On satellite scatterometer model functions

Abstract

Abstract. Satellite scatterometer model functions are the empirical correlations between observed backscatter and geophysical parameters. One problem is the choice of "surface truth" for the latter. Currently, buoy data and general circulation model surface analyses have been used to establish 10 rn neutrally stratified wind model functions. The correlation between backscatter and surface winds has been the foundation of the scatterometers. The backscatter is anisotropic with respect to wind direction, allowing wind direction information to be derived from multiple looks at the same patch of ocean. Here the evolution of a model function for the pressure gradient (hence surface geostrophic wind) is given. The European Center for Medium-Range Weather Forecasts surface pressure analyses are used to establish the pressure-based model function. There are limitations and advantages to both geophysical data sets. IntroductionThe satellite scatterometers have experienced enormous success in producing dense high quality data for marine surface wind, stress, and pressure analyses. This is due to the remarkable correlation between surface roughness in the 1-10 cm capillary/short-gravity-wave regime and the wind vector. The latter are obtained from measurements at near-surface heights by buoys or from the general circulation model (GCM) surface wind products. While this agreement between the radar backscatter cross-section o-0, averaged over a 25 km or 50 km footprint, and the point measurement of the wind by a buoy was perhaps unexpected, it is now well documented. So much so, that it is often necessary to remind ourselves that the radar signal is from surface roughness and is subject to variation independent of surface stress or wind. These possible effects include, for instance, wave-wave interaction, nonequilibrium, attenuation of the radar signal, parasitic capillaries, rain, slicks, white water, and other factors [e.g., Brown, 1983].It appears that the process of averaging •r 0 over 25-50 km diameter footprints yields a stable signal corresponding to some steady state, uniform surface roughness characteristic. This roughness evidently responds to the wind magnitude with a very short lapse time (less than a few minutes maximum). While the wind generation of water waves theory is incomplete, we can assume that the waves are generated by an instability in the velocity/density profile at the water-air interface. Energy is transferred from the wind as the momentum flux to generate the waves [e.g., Kraus and Businger, 1999, pp. 124-136]. This momentum flux is by definition the stress at the surface. Hence the capillary and short gravity waves are related to the drag of the wind on the water. Energy is transferred from the wind to the sea surface in the form of the momentum flux. We assume that a fixed part of the energy goes into generating the short waves with some proportion going into dissipation, long waves, and other effects. The wavelength of the short waves evidently produces a strong response to an active microwave ...

Validation of ERS-2 SAR offshore wind-speed maps in the North Sea

“…The correlation analysis between the coarse grid cells of the scatterometer and, for example, buoy data can suffer from smaller scale gradients within the cells that are averaged out. Furthermore calibration data from buoys and ships may suffer errors due to flow distortion, tilting and displacement (Brown 2000a,b, Atlas et al 2001. Improvements to the current algorithms are ongoing as the global dataset collocated with buoys and model data becomes larger.…”

An assessment of SeaWinds on QuikSCAT wind retrieval

“…These factors generally cause the selected wind flow to appear ''noisy''. For a discussion on GMF inaccuracies and the limitations of the comparison data, see Brown [2000].…”

An SST‐dependent Ku‐band geophysical model function for RapidScat

“…A GMF can been written as: where is the stress‐equivalent wind speed at a height of 10 m [ Edson , ; Liu and Tang , ; De Kloe et al ., ], is the wind direction with respect to the radar range direction, is the radar emitted and received polarization, is the incidence angle, and is the radar wavelength [ Wentz et al ., ; Wentz and Smith , ; Brown , ; Gohil et al ., ; Hersbach et al ., ; Ricciardulli and Wentz , ]. Several studies have shown that it is necessary to take SST into account for a Ku‐band GMF, while for C‐band SST effects are much smaller [ Grodsky et al ., ; Bentamy et al ., ; Bentamy et al ., ; Wang et al ., ].…”