L’albédo spectral comme outil permettant d’estimer la propagation du rayonnement solaire dans et sous la banquise

The interaction between solar radiation and the sea ice cover plays an important role for polar climate and ecosystems. Solar radiation absorbed within sea ice is an important component of its energy budget and contributes in large part to surface melting during spring. Solar radiation transmitted at the bottom of the ice and in the ocean determines the growth of the photosynthetic organisms which forms the basis of the polar marine food chain. While this interaction, named sea ice solar radiative transfer, plays a key role for polar climate and ecosystems, its seasonal evolution in response to environmental forcing is neither sensed nor understood in a fundamental way. In particular, the process of scattering, fundamental to the description of light behavior in sea ice, cannot be monitored effectively with the current methods. This process of scattering describes how light bounces on the multiple interfaces of snow and sea ice porous microstructure as it travels through it. On the one hand, the scattering properties evolution throughout the season is not directly accounted for when estimating under-ice light availability for ecosystems by satellite. On the other hand, with no effective measurement method, the spatial and temporal coverage of scattering properties evolution is sparse. Consequently, the way in which environmental forcing successively affects the microstructure, the scattering properties, and finally absorption and transmission is still not well understood and parametrized in large-scale numerical models. In this thesis, we aim to use spectrally resolved albedo in the visible range as a tool to estimate the vertically resolved scattering properties and predict light propagation in and through sea ice. Spectral albedo is easy to measure and non-destructive. It has a wide spatial and temporal coverage both from the ground and from satellite. Thus, vertically resolved scattering properties inverted from spectral albedo could be used to improve remote under-ice light predictions and help parametrize radiative transfer in a more fundamental way. To demonstrate the validity of the technique we relied on a lookup table of Monte Carlo simulations covering the variability of snow-covered and bare first-year sea ice from winter to summer. In the first chapters, we used the simulated lookup table to demonstrate that spectral albedo contains information on the vertically resolved scattering properties of snow and /or sea ice above the freeboard. We demonstrated, using the framework of optical thickness, that properties above freeboard are sufficient to predict solar heat deposition in and transmittance under first-year sea ice in most scenarios. In the second chapter, we present an inversion algorithm relying solely on spectral albedo to provide vertically resolved scattering properties and estimate under-ice transmittance. To do so, the algorithm compares spectral albedo to simulations in the lookup table. The algorithm was validated from the ground using data from 5 field campaigns. In one of these campaigns, we demonstrated that scattering properties from spectral albedo corresponded to in situ measurement with the active probe. For the five campaigns, transmittances obtained from spectral albedo were in good agreement with measurements and better than assessments from the state-of-the-art method. This improved performance is explained by the ability to sense the scattering properties. In the third chapter, we obtained vertically resolved scattering properties of sea ice over a complete season by inverting spectral albedo from autonomous stations in central Arctic. We demonstrated that the scattering properties drop by half when the snow melts, and by another half when snow disappears, leaving a bare ice surface. Aside from those specific events, we could not show any further relation between evolution of scattering properties, temperature and snow age.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.