Search

Abstract An assessment of variance in ocean current signal and noise shared by in situ observations (drifters) and a large gridded analysis (GlobCurrent) is sought as a function of day of the year for 1993–2015 and across a broad spectrum of current speed. Regardless of the division of collocations, it is difficult to claim that any synoptic assessment can be based on independent observations. Instead, a measurement model that departs from ordinary linear regression by accommodating error correlation is proposed. The interpretation of independence is explored by applying Fuller's (1987) concept of equation and measurement error to a division of error into shared (correlated) and unshared (uncorrelated) components, respectively. The resulting division of variance in the new model favours noise. Ocean current shared (equation) error is of comparable magnitude to unshared (measurement) error and the latter is, for GlobCurrent and drifters respectively, comparable to ordinary and reverse linear regression. Although signal variance appears to be small, its utility as a measure of agreement between two variates is highlighted. Sparse collocations that sample a dense (high resolution) grid permit a first order autoregressive form of measurement model to be considered, including parameterizations of analysis-in situ error cross-correlation and analysis temporal error autocorrelation. The former (cross-correlation) is an equation error term that accommodates error shared by both GlobCurrent and drifters. The latter (autocorrelation) facilitates an identification and retrieval of all model parameters. Solutions are sought using a prescribed calibration between GlobCurrent and drifters (by variance matching). Because the true current variance of GlobCurrent and drifters is small, signal to noise ratio is near zero at best. This is particularly evident for moderate current speed and for the meridional current component. Highlights • Signal/noise of two ocean current datasets are compared by day of the year and speed. • A first order autoregressive form of regression model and its solution are derived. • Fuller's (1987) equation/measurement errors are applied to shared/unshared errors. • Shared true variance is smaller, and equation error larger, than expected. • Measurement error is saturated at ordinary and reverse least squares bounds. [ABSTRACT FROM AUTHOR]

The complexity of real-world geophysical systems is often compounded by the fact that the observed measurements depend on hidden variables. These latent variables include unresolved small scales and/or rapidly evolving processes, partially observed couplings, or forcings in coupled systems. This is the case in ocean–atmosphere dynamics, for which unknown interior dynamics can affect surface observations. The identification of computationally-relevant representations of such partially-observed and highly nonlinear systems is thus challenging and often limited to short-term forecast applications. Here, we investigate the physics-constrained learning of implicit dynamical embeddings, leveraging neural ordinary differential equation (NODE) representations. In particular, we restrict the NODE representation to linear–quadratic dynamics and enforce a global boundedness constraint, which promotes the generalization of the learned dynamics to arbitrary initial conditions. The proposed architecture is implemented within a deep learning framework, and its relevance is demonstrated with respect to state-of-the-art schemes for different case studies representative of geophysical dynamics. • Neural ODEs are studied for the modeling of partially observed dynamical systems. • The models are constrained to remain bounded using the Schlegel boundedness theorem. • This framework allows deriving dynamical models with good asymptotic behaviors. • These models also have good generalization properties for unprecedented initial conditions. [ABSTRACT FROM AUTHOR]

Measurements of wave heights in marginal ice zones are limited to very few in situ data. Here we revisit the linear and quasilinear theories of Synthetic Aperture Radar imaging of waves in the particular case of waves in sea ice. Instead of only working with spectra, we have developed an iterative nonlinear algorithm to estimate phase-resolved deterministic maps of wave-induced orbital velocities, from which elevation spectra can be derived. Application of this algorithm to Sentinel 1A wave mode images in the Southern Ocean shows that it produces reasonable results for swells in all directions except when they propagate at a few degrees off the range direction. The estimate of wave parameters is expected to work best when the shortest wave components, those which cause a pixel displacement of the order of the dominant wavelength in azimuth, can be neglected. Otherwise short waves produce a blurring of the image, increasing exponentially with the azimuthal wavenumber and reducing the estimated wave amplitude. Given the expected spatial attenuation of waves in ice-covered regions, our deterministic method should apply beyond a few tens of kilometers in the ice, without any correction for short wave effects. In situ data collected around the ice edge as part of the 2015 SeaState DRI cruise in the Beaufort confirm the progressive image blurring caused by such short waves, and the apparent reduction in the wave modulation. When short waves propagate from the open ocean towards the ice, this blurring can produce an unrealistic apparent increase of wave height, from the open ocean up to a few tens of kilometers inside the ice. [ABSTRACT FROM AUTHOR]

In this study we show that spectral and phase-resolving wave models, used for coastal engineering applications, can provide significant differences in the local wave heights and wave power magnitude in the presence of strong bottom-induced refraction. We present here methodologies to compare, in a consistent manner, model outputs to an alternative source of spatial data derived from Synthetic Aperture Radar (SAR) satellite images. Normalized amplitude-derived data are computed from SAR images, compared to both normalized amplitude-derived map from a phase-resolving model outputs and normalized significant wave height-derived map from a spectral model. Two methodologies of SAR images analysis are developed in this paper. While one appears unsuitable due to stochastic wave group modulations, the second methodology provides encouraging similarities between the SAR data and the phase-resolving model outputs. • Spectral and phase-resolving wave models behave differently in coastal areas with strong bottom-induced refraction. • SAR data may represent a new source of data to investigate the spatial variability of the wave field in coastal waters. • Methodologies are developed to derive comparable models and SAR data. • Encouraging similarities are observed between the phase-resolving model and SAR data for the site and situations considered. [ABSTRACT FROM AUTHOR]

Abstract: A method is proposed to retrieve and interpret fine spatial variations of the sea surface roughness in sun glitter imagery. Observed sun glitter brightness anomalies are converted using a transfer function determined from the smoothed shape of sun glitter brightness. The method is applied to MODIS and MERIS sun glitter imagery of natural oil seeps and the catastrophic Deepwater Horizon oil spill in the Gulf of Mexico. The short-scale roughness variations in the presence of mineral oils slicks are consistently extracted and compared to variations associated with the biogenic slicks. In doing so, the wind speed dependency on the roughness anomalies is also considered. A comparison to normalized radar cross section (NRCS) anomalies taken from the corresponding high resolution ASAR images is performed, and similarities as well as differences are investigated. The results document significant benefit from the synergetic use of sun glitter and radar imagery for detection and monitoring of surface slicks. [Copyright &y& Elsevier]

Abstract: The variability of small-size iceberg distributions is revealed from a novel analysis of satellite altimeter data. A strong annual cycle is modulated by pulse-like events confined to single ocean basins, with dense iceberg populations in the South Atlantic in 2004–2005, and in the South Pacific in 2008. Anomalies in sea surface temperatures of the order of 1°C may be related to the iceberg distribution. Icebergs also appear very strongly associated with anomalies in the heights of ocean waves. A preliminary parameterization of wave blocking by icebergs significantly reduces wave model errors in the region south of 45° South, and has a perceptible influence on all the west coasts of the Southern hemisphere. [Copyright &y& Elsevier]

The Sentinel-2 MultiSpectral Instrument (MSI) collects multiple spectral band images, corresponding to specific sensing wavelengths and spatial resolutions, i.e. 10 m , 20 m and 60 m , respectively. Images are collected one at the time with a given time-lag between observations. Under favorable conditions, spatio-temporal characteristics of propagating ocean surface waves can thus uniquely be retrieved. A method for surface current vector field reconstruction is then developed. Demonstrated over different deep ocean regions, the retrieved surface current fields well compare with medium-resolution ocean circulation model or derived-velocities from altimeter measurements. At finer scales, the surface wave-conservation law is recovered, with the associated relationship between current vorticity and wave-ray curvature. Over shallow water regions, the wave propagation properties well follow sea depth variations, consistent with ETOPO1 data. Finally, time-lag between detector bands can also be exploited to estimate speed and direction properties of detected surface wave breaking whitecaps. An analysis of velocity reconstruction errors further reveals that Sentinel-2 MSI inter-channel co-registration is realized with an accuracy better than 0.1 pixel. Overall, these results confirm very promising capabilities of optical imagery to provide direct surface current velocity measurements from space, over relatively large areas, O(100 km), with a spatial resolution down to O(1 km). • A method is developed for surface current reconstruction from Sentinel-2 images •It is based on cross-spectral phase estimation between images from different bands •Retrieved current fields well compare with model and altimeter measurements •Over shallow water regions, sea depth is reconstructed •Inter-channel time delay is used to estimate speed and direction of large whitecaps [ABSTRACT FROM AUTHOR]