Michel Gay - Academia.edu (original) (raw)

Papers by Michel Gay

Journal of Glaciology, 1999

A digital image-processing approach is proposed which allows the extraction of two-dimensional po... more A digital image-processing approach is proposed which allows the extraction of two-dimensional polycrystalline ice microstructure (grain boundaries) from thin sections observed between cross-polarisers. It is based on image segmentation of colour images. The method is applied to the preliminary analysis of the shallow ( Holocene) ice of the European Project for Ice Coring in Antarctica (EPICA) ice core at Dome Concordia. Structural parameters, such as the mean cross-sectional area, shape anisotropy and grain morphology, are obtained. The interest and limitations of this automatic procedure are discussed.

$Research paper thumbnail of Fracturing of ice under compression creep as revealed by a multifractal analysis$

Journal of Geophysical Research, 1998

Annals of Glaciology, 2000

The shallow-ice microstructure at Dome Concordia, Antarctica, has been studied between 100 m and ... more The shallow-ice microstructure at Dome Concordia, Antarctica, has been studied between 100 m and 580 m. An original digital-image-processing technique has been specially developed to extract ice microstructure (grain boundaries) from thin sections prepared during the two first scientific EPICA field seasons (1997/98 and 1998/99). Using this, not only the mean crystal size, but also crystal-size distributions and shape anisotropy were determined. The mean crystal-size profile as well as crystal-size distributions reveal normal grain growth up to 430 m. Between 430 m and 500 m, a marked decrease of crystal size is observed and compared with a similar trend obtained in the "old" Dome C ice core formerly associated with the Holocene/Last Glacial transition (Duval and Lorius, 1980). This seems to indicate a slightly lower accumulation rate (by <10%) at Dome C. The shapes of the crystal-size distributions, though very similar, do evolve with depth and seem to be sensitive to ...

International audienceAt the Zackenberg site, sea ice starts to move between June and September r... more International audienceAt the Zackenberg site, sea ice starts to move between June and September resulting in icebergs flowing freely on the sea. Splitting into smaller parts, they reduce in size. Icebergs represent a risk for maritime transport and needs to be studied. In order to determine iceberg density per surface unit, size distribution, and movement of icebergs, we need to observe, detect, range and track them. The use of SAR images is particularly well adapted in regions where cloud cover is very present. We focused our study on the Yung Sund fjord in Greenland, where lots of icebergs and sea ice are generated during the summer. In the beginning of July, sea ice breaks up first, followed by icebergs created by the different glaciers based in the ocean. During our investigation, we noticed that the iceberg and sea ice were drifting very fast and thus, we needed to adapt our methodology. To achieve our goal, we collected all remote sensing data available in the region, principa...

snow and radar images are distorted by foldover and shadow due to the relief [1]. These irreversi... more snow and radar images are distorted by foldover and shadow due to the relief [1]. These irreversible degradations affect the data which have to be rectified in order to be combined either in optical geometry or in radar geometry. Some processing softwares, such as those dedicated to differential SAR interferometry (D–InSAR), include geocoding tools which allow to resample SAR data in ground geometry. The inverse operation consists in projecting georeferenced data such as optical images or ground measurements in each specific radar geometry. To perform it for SAR images in mountainous areas is not straightforward since the two geometries are not linked by a bijective transformation. Specific tools, which takes SAR sensor geometry into account, are required. Very few softwares offer the end-user with convenient tools to perform this task. In this paper, a convenient and generic approach is proposed to perform direct and inverse geocoding for the use of SAR images in high relief areas....

HAL is a multi-disciplinary open access archive for the deposit and dissemination of scientific r... more HAL is a multi-disciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. Efficient Bayesian data assimilation via inverse regression Benoit Kugler, Florence Forbes, Sylvain Douté, Michel Gay

IGARSS 2020 - 2020 IEEE International Geoscience and Remote Sensing Symposium, 2020

From simple scattering mechanism extraction and throughout more complex applications (e.g., land ... more From simple scattering mechanism extraction and throughout more complex applications (e.g., land classification, disaster monitoring), polarimetry has become a key element for remote sensing. For the particular case of bistatic polarimetry, the development of a theoretical basis has not been yet aligned with a comprehensive experimental validation. At the moment, an exhaustive search across the polarimetric scientific literature will reveal that for true bistatic geometries (i.e., significant angular separation between transmitter and receiver), only a small number of qualitative investigations have been made and there is still work to be done. In the current paper, one of the most popular polarimetric decomposition methods (H -α) is applied to dual-pol VV-VH data, in both bistatic (space-surface geometry with ground-based receiver) and monostatic configurations. Images from both geometries are displaying a common, urban scene. Comparing the obtained results, objective observations ...

La majeure partie de l'eau douce de la Terre est stockee dans les calottes polaires et 1 les ... more La majeure partie de l'eau douce de la Terre est stockee dans les calottes polaires et 1 les vallees glaciaires. Les bilans nets entre les composantes d'entree et de sortie des masses de 2 neige et de glace dependent des precipitations, de l'ablation, fonte, sublimation, et evaporation. 3 Ils reagissent fortement au rechauffement atmospherique et aux changements du regime des 4 precipitations. Stocker et al. (2013) notent que l'incertitude la plus importante dans l'estimation 5 de la perte de masse de glace provient de l'Antarctique et que les donnees d'observation des 6 interactions glace-ocean autour des calottes polaires de l'Antarctique et du Groenland restent faibles. 7 Les mesures d'observation satellitaire Radar a Synthese d'Ouverture (RSO) obtenues par "tracking" 8 ou interferometrie sont une reponse sans precedent a la determination des changements d'elevation 9 et des deplacements de la surface de la glace et des glacie...

European geosciences union general assembly, 2016

This paper presents a multilayer snowpack electromagnetic backscattering model, based on Dense Me... more This paper presents a multilayer snowpack electromagnetic backscattering model, based on Dense Media Radiative Transfer (DMRT). This model is capable of simulating the interaction of electromagnetic wave (EMW) at X-band and Ku-band frequencies with multilayer snowpack. The air-snow interface and snow-ground backscattering components are calculated using the Integral Equation Model (IEM) by [1], whereas the volume backscattering component is calculated based on the solution of Vector Radiative Transfer (VRT) equation at order 1. Case study has been carried out using measurement data from NoSREx project [2], which include SnowScat data in X-band and Ku-band, TerraSAR-X acquisitions and snowpack stratigraphic in-situ measurements. The results of model simulations show good agreement with the radar observations, and therefore allow the DMRT model to be used in various applications, such as data assimilation [3].