Enhanced Ocean Scatterometry

Abstract

An ocean scatterometer is an active microwave instrument which is designed to determine the normalized radar cross section (NRCS) of the sea surface. Scatterometers transmit pulses towards the sea surface and measure the reflected energy. The primary objective of spaceborne scatterometers is to measure near-surface winds over the ocean. This is made possible by observing the ocean with different azimuth views and by using a geophysical model function (GMF) which relates wind and backscatter. Nowadays satellites measure wind fields over the oceans worldwide on a daily basis to improve weather forecasts. Current Geophysical Model Functions are derived either from collocated backscatter and buoy-based wind measurements or by empirically fitting satellite data and NWP (Numerical Weather Prediction) model winds. These functions have been shown to be accurate to approximately ± 1.5 m/s for winds within the range 5÷15 m/s. The characterization of an empirically derived relationship between normalized radar cross-section and wind vector strongly depends on the data set from which such a relationship is derived. This limits the domain of validity of the model to a specific frequency, a specific polarization of the electromagnetic wave, and a confined range of incidence angles and wind speeds. In particular, given the low probability of having scatterometer observations and collocated independent wind vector measurements during high-wind events (such as hurricanes and typhoons), current empirical models are not well defined for high winds. Furthermore, the relatively coarse spatial resolution of existing spaceborne scatterometers (50 km x 50 km) results in the averaging of high and moderate surface winds within the cell. Nowadays, wind dependent correction terms are used, operationally, by the European Centre for Medium-Range Weather Forecasts (ECMWF) to overcome the underestimation of high wind speeds resulting from the CMOD geophysical model function. In addition, being measurement-dependent, the use of empirical functions makes it very difficult to distinguish errors associated with uncertainties of the observing system from errors associated with uncertainties of the models. In some cases, receive chain saturations of the observing systems are confused with limited sensitivity of normalized radar cross section to wind speed. In other cases, poor cross-talk performance of the instrument generates wrong polarization relationships between VV, HH and VH scattering products. Today, there is no integrated approach valid across different frequencies, polarizations, incidence angles and wind speeds that is used operationally to model the relationship between ocean scattering and wind vector. This represents the key objective of the present thesis. Radar backscatter models, based on a description of the underlying physical phenomenon, have the big potential of providing a more general and understandable relation between measured microwave backscatter and surface wind field than the empirical models. The growing interest in achieving a better understanding of the physics that governs the scattering of microwave radiation from sea is triggered by extremely rich and varied data sets collected by microwave sensors on numerous space-borne satellite missions. Additional missions with advanced microwave sensors are planned for launch in the near future; they ask not only for better spatial resolution, radiometric resolution and stability but also for wider swaths and multi-polarisation observation capabilities. Higher spatial resolutions, can better describe the spatial variations in hurricanes and coastal wind fields, whereas wider swaths can sensibly reduce the time between consecutive observations of the same area on the Earth. Wider swaths imply wider ranges of incidence angles to be explored, and then new challenges arise in the modelling of the interaction of the electromagnetic and oceanic waves. Current operating scatterometers use only co-polar scattering (VV or HH) to retrieve wind speeds and directions. The main reason behind this design choice is associated with the fact that the Signal-to-Noise Ratio (SNR) in co-polarization is expected to be higher than in cross-polarization for most winds. However, airborne measurements over hurricanes, performed at C-band and Ku-band, have confirmed that co-polar scattering suffers from problems of incidence- and azimuth angle-dependent signal saturation and dampening, which make it only weakly sensitive to wind speed variations above 25 m/s. This shortcoming impairs the ability to provide accurate hurricane warnings. Errors in wind sometimes prevent communities from correctly identifying the most vulnerable regions where emergency preparations are needed. The addition of VH polarisation to the standard VV and HH polarisations, can significantly improve the retrieval of the wind speed in case of extreme weather events, such as hurricanes. Analysis of RADARSAT-2 quad-polarisation data with collocated in situ ocean wind measurements has recently revealed that the cross-polarised radar backscatter does not saturate at high wind speeds. As a result, the wind speed retrieved with cross-polarised backscattering is much more accurate than that retrieved with co-polarised data in hazardous storm conditions. In this thesis, we present an analytical physical model for accurate simulation of full-polarimetric microwave sea-surface scattering and Doppler signatures. This model combines an adequate sea surface description with advanced electromagnetic theories to simulate both monostatic and bistatic scattering over a wide range of wind speeds, radar frequencies, incidence angles, different polarisations and arbitrary radar look direction with respect to the wind direction. Results will be compared with real measurements from ASAR, Sentinel-1, Radarsat-2, ASCAT and well established empirical Geophysical Model Functions showing good agreement. Being capable of simulating full-polarimetric Doppler spectra of microwave backscatter from ocean surface, this model will be used to explore ocean surface motion retrievals, thus supporting the definition of future ocean Doppler scatterometers, capable of simultaneous measurement of Ocean Vector Wind (OVW) and Ocean Vector Motion (OVM) on a global scale. Measurements of powerful, complex and highly variable ocean surface currents are fundamental for a variety of applications, such as the monitoring of changes in coastal regions, risk management for coastal and off-shore structures, ship routing, anthropogenic and natural pollution and offshore renewable energy monitoring. Ocean surface currents are complex and, in coastal areas, highly dynamic, and therefore need to be monitored with short time sampling (possible on a daily basis) on a global scale. To this aim, Ocean Doppler scatterometry provides simultaneous and accurate measurements of wind fields and ocean motion vectors that can be used to generate global surface ocean current maps at a spatial resolution of 25 km (i.e. 12.5 km spatial sampling) on a daily basis (thanks to the very large swath illuminated). These maps will allow gaining some insights on the upper ocean dynamics at mesoscale.