RapidScat is a Ku-band scatterometer that was launched September 2014 and is currently operating on the International Space Station. It estimates ocean vector winds through accurate measurement of the normalized radar coefficient (σ0) of the ocean surface. In order to ensure the accuracy of σ0 measurements and consistency with previous Ku-band scatterometers, post-launch calibration and validation is necessary. Calibration and validation is performed using natural land targets, namely the Amazon and Congo rainforests, to complement calibration efforts over the ocean. The σ0 response of the targets is estimated with respect to viewing angle and time of year using previous Ku-band scatterometers. Taking advantage of the ISS orbit, the diurnal response of each target is estimated using RapidScat. Normalizing factors for incidence angle, azimuth angle, local time of day, and time of year are derived from these measured responses. RapidScat σ0 measurements are found to be consistent throughout its mission life with instrumental drift less than 0.3 dB. The effectiveness of slice balancing is evaluated and found to be highly dependent on the pitch of the ISS. Understanding of the diurnal backscatter response and incidence response allow comparison of RapidScat measurements with measurements from the QuikSCAT, NSCAT, and Oceansat-II scatterometers. RapidScat σ0 is found to be biased low compared to QuikSCAT by 0.1--0.3 dB.

Satellite scatterometers have recently gained popularity due to their unique ability to measure global geophysical data on a daily basis. Increased interest in scatterometry mandates improved design and calibration of these instruments. This dissertation presents new techniques for scatterometer calibration and addresses issues related to the design of future instruments and applications. First, the use of a calibration ground station is considered. A new methodology is established for calibration of SeaWinds, NASA's current scatterometer, using a receive-only ground station. Principles of the methodology are implemented, new analysis techniques developed, and important results obtained for instrument timing, frequency, power, position, and pointing. Second, an investigation into methods for calibration of measurement surface location is conducted. Two new approaches are proposed and results of both approaches using SeaWinds data are provided. Third, measurement correlation, a critical issue related to new scatterometer designs, particularly those which significantly oversample the surface is considered. General statistical expressions for measurement correlation are derived and analysis of the effects on data variance is presented. Finally, a new data simulation model is developed to support instrument and application development. New applications require sophisticated models which are general, yet accurate, enabling them to rapidly and easily simulate data from multiple instruments. The model generates data which is statistically equivalent (in a mean and variance sense) to actual scatterometer measurements by separately accounting for the two main forms of variation present in scatterometer data, multiplicative fading and additive noise, and also accounting for correlation between measurements. The model is valuable for a variety of data applications including image generation and high resolution wind retrieval.

The Ku-band SeaWinds-on-QuikSCAT scatterometer (QuikSCAT) operated continuously from 1999 to 2009. Though its primary mission was to estimate global ocean winds, QuikSCAT has proven useful in a variety of geophysical studies using land backscatter measurements. The end of the primary QuikSCAT mission in 2009 has prompted interest for continuing the QuikSCAT land dataset with other scatterometers. The Oceansat-2 scatterometer (OSCAT), launched in 2009, is a viable candidate for continuing the QuikSCAT time series because of the similarities of both sensors in function and design. An important difference in the sensors is that they operate at slightly different incidence angles. Continuing the time series requires careful cross-calibration of the two sensors. Because the sensor datasets overlapped by only a few weeks in late 2009, the amount of simultaneous data is insufficient to describe temporal and locational variations in the relative calibration, or difference between QuikSCAT and OSCAT measurements. To overcome this limitation, we perform direct and model-based comparisons of temporally-disjoint QuikSCAT and OSCAT global land measurements to describe the relative calibration. Using homogeneous rainforest targets, we also identify drift and azimuthal biases in the OSCAT dataset and present suggestions for removing them. The relative calibration is found to vary locationally by several tenths of a decibel over certain regions. Evidence is presented that suggests the relative calibration is dependent on environmental factors such as vegetation density and freeze-thaw status and results from the different incidence angles of the measurements.

The Antarctic ice sheet is a geophysically - and in an age of growing concern about global warming, geopolitically - important portion of Earth. The composition and dynamics of the Antarctic ice sheet influence global climate patterns, global sea level and the planet's radiation budget. Recent evidence also suggests that the long term stability of portions of the ice sheet may be in jeopardy. In this thesis I use data from three Ku-band space-borne scatterometers to monitor changes in the backscatter signature of the Antarctic ice sheet from 1978 through 2007. Significant changes in backscatter, which result from geophysical changes in the ice sheet itself, are found over much of the Antarctic continent, especially in West Antarctica and along much of the coasts. Less drastic changes, including regular seasonal variations, are observed over much of the ice shelf. Possible scattering mechanisms are proposed and discussed. A secondary result is the demonstration of the stability of NASA's QuikSCAT scatterometer, data from which is used extensively in this thesis and in many other publications. It is shown that QuikSCAT's observation geometry and backscatter instrumentation have remained consistent to great precision throughout its nearly nine-year long mission.

A resposta oceânica a perturbações com períodos e comprimentos significativamente maiores que o período inercial e que o raio de deformação de Rossby se dá na forma de ondas de Rossby planetárias. Geralmente, as perturbações são atribuídas a variações no rotacional do vento via bombeamento de Ekman. A passagem dessas ondas causa deformação das isopicnais, podendo resultar em anomalias da temperatura da superfície do mar (TSM) por advecção vertical. Dependendo de como ocorre a interação ar-mar, anomalias de TSM podem alterar o campo de ventos ou serem alteradas por ele através de fluxo de calor. Este trabalho utiliza dez anos de dados de temperatura da superfície do mar, velocidade e direção dos ventos e anomalia da altura do mar obtidos por satélites para identificar regiões do oceano onde há forçamento direto do vento na geração de ondas planetárias que se propagam linearmente. Mapas de correlação cruzada entre essas variáveis permitiram identificar onde a interação entre o oceano e a atmosfera é linear. Um modelo simples de uma camada e meia forçado apenas pelo bombeamento de Ekman foi utilizado para testar se, nestas regiões, a variabilidade atmosférica seria suficiente para explicar a variabilidade das ondas de Rossby estimadas pelos dados altimétricos. A interação entre a TSM e a intensidade do vento no Atlântico sul tropical é distinta das demais bacias oceânicas. Das correlações entre a TSM e o rotacional da tensão de cisalhamento do vento, observou-se que a dinâmica de Ekman não é marcante no Índico. Nas regiões tropicais do Atlântico e do Pacífico, as previsões do modelo foram similares às observações. Por fim, foram obtidas evidências de geração e retroalimentação de ondas planetárias nas bordas leste do Atlântico e do Pacífico. / Rossby waves are the ocean response to perturbations whose temporal and spatial scales are significantly longer than both the inertial period and the Rossby radius of deformation. These perturbations are, more often than not, attributed to variations in the wind stress curl {\\em via} Ekman pumping. The waves cause isopycnal displacement which due to vertical advection may result in sea surface temperature (SST) anomalies. Depending on the ocean--atmosphere interaction, SST anomalies can either change the wind field or be changed by it due to the heat flux. This study makes use of ten years of satellite derived SST, wind vector, and sea surface height anomaly data to identify regions where there is direct wind forcing of linear Rossby waves. Cross-correlation maps between these variables show where linear interactions occur. A simple 1½ layer model forced by Ekman pumping was used to check if, in those regions, atmospheric variability alone can explain the observed Rossby wave variability as estimated from radar altimeter data. The interaction between SST and wind magnitude in the South Atlantic is distinct from all other ocean basins. SST and wind stress curl correlations show that the Ekman dynamics is not dominant in the Indian Ocean. In the tropical Atlantic and Pacific the model predictions are similar to the observations. Finally, evidence of genesis and feedback of planetary waves is presented for the eastern boundaries of the Atlantic and Pacific oceans.

A resposta oceânica a perturbações com períodos e comprimentos significativamente maiores que o período inercial e que o raio de deformação de Rossby se dá na forma de ondas de Rossby planetárias. Geralmente, as perturbações são atribuídas a variações no rotacional do vento via bombeamento de Ekman. A passagem dessas ondas causa deformação das isopicnais, podendo resultar em anomalias da temperatura da superfície do mar (TSM) por advecção vertical. Dependendo de como ocorre a interação ar-mar, anomalias de TSM podem alterar o campo de ventos ou serem alteradas por ele através de fluxo de calor. Este trabalho utiliza dez anos de dados de temperatura da superfície do mar, velocidade e direção dos ventos e anomalia da altura do mar obtidos por satélites para identificar regiões do oceano onde há forçamento direto do vento na geração de ondas planetárias que se propagam linearmente. Mapas de correlação cruzada entre essas variáveis permitiram identificar onde a interação entre o oceano e a atmosfera é linear. Um modelo simples de uma camada e meia forçado apenas pelo bombeamento de Ekman foi utilizado para testar se, nestas regiões, a variabilidade atmosférica seria suficiente para explicar a variabilidade das ondas de Rossby estimadas pelos dados altimétricos. A interação entre a TSM e a intensidade do vento no Atlântico sul tropical é distinta das demais bacias oceânicas. Das correlações entre a TSM e o rotacional da tensão de cisalhamento do vento, observou-se que a dinâmica de Ekman não é marcante no Índico. Nas regiões tropicais do Atlântico e do Pacífico, as previsões do modelo foram similares às observações. Por fim, foram obtidas evidências de geração e retroalimentação de ondas planetárias nas bordas leste do Atlântico e do Pacífico. / Rossby waves are the ocean response to perturbations whose temporal and spatial scales are significantly longer than both the inertial period and the Rossby radius of deformation. These perturbations are, more often than not, attributed to variations in the wind stress curl {\\em via} Ekman pumping. The waves cause isopycnal displacement which due to vertical advection may result in sea surface temperature (SST) anomalies. Depending on the ocean--atmosphere interaction, SST anomalies can either change the wind field or be changed by it due to the heat flux. This study makes use of ten years of satellite derived SST, wind vector, and sea surface height anomaly data to identify regions where there is direct wind forcing of linear Rossby waves. Cross-correlation maps between these variables show where linear interactions occur. A simple 1½ layer model forced by Ekman pumping was used to check if, in those regions, atmospheric variability alone can explain the observed Rossby wave variability as estimated from radar altimeter data. The interaction between SST and wind magnitude in the South Atlantic is distinct from all other ocean basins. SST and wind stress curl correlations show that the Ekman dynamics is not dominant in the Indian Ocean. In the tropical Atlantic and Pacific the model predictions are similar to the observations. Finally, evidence of genesis and feedback of planetary waves is presented for the eastern boundaries of the Atlantic and Pacific oceans.

Spaceborne microwave sensors are capable of estimating various properties of many geophysical phenomena, including the age and extent of Arctic sea ice and the relative soil moisture over land. The measurement and classification of such geophysical phenomena are used to refine climate models, localize and predict drought, and better understand the water cycle. Data from the active Ku-band scatterometers, the Quick Scatterometer (QuikSCAT), and the Oceansat-2 Scatterometer (OSCAT), are here used to classify areas of first-year and multiyear Arctic sea ice using a temporally adaptive threshold on reported radar backscatter values. The result is a 15-year data record of daily ice classification images. An additional ice age data record is produced using the C-band Advanced Scatterometer (ASCAT) and the Special Sensor Microwave Imager Sounder (SSMIS) with an alternate classification methodology based on Bayesian decision theory. The ASCAT/SSMIS classification methodology results in a record which is generally consistent with the QuikSCAT and OSCAT classifications, which conclude in 2014. With multiple ASCAT and SSMIS sensors still operational, the ASCAT/SSMIS ice classifications can continue to be produced into the future. In addition to ice classification, ASCAT is used to estimate the relative surface soil moisture at high-resolution (4.45 — 4.45 km per pixel). The soil moisture estimates are obtained using enhanced resolution image reconstruction techniques and an altered version of the Water Retrieval Package (WARP) algorithm. The high-resolution soil moisture estimates are shown to agree well with the existing lower resolution WARP products while also revealing finer details.

Originally designed for wind velocity estimation over the ocean, scatterometers have since been applied to climate studies of the Earth's cryosphere and bioshere. As an integral part of climatological studies of the planet, the NASA Scatterometer Climate Record Pathfinder (SCP) supplies scatterometer-based products designed to aid researchers in climatological studies of the planet. In this thesis, necessary steps are taken to facilitate data from the Oceansat-2 Ku-band scatterometer (OSCAT) to be used in extending the Ku-band SCP dataset of conically scanning pencil-beam scatterometers begun by the Seawinds scatterometer flown on the QuikSCAT mission 1999-2009. As a standard SCP product, a temporal resolution enhancement technique for the scatterometer image reconstruction (SIR) algorithm is applied to OSCAT data. A relative cross-calibration method is developed to ensure consistency amongst datasets of conically scanning pencil-beam scatterometers in the SCP data time series. By application of the method, both raw data and SIR image data of OSCAT is cross-calibrated with QuikSCAT. To enable creation of SCP products requiring knowledge of the spatial response function (SRF) with OSCAT data, a method of estimating the SRF of pencil-beam scatterometers is developed. The estimation method employs rank-reduced least-squares to invert the radar equation using measurements over islands. A simulation is performed to validate the efficacy of the method and provide optimum choice of island size and number of singular values used in rank-reduced least-squares. The utility of the SRF estimates is demonstrated by applying an estimate of the OSCAT SRF to SIR image construction with OSCAT data.

RapidScat 2.5 km ultrahigh resolution (UHR) wind estimation is introduced and validated it in near-coastal regions. In addition, this thesis applies direction interval retrieval techniques and develops a new wind processing method to enhance the performance of RapidScat UHR wind estimation in the nadir region. The new algorithm is validated with L2B wind estimates, Numerical Weather Prediction (NWP) wind products, and buoy measurements. The wind processing improvements produce more spatially consistent UHR winds that compare well with the wind products mentioned above. Hawaii regional climate model (HRCM), QuikSCAT, and ASCAT wind estimates are compared in the lee of the Big Island with the goal of understanding UHR scatterometer wind retrieval capabilities in this area. UHR wind vectors better resolve fine resolution wind speed features compared to L2B, but still do not resolve the expected wind direction features. A comparison of scatterometer measured σ 0 and HRCM and NWP predicted σ 0 suggests that scatterometers can detect a reverse flow in the lee of the island. Differences between scatterometer measured σ 0 and HRCM predicted σ 0 indicate error in the placement of key reverse flow features by the model. Coarse initialization fields and a large fixed size median filter window are also shown to impede UHR wind retrieval in this area.

Field-wise wind estimation (also known as model-based wind estimation) is a sophisticated technique to derive wind estimates from radar backscatter measurements. In contrast to the more traditional method known as point-wise wind retrieval, field-wise techniques estimate wind field model parameters. In this way, neighboring wind vectors are jointly estimated, ensuring consistency. This work presents and implementation for field-wise wind retrieval for the SeaWinds scatterometer on the QuikSCAT satellite. Due to its sophistication, field-wise wind retrieval adds computational complexity and intensity. The tradeoffs necessary for practical implementations are examined and quantified. The Levenberg-Marquardt algorithm for minimizing the field-wise objective function is presented. As the objective function has several near-global local minima, several wind fields represent ambiguous wind field estimates. A deterministic method is proposed to ensure sufficient ambiguities are obtained. An improved method for selecting between ambiguous wind field estimates is also proposed. With a large set of Sea-Winds measurements and estimates available, the σ° measurement statistics are examined. The traditional noise model is evaluated for accuracy. A data-driven parameterization is proposed and shown to effectively estimate measurement bias and variance. The parameterized measurement model is used to generate Cramer-Rao bounds on estimator performance. Using the Cramer-Rao bound, field-wise and point-wise performances are compared.