Abstract: Obligate intracellular bacteria of the Rickettsiella genus are emerging as both widespread and biologically diverse in arthropods. Some Rickettsiella strains are highly virulent entomopathogenic agents, whereas others are maternally inherited endosymbionts exerting very subtle manipulations on host phenotype to promote their own spread. Recently, a variety of Rickettsiella strains have been reported from ticks, but their biology is entirely unknown. In the present study, we examined the incidence and diversity of Rickettsiella in 11 geographically distinct populations of the polar seabird tick Ixodes uriae. We found Rickettsiella in most tick populations with a prevalence ranging from 3 to 24 %. 16S ribosomal RNA (rRNA) and GroEL gene sequences revealed an unexpected diversity of Rickettsiella, with 12 genetically distinct Rickettsiella strains present in populations of I. uriae. Phylogenetic investigations further revealed that these Rickettsiella strains do not cluster within a tick-specific clade but rather exhibit distinct evolutionary origins demonstrating frequent horizontal transfers between distantly related arthropod species. Tick rearing further showed that Rickettsiella are present in eggs laid by infected females with no evidence of abortive development. Using this data set, we discuss the potential biological significance of Rickettsiella in seabird ticks. Most notably, we suggest that these organisms may not be pathogenic forms but rather use more subtle adaptive strategies to persist within tick populations.

Abstract: Glaciers of the southern hemisphere sub-polar regions between 45 and 60°S have declined dramatically over the last century. The islands of Kerguelen archipelago (49°S, 69°E) represent a unique location in regions where few data are available to understand glacier retreat. Situated at low altitudes and close to the ocean, their glaciers have shown particular sensitivity to atmospheric and oceanic variations. Thus, since the 1960s, the Cook Ice Cap (~400km²) has retreated spectacularly, losing 20% of its area in 40 years. The aim of my thesis was to assess the present and future state of the ice cap, and to understand the causes of this decline while putting them in a global context. To do so, a meteorological and glaciological network was set up in 2010 on Kerguelen archipelago and field campaigns have been carried out annually since then. Analysis of these measurements confirms the negative mass balance of Cook Ice Cap. In parallel, the study of the albedo over the whole ice cap from MODIS satellite images (MODerate resolution Imaging Spectroradiometer) gives us access to the evolution of the snow line since 2000, highlighting an important reduction of Cook Ice Cap accumulation area over the last decade. Mass balance modelling of the Cook Ice Cap using a degree-day model coupled to a simple ice motion routine further reveals that its retreat is mainly due to a strong decrease in precipitation over the Kerguelen Islands since the 1960s. In order to put the decline of the cryosphere on Kerguelen in a global context, climatic trends over the whole sub-polar regions are studied, revealing that the sub-Antarctic area is currently the one where glacier retreat is the strongest. To understand these variations, we analyse a complete set of field and satellite observations and modelling results : reanalyses, models from the CMIP5 (Coupled Model Intercomparison Project phase 5) experiment, atmospheric and oceanic temperature and precipitation observations, etc. The latter show warming and quasigeneralised drying of the whole 40°S – 60°S area, linked to the southward shift of storm tracks in response to the more frequent positive phases of the Southern Annual Mode (SAM). Recent glacier retreat on Kerguelen archipelago, and for other glaciers and ice caps located at similar latitudes, is thus mainly due to a deficit of accumulation caused by the SAM, and amplified by atmospheric warming. The future evolution of Cook Ice Cap mass balance is evaluated using the MAR (Modèle Atmosphérique Régional) model, forced at its boundaries by CMIP5 models. Recent mass balance simulations are first carried out using ERA-Interim and NCEP1 reanalyses, and compared to in situ observations. In parallel, one-year simulations are produced with the precipitation desagregation scheme SMHiL (Surface Mass balance High resolution downscaLing) on MAR outputs, at various scales, in order to evaluate the impact of downscaling on precipitation. An evaluation of CMIP5 models over the recent period against ERA-Interim is then carried out, considering certain key climatic variables. The model closest to ERA-Interim as well as the two most extreme models are then used to force the MAR model over the next century, and surface mass balance outputs are critically analysed. The analysis of the decline of the Kerguelen ice cap using different tools and techniques brought new insights on the link between glaciers and climate, highlighting the major role of the SAM, but also raised new questions.

Abstract: We combine the surface-elevation and surface-mass change derived from Envisat data and GRACE solutions, respectively, to estimate regional changes in air and ice content of the surface of the Antarctic Ice Sheet (AIS) between January 2003 and October 2010. This leads, upon certain assumptions, to the separation of the rates of recent snow-accumulation change and that of ice-mass change. We obtain that the height of ice in Thwaites and Pine Island glaciers sectors decreases ( ≤ − 15.7 cm / yr ) while that in the Kamb glacier sector increases (≥5.3 cm/yr). The central part of the East AIS is mostly stable while the whole Dronning Maud Land coast is dominated by an increase in snow accumulation. The Kemp land regions show an ice-mass gain that accounts for 67–74% of the observed rates of elevation change in these regions. A good agreement is obtained over 68% of the investigated area, mostly in the East AIS, between our estimated rates of snow accumulation change and the predicted rates of the monthly surface mass balance derived from a regional atmospheric climate model.