Abstract's details
Inclusion of the ocean's vertical velocity variance into the sea-state-bias correction
CoAuthors
Event: 2023 Ocean Surface Topography Science Team Meeting
Session: Instrument Processing: Propagation, Wind Speed and Sea State Bias
Presentation type: Type Oral
Contribution: PDF file
Abstract:
The sea-state-bias (SSB) correction is one of the largest sources of error in sea-surface-height estimation from nadir altimetry. The SSB is often considered to consist of three contributions: 1) a difference in reflectivity of wave crests and troughs (the electromagnetic bias), 2) non-linear wave shapes (skewness bias) and 3) instrumental and processing effects. Empirical models to correct for the SSB now depend on the significant wave height (SWH) and the normalised radar cross section (NRCS), or its derived level-2 parameter, stress-equivalent wind speed. Despite attempts to improve the model using external data or gradients of retracked parameters, the uncertainty of the SSB correction still lingers around 2 cm. With requirements for upcoming altimeter missions (for example Sentinel-3 Next Generation Topography) being pushed towards this level, an improved correction is required.
Following other works, we agree that the NRCS and SWH are insufficient to describe the complex nature of the ocean surface. The NRCS is closely connected to the mean-square slope, which is primarily driven by short wind waves (<5 m wavelength). The SWH depends primarily on long wind waves and swell, but does not discriminate between both, while they both contribute differently to the electromagnetic and skewness biases. The absence or presence of swell would therefore already introduce problems for the two-parameter model. Additional complexity arises from the presence of surface currents, or near-shore ocean-bottom topography (wave breaking, refraction) and wind-speed variations (wave age). These issues cannot be overcome by traditional processing approaches.
Buchhaupt et al. (2023) already hinted on the connection between the SSB and the ocean's vertical velocity variance. The velocity variance is dominated by medium to long wind waves, while the contribution of swell (if present) is typically in the order of 10%. The addition of the velocity variance parameter is therefore expected to improve the SSB model. Focused SAR processing for nadir altimeters allows to estimate the along-track autocorrelation. The along-track auto-correlation depends on the sea state and is under moderate conditions dominated by the line-of-sight-projected velocity variance of the ocean surface.
In this study, we apply focused SAR processing to Sentinel-6 data at crossovers. We introduce the estimated autocorrelation as a new parameter in the SSB model for the low-resolution sea-surface-height product. The results are compared to the traditional two-parameter models. Some limitations to the suggested methodology are discussed. Further improvements are expected by separating wind-wave and swell parameters based on the analysis of waveform tails.
Following other works, we agree that the NRCS and SWH are insufficient to describe the complex nature of the ocean surface. The NRCS is closely connected to the mean-square slope, which is primarily driven by short wind waves (<5 m wavelength). The SWH depends primarily on long wind waves and swell, but does not discriminate between both, while they both contribute differently to the electromagnetic and skewness biases. The absence or presence of swell would therefore already introduce problems for the two-parameter model. Additional complexity arises from the presence of surface currents, or near-shore ocean-bottom topography (wave breaking, refraction) and wind-speed variations (wave age). These issues cannot be overcome by traditional processing approaches.
Buchhaupt et al. (2023) already hinted on the connection between the SSB and the ocean's vertical velocity variance. The velocity variance is dominated by medium to long wind waves, while the contribution of swell (if present) is typically in the order of 10%. The addition of the velocity variance parameter is therefore expected to improve the SSB model. Focused SAR processing for nadir altimeters allows to estimate the along-track autocorrelation. The along-track auto-correlation depends on the sea state and is under moderate conditions dominated by the line-of-sight-projected velocity variance of the ocean surface.
In this study, we apply focused SAR processing to Sentinel-6 data at crossovers. We introduce the estimated autocorrelation as a new parameter in the SSB model for the low-resolution sea-surface-height product. The results are compared to the traditional two-parameter models. Some limitations to the suggested methodology are discussed. Further improvements are expected by separating wind-wave and swell parameters based on the analysis of waveform tails.