The long term variability of the Black Sea physical processes (e.g. temperature, main circulation, cold intermediate layer, sea level) and its relation to atmospheric conditions and large scale climate patterns are investigated using an eddy-resolving tridimensional model in combination with statistical tools (e.g Empirical Orthogonal Functions, Self Organizing Maps). First, the ability of the model to represent the interannual dynamics of the system is assessed by comparing the modeled and satellite sea surface temperature (SST) and sea level anomaly (SLA) decomposed into their dominant Empirical Orthogonal Functions (EOFs). The correlation between the spatial and temporal EOFs modes derived from model and satellite is usually satisfactory and this gives some confidence in using the model as a tool to investigate not only the SST and SLA dynamics but also the dynamics of connected variables.Then, the long term variability of the Black Sea hydrodynamics is assessed by decomposing into their dominant EOFs modeled SST, SLA and selected key hydrodynamical variables associated to the main circulation and vertical structure of the water column. Significant correlations between the EOFs associated to these variables are investigated in order to link the variability of surface fields and the internal dynamics of the system.In particular, the intensity of the general cyclonic circulation (the Rim current) is shown to impact strongly on (1) the mean sea level, (2) the SST response to air temperature (AT), (3) In order to appraise the variability of atmospheric conditions over the Black Sea during 1962-2000 and their role in driving the hydrodynamics, a self-organizing maps technique is used in order to identify spatial recurrent patterns of atmospheric fields (ie. AT, wind stress and curl).The impact on these patterns of large scale climatic variability over the north Atlantic, the Eurasia and Pacific Ocean (estimated by respectively the north Atlantic oscillation (NAO), the east Atlantic/west Russia oscillation (EA/WR) and the El Niño southern oscillation (ENSO) indexes) is assessed. Distinct time scales of influence of the large scale teleconnection patterns on the AT are identified : EA/WR drives the short scale (1-5 years) variations of SST, while the long term (>5 year) trends of the NAO drive the long term SST trends.The drastic changes that have occurred in the Black Sea deep sea ecosystem at the end of the 80's are connected to an intensification of the general circulation which has promoted an export of riverine materials from the eutrophicated north-western shelf to the deep sea.Finally, in the last two decades, we find an increased duration of persistent atmospheric anomalies regime that have the potentials to drive the system away from its average state as occurred in the late 80s. If persistent in the future, such long lasting atmospheric anomalies may have a significant impact on the ecosystem functioning.