Skin and bulk sea surface temperature estimates from passive microwave and thermal infrared satellite imagery and their relationships to atmospheric forcing

Abstract. In the context of stand alone ocean models, the atmospheric forcing is generally computed using atmospheric parameters that are derived from atmospheric reanalysis data and/or satellite products. With such a forcing, the sea surface temperature that is simulated by the ocean model is usually significantly less accurate than the synoptic maps that can be obtained from the satellite observations. This not only penalizes the realism of the ocean long-term simulations, but also the accuracy of the reanalyses or the usefulness of the short-term operational forecasts (which are key GODAE and MERSEA objectives). In order to improve the situation, partly resulting from inaccuracies in the atmospheric forcing parameters, the purpose of this paper is to investigate a way of further adjusting the state of the atmosphere (within appropriate error bars), so that an explicit ocean model can produce a sea surface temperature that better fits the available observations. This is done by performing idealized assimilation experiments in which Mercator-Ocean reanalysis data are considered as a reference simulation describing the true state of the ocean. Synthetic observation datasets for sea surface temperature and salinity are extracted from the reanalysis to be assimilated in a low resolution global ocean model. The results of these experiments show that it is possible to compute piecewise constant parameter corrections, with predefined amplitude limitations, so that long-term free model simulations become much closer to the reanalysis data, with misfit variance typically divided by a factor 3. These results are obtained by applying a Monte Carlo method to simulate the joint parameter/state prior probability distribution. A truncated Gaussian assumption is used to avoid the most extreme and non-physical parameter corrections. The general lesson of our experiments is indeed that a careful specification of the prior information on the parameters and on their associated uncertainties is a key element in the computation of realistic parameter estimates, especially if the system is affected by other potential sources of model errors.

Download Full-text

Deep Learning for Sea Temperature Eddy signature Classification

10.5194/egusphere-egu21-7635 ◽

2021 ◽

Author(s):

Evangelos Moschos ◽

Alexandre Stegner ◽

Olivier Schwander ◽

Patrick Gallinari

Keyword(s):

Remote Sensing ◽

Deep Learning ◽

Surface Temperature ◽

Sea Surface Temperature ◽

Mediterranean Sea ◽

Satellite Imagery ◽

Mesoscale Eddies ◽

Sea Surface ◽

Sea Temperature ◽

Cloud Coverage

Mesoscale eddies are oceanic vortices with radii of tens of kilometers, which live on for several months or even years. They carry large amounts of heat, salt, nutrients, and pollutants from their regions of formation to remote areas, making it important to detect and track them. Using satellite altimetric maps, mesoscale eddies have been detected via remote sensing with advancing performance over the last years [1]. However, the spatio-temporal interpolation between satellite track measurements, needed to produce these maps, induces a limit to the spatial resolution (1/12&#176; in the Med Sea) and large amounts of uncertainty in non-measured areas.Nevertheless, mesoscale oceanic eddies also have a visible signature on other satellite imagery such as Sea Surface Temperature (SST), portraying diverse patterns of coherent vortices, temperature gradients, and swirling filaments. Learning the regularities of such signatures defines a challenging pattern recognition task, due to their complex structure but also to the cloud coverage which can corrupt a large fraction of the image.We introduce a novel Deep Learning approach to classify sea temperature eddy signatures [2]. We create a large dataset of SST patches from satellite imagery in the Mediterranean Sea, containing Anticyclonic, Cyclonic, or No Eddy signatures, based on altimetric eddy detections of the DYNED-Atlas [3]. Our trained Convolutional Neural Network (CNN) can differentiate between these signatures with an accuracy of more than 90%, robust to a high level of cloud coverage.We furtherly evaluate the efficiency of our classifier on SST patches extracted from oceanographic numerical model outputs in the Mediterranean Sea. Our promising results suggest that the CNN could complement the detection, tracking, and prediction of the path of mesoscale oceanic eddies.[1] Chelton, D. B., Schlax, M. G. and Samelson, R. M. (2011). Global observations of nonlinear mesoscale eddies. Progress in oceanography, 91(2),167-216.[2] E. Moschos, A. Stegner, O. Schwander and P. Gallinari, "Classification of Eddy Sea Surface Temperature Signatures Under Cloud Coverage," in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, vol. 13, pp. 3437-3447, 2020, doi: 10.1109/JSTARS.2020.3001830.[3] https://www.lmd.polytechnique.fr/dyned/

Download Full-text