water extent and depth of the world's wetlands are currently limited by traditional remote sensing methods, which have difficulty observing surface water beneath dense vegetation and clouds. A novel remote sensing technique known as GNSS Reflectometry (GNSS-R) has shown great potential in the detection of terrestrial surface water beneath vegetation. The Cyclone Global Navigation Satellite System (CYGNSS) is a GNSS-R small satellite constellation that exhibits sub-daily revisit rates over tropical wetlands. In this work, we present a retrieval algorithm to predict water depth and surface water extent using CYGNSS observations over the Everglades. We test our algorithm over three regional approaches and varying smoothing filters. Results indicate that CYGNSS signal-to-noise ratio (SNR) is highly correlated with both water depth and extent over shallow, vegetated water.

GNSS Reflectometry (GNSS-R) measurements are very sensitive to the presence of inland waters such as wetlands, floods, rivers and lakes. This paper reviews the basic characteristics of a GNSS-R ‘water detection’ research product, including resolution and temporal sampling of wetlands, and discusses the main known sources of errors. Additionally, a summary of GNSS-R applicability to the study of lakes is provided.

This work demonstrates the creation of dynamic inland water body masks at spatial resolutions ranging from 1 to 3 km through the use of a recently developed coherence detector for the delay-Doppler maps produced by the cyclone global navigation satellite system (CYGNSS) constellation. The use of the coherence of the observed measurements reduces many of the uncertainties associated with previous signal-to-noise ratio-based water body detection approaches for CYGNSS. Using data from January 2018 to February 2020 and producing maps representing time intervals ranging from 3 months to 2 years, the water body masks created are found to be associated with a probability of detection that exceeds 80% as compared to the Pekel water mask developed from Landsat observations. The analysis presented in this work highlights the potential of using spaceborne Global Navigation Satellite Systems Reflectometry (GNSS-R) systems for dynamic inland water body mapping.

The accuracy of spaceborne ocean surface altimetry depends on precise knowledge of the altimeter satellite orbit. Previous studies of the potential utility of CYGNSS GNSS Reflectometry (GNSS-R) measurements for ocean altimetry have identified its orbit error as a limiting factor. To address this, a recent firmware upgrade to the CYGNSS satellites has enabled the downlink of additional GPS raw measurements from the navigation receiver and increased numerical precision of open-loop GNSS-R tracking information. In this paper, we present improvements to the accuracy of both the orbit determination and the ocean surface height retrievals as a result of these recent upgrades. JPL's GipsyX software is used to process 170 days of GPS navigation measurements for one of the 8 CYGNSS satellites. Incorporating high fidelity dynamic models and antenna group delay corrections, daily orbit overlaps throughout the period show mean RMS differences of 2.5 cm in height, 5.9 cm in cross-track, and 10 cm in along-track. Applying these new orbits reduced the standard deviation of retrieved sea surface height anomalies from 2.2 m to 1.9 m using 4 second smoothed measurements. Further improvements are expected using new ionospheric corrections and re-tracking methods currently in development.

The Cyclone Global Navigation Satellite System (CYGNSS) has recently shown exciting potential for GNSS reflectometry (GNSS-R) to resolve small-scale and dynamic hydrological features over land (such as rivers, lakes, wetlands, and urban flooding), even when obstructed by dense vegetation. Since CYGNSS is a small satellite constellation, sub-daily measurement frequencies provide a unique opportunity to observe short timescale changes. However, since CYGNSS observations occur in sparse, quasi-random tracks, it is more difficult to understand the true observability of events as compared to imaging instruments. Changes in SNR that would indicate a dynamic change in the scene are confounded by inherent variability due to other sources, including vegetation, geometry changes, instrument gain calibration, and surface roughness due to wind. While the literature does detail the use of CYGNSS measurements to map surface water changes, the lack of valid ground data makes it difficult to quantify the true accuracy. Here, we present results from a simulation study currently underway that utilizes a GNSS-R coherent scattering model to understand the observability and accuracy of CYGNSS measurements of dynamic changes in inland water bodies. The goal is to quantify the sensitivity and resolution of observations of dynamic spatial and temporal variations of inland water bodies.

Spaceborne GNSS Reflectometry (GNSS-R) measurements over inland waters have exhibited strong coherent scattering. The strong reflected signal from a relatively small spatial extent (several km) is highly sensitive to surface waves. This sensitivity can be leveraged to estimate wave height profiles across inland waters. Coupled with a wind wave model, retrievals of wind vector can be performed using a forward model approach. The surface waves play a significant role in nearshore ecosystems, affecting sediment resuspension, biomass production, and fish habitat, among others. This paper details a novel approach to estimating surface wave profiles and wind vectors using the passive, bistatic radar receiver aboard CYGNSS. The first ever retrieval of wind vector and wave height of an inland water body using spaceborne GNSS-R will be shown using raw signals recorded onboard CYGNSS.

The enduring activity of CYGNSS and TDS-1 observatories has provided clear evidence that ground reflected GNSS signal can be exploited for mapping of geophysical quantities in climate and global monitoring applications. One main challenge is to find out useful variables for determining the level of coherence of the scattered wave in the presence of complex natural landscapes such as water basins, river floods, mixed ice and water. In this study, we approach the concept of wave coherence by analyzing the persistence of signal energy across principal directions, determined via generalized eigenvalue decomposition of the DDM delay waveform correlation matrix. An example shows that the dimensional spread of the eigenvalues is finely sensitive to coherence of the ground reflected wave.

Spaceborne GNSS reflectometry has shown ability to observe the global inland water distribution. It allows generating dynamic maps of rivers, wetlands and inundations using the large variation in the received power while the ground track crosses those objects. Here, we analyze the change of the reflected signal's power and coherence due to surface roughness and cover (ice) observed in the CYGNSS delay-Doppler maps for Qinghai Lake, China. It is shown that a significant attenuation of the reflected coherent signal and emergence of the diffuse component occurs due to wind-generated surface waves. The comparable attenuation effect of the coherent component is observed in the case of the frozen lake. It takes place due to the lower value of the average Fresnel reflection coefficient of ice. A similar effect may occur for reflections from wetlands, when open water and water covered by vegetation could produce comparable reflected powers, however, due to different mechanisms. Understanding the relative scattering characteristics of wetlands in contrast with rougher bodies such as lakes is important for the development of algorithms to detect their dynamic changes.

Ocean surface altimetry with GNSS reflectometry (GNSS-R) has the potential to improve the observation and retrieval of mesoscale ocean current flows by increasing the spatial density of ocean surface observations. The leading challenges in current GNSS-R altimetry investigations include precise delay retracking, correction of ionospheric effects, and spacecraft receiver positioning. Here, we present improved methods to account for these challenges. A reflection-model-based approach is presented for delay retracking that uses simulated delay-Doppler maps (DDMs) to retrieve the specular delay from measured DDMs. Global ionosphere maps are used to estimate the group delay effect along the direct and reflection paths. Precise orbit determination techniques are used to improve estimates of the receiving spacecraft position and reduce systematic intersatellite biases. We analyze altimetric height retrievals from two cases, strongly coherent or diffuse reflections. We study the residual systematic error signals with a focus on understanding the errors identified above. The strong coherent observations have narrow, symmetric delay waveforms and provide higher than average signal-to-noise ratio. The DDMs of diffuse observations, originate from relatively rough ocean surfaces and have the characteristic horse-shoe shape. Only measurements from the ocean and seas surrounding Indonesia are considered as this area provides a significant number of strongly coherent reflections. More than 50 000 single-point observations and week-long averaged results are presented to quantify the noise and systematic characteristics of the retrieved surface heights.

This paper presents recent work being done to produce a new GNSS Reflectometry (GNSS-R) ocean altimetry product using data from NASA's CYGNSS Mission, a constellation of 8 small satellites that measure reflected GPS signals off of the surface of the ocean. Although CYGNSS is not intended or optimized for altimetry, creation of a research product will allow assessment at the synoptic level over the life of the CYGNSS mission. The algorithms being developed here are for generation of large-scale production runs, unlike previous reports that have have examined limited sets of data. The CYGNSS Level 1 data is processed using a modified CYGNSS End-to-end Simulator (E2ES) which provides a full-forward-model waveform fit (along both delay and Doppler space) and precise delay retrieval. The delay is converted into estimated sea surface height (SSH) using auxiliary data sources for corrections, such as ionospheric, tropospheric, tidal, orbital and instrument. The resulting data product will enable thorough characterization of altimetry accuracy and remaining error sources.

Spaceborne GNSS Reflectometry (GNSS-R) measurements have shown strong coherent scattering over inland waters. It has been recognized that GNSS-R could be utilized for monitoring the global surface water distribution by making dynamic maps of wetlands as well as rapid response to flood events. Using the strength of the reflected signals, one can make maps that reveal the presence of water over land. In this paper, we used simulations to analyze the accuracy of these maps. The CYGNSS End-to-end Simulator (E2ES) was extended to include coherent scattering in the heterogeneous scenes where the region around the specular point is composed of both land and water in complex geometries. The simulation is then used to evaluate the accuracy of a simple fractional water in footprint approach to mapping wetland extent. We find that scattering from outside the first Fresnel zone and CYGNSS measurement processing effects significantly impact the accuracy of this approach. However, the accuracy can be improved by combining multiple measurements into a gridded map.

The objective of this paper is to model and study the sensitivity of bistatic microwave scattering versus changes in wetland characteristics as observed by a GNSS-R satellite system such as CYGNSS. We develop a simplified scattering model starting from the Water Cloud Model traditionally used in monostatic radar problems. Vegetation is idealized as a cloud of randomly oriented scattering elements over a rough surface representing either soil or water. The bistatic scattering coefficient is modeled as the incoherent sum of soil, water and vegetation scattering weighted by the fraction of each contribution within the CYGNSS footprint. The model is tested against CYGNSS observations across the Everglades National Park for which high-resolution land-cover and water depth maps are available. We show that our simplified model is able to capture to first order the variability of bistatic scattering versus changes in water depth and water fraction. This effort is a step forward towards the development of an effective algorithm to map the dynamic state of tropical wetlands and other regions subject to flooding using CYGNSS measurements.

The CYGNSS mission is a new GNSS-R ocean remote sensing constellation of eight small spacecraft. The measurements from CYGNSS allow for the first time a large, densely sampled, and calibrated GNSS-R dataset to be used for ocean surface altimetry. Re-tracking of the reflected signal correlation waveform is one of the most significant performance limiting components of ocean altimetry retrievals. This work demonstrates the use of empirical methods and physical scattering delay/delay-Doppler waveform models, fit to the measured correlation functions in a least-squares sense, to retrieve the specular reflection delay. We examine the performance of waveform modelling methods as compared to single-point re-tracking methods for the first time with a spaceborne dataset. Particular attention is given to the presence of narrow waveforms in the data set, indicative of coherent, rather than purely diffuse reflection. Re-tracking performance is quantified with statistics of the accuracy and precision and examples are presented for returns demonstrating both diffuse and coherent characteristics.

The Cyclone Global Navigation Satellite System (CYGNSS) mission was designed to study hurricane intensification by measuring wind speeds in tropical cyclones. However, the delay-Doppler maps (DDM) that are produced can be used to estimate the sea surface height (SSH) at the specular reflection point on the ocean surface. Proof-of-concept studies that DDMs are suitable to solve for SSH have been recently reported (Clarizia et al., 2016; Mashburn et al., 2018), based on data acquired by the demonstration satellite experiment Tech Demo Sat -1 (TDS-1) carrying a GNSS-R receiver similar to the ones onboard CYGNSS. Although the precision of each 1-sec averaged SSH is considerably lower than that of the existing satellite altimeters, by virtue of the dense coverage and frequent revisit time exhibited by the constellation of 8 microsats, the error may be smoothed down considerably by optimal interpolation (Li et al., 2016). Hence the CYGNSS dataset presents a potential opportunity to sample the tropical oceans, and investigate the sensitivity of the SSH measurements to mesoscale eddies.

NASA's Soil Moisture Active Passive (SMAP) mission has been tuned to perform a Global Navigation Satellite Systems Reflectometry (GNSS-R) experiment. The motivation of this study is to assess the capabilities of GNSS-R for soil moisture determination, biomass monitoring and cryosphere studies. The use for first time of the Polarimetric Ratio (PR) from a spaceborne platform shows significant sensitivity to soil moisture. Additionally, the investigation for first time of the leading and trailing edges width sensitivity to Above Ground Biomass (AGB) and rough topography shows promising results. Better understanding of these effects in the reflected waveforms will improve the development of retrieval algorithms.

Global Navigation Satellite Systems Reflectometry (GNSS-R) ocean applications includes scatterometry and altimetry. In this work, an investigation is performed on polarimetric scatterometry over ocean surface using data from a GNSS-R experiment on-board the Soil Moisture Active Passive (SMAP) mission, and on ocean surface topography from the Cyclone Global Navigation Satellite System (CyGNSS) mission using new retrieval algorithms. The former one provides global coverage because of the Sun Synchronous Orbit (SSO), while the latter one focuses on tropical latitudes providing a spatial sampling of 32 swaths. First results from SMAP over the Artic Sea show clearly sea ice effects on the reflected waveforms.

Spatial and temporal variations in near-surface soil moisture are important to measure for climate studies, numerical weather forecasts, and drought monitoring. Several previous studies have shown success in using ground-reflected Global Navigation Satellite System (GNSS) signals as a form of bistatic radar to sense soil moisture. However, the ability of this type of data to sense soil moisture variations from space is still a nascent field of study. In the past two years, three satellites have been launched that were either designed to capture ground-reflected GNSS signals or have been modified to record these signals. The data provided by these satellites are giving scientists an unprecedented opportunity to investigate their ability to detect changes in Earth's land surface, including but certainly not limited to near-surface soil moisture. This paper will present spaceborne observations of ground-reflected GNSS signals and evaluate their sensitivity to near-surface soil moisture. This sensitivity will be compared to empirical and theoretical sensitivities of monostatic L-band radar measurements to soil moisture. We will also comment on possibilities for retrieval algorithm development, using techniques employed for monostatic radar as a guide.

It is now recognized that the enormous challenge of scientifically understanding the Earth system requires careful strategic decisions on what missions are deployed. In a recent report, the National Research Council developed a “value framework” for Earth observing systems with a focus on prioritizing NASA observations that merit long-term continuity. In this paper, we refer to this framework to discuss how high value observations arise from opportunistic use of signals generated by Global Navigation Satellite Systems (GNSS) such as GPS. The increasing number of GNSS constellations internationally, likely to be permanently deployed, suggests that the geosciences community will benefit by adopting these signals for a variety of remote sensing needs. We describe recent progress in using these observations scientifically and developing technology to exploit them. We conclude that dedicated constellations of GNSS science instruments in low Earth orbit capable of receiving both direct and reflected GNSS signals will provide excellent science return in a broad range of areas, and constitute a high value Earth observing system.

It is important to closely monitor the state of the world's wetlands, as climate change and human encroachment in a rapid global urbanization trend threaten to cause large-scale wetland collapse. Because wetlands are often difficult to observe in situ, remote sensing is the only viable way to map wetland extent globally. However, current remote sensing methods suffer limitations in capturing wetland extent, and more importantly, wetland dynamics at appropriate spatial and temporal scales. GNSS-Reflectometry could help fill the current observation gap, as experimental data show that ground-reflected GNSS signals are very sensitive to changes in inundated areas. Furthermore, because this technique only requires a custom developed receiver and antenna system, a constellation of such instruments can potentially be launched at relatively low cost, providing global observations at sub-daily intervals. One challenge remains, however, which is quantitatively formulating the geophysical product of reflections over the land surface in various states of inundation. Here, we use a novel reflection dataset, derived from the SMAP radar receiver, to elucidate the sensitivity of reflections to small land surface features and their seasonal variations. Additionally, we quantify the dynamic range of reflections over both open and closed wetlands, and suggest an algorithm for wetland type classification.

Global Navigation Satellite System (GNSS) signals can be used as a kind of bistatic radar, with receivers opportunistically recording ground-reflected signals transmitted by the GNSS satellites themselves. The ground-reflected signals are sensitive to changes in surface permittivity, which for L-band is primarily a function of the moisture content of the surface. Here, we investigate the ability of GNSS signals, as recorded by a GPS receiver flown on a satellite, to measure changes in wetland extent and flood inundated area. We find that the ground-reflected signals give similar results as flood-inundation maps derived from other sources. Reflected power increases of over 10 dB in the vicinity of wetlands indicates that these signals could successfully map changes in wetlands around the globe.

GEROS-ISS (GEROS hereafter) stands for GNSS REflectometry, Radio Occultation and Scatterometry onboard the International Space Station. It is a scientific experiment, proposed to the European Space Agency (ESA) in 2011 for installation aboard the ISS. The main focus of GEROS is the dedicated use of signals from the currently available Global Navigation Satellite Systems (GNSS) for remote sensing of the System Earth with focus to Climate Change characterisation. The GEROS mission idea and the current status are briefly reviewed.

The Global Navigation Satellite System (GNSS) reflectometry, i.e. GNSS-R, is a novel remote-sensing technique first published in [1] that uses GNSS signals reflected from the Earth's surface to infer its surface properties such as sea surface height (SSH), ocean winds, sea-ice coverage, vegetation, wetlands and soil moisture, to name a few. This communication discusses the scientific value of GNSS-R to (a) furthering our understanding of ocean mesoscale circulation toward scales finer than those that existing nadir altimeters can resolve, and (b) mapping vegetated wetlands, an emerging application that might open up new avenues to map and monitor the planet's wetlands for methane emission assessments. Such applications are expected to be demonstrated by the availability of data from GEROS-ISS, an ESA experiment currently in phase A [2], and CyGNSS [3], a NASA mission currently in development. In particular, the paper details the expected error characteristics and the role of filtering played in the assimilation of these data to reduce the altimetric error (when averaging many measurements).

The value of Global Positioning Satellites (GPS) measurements to atmospheric science, space physics, and ocean science, is now emerging or showing a potential to play a major role in the evolving programs of NASA, NSF and NOAA. The objective of this communication is to identify and articulate the key scientific questions that are optimally, or perhaps uniquely, addressed by GPS or GPS-like observations, and discuss their relevance to existing or planned national Earth-science research programs. The GPS-based ocean reflection experiments performed to date have demonstrated the precision and spatial resolution suitable to altimetric applications that require higher spatial resolution and more frequent repeat than the current radar altimeter satellites. GPS radio occultation is promising as a climate monitoring tool because of its benchmark properties: its raw observable is based on extremely accurate timing measurements. GPS-derived temperature profiles can provide meaningful climate trend information over decadal time scales without the need for overlapping missions or mission-to-mission calibrations. By acquiring data as GPS satellites occult behind the Earth's limb, GPS also provides high vertical resolution information on the vertical structure of electron density with global coverage. New experimental techniques will create more comprehensive TEC maps by using signals reflected from the oceans and received in orbit. This communication will discuss a potential future GNSS Earth Observing System project which would deploy a constellation of satellites using GPS and GPS-like measurements, to obtain a) topography measurements based on GPS reflections with an accuracy and horizontal resolution suitable for eddy monitoring, and h) climate-records quality atmospheric temperature profiles. The constellation would also provide for measurements of ionospheric elec tron density. This is a good example of an interdisciplinary mission concept, with broad science objectives of high societal relevance, al l resting on common cost-effective technology.

A bistatic electromagnetic wave scattering model for the sea surface is developed to examine its wind dependence property over a wide range of incident angles along the specular direction. This is done by combining an existing scattering model with a sea spectrum recently reported in the literature. In general, electromagnetic wave scattering from a rough surface is dependent on the Fourier transform of the nth power of its height correlation function which can be computed numerically from the surface spectrum. This transform relation indicates that scattering is sensitive not only to the surface spectrum but also to its convoluted properties. Generally, surface scattering is sensitive only to a portion of the surface correlation measured from the origin. The size of this portion is a function of three variables (the incident angle, the surface height standard deviation, and the exploring wavelength) and the rate of decay of the correlation function. The decay rate near the origin of the sea surface correlation is very small, so much so that at L-band this portion is too wide for a two-term approximation of the correlation function. This is true in spite of the fact that the sea surface has a very large rms height. Thus, a scattering model based on geometric optics is generally not applicable at L-band especially at large angles of incidence. An additional finding is that in specular scattering wind dependence is stronger at larger angles of incidence for incident angles between 0 and 70/spl deg/ over the wind speed range of 4 m/s-20 m/s.