We review progress in Baltic Sea physical oceanography (including sea ice and atmosphere-land interactions) and Baltic Sea modelling, focusing on research related to BALTEX Phase II and other relevant work during the 2003-2014 period. The major advances achieved in this period are: Meteorological databases are now available to the research community, partly as station data, with a growing number of freely available gridded datasets on decadal and centennial time scales. The free availability of meteorological datasets supports the development of more accurate forcing functions for Baltic Sea models. In the last decade, oceanographic data have become much more accessible and new important measurement platforms, such as FerryBoxes and satellites, have provided better temporally and spatially resolved observations. Our understanding of how large-scale atmospheric circulation affects the Baltic Sea climate, particularly in winter, has improved. Internal variability is strong illustrating the dominant stochastic behaviour of the atmosphere. The heat and water cycles of the Baltic Sea are better understood. The importance of surface waves in air-sea interaction is better understood, and Stokes drift and Langmuir circulation have been identified as likely playing an important role in surface water mixing in sea water. We better understand sea ice dynamics and thermodynamics in the coastal zone where sea ice interaction between land and sea is crucial. The Baltic Sea's various straits and sills are of increasing interest in seeking to understand water exchange and mixing. There has been increased research into the Baltic Sea coastal zone, particularly into upwelling, in the past decade. Modelling of the Baltic Sea-North Sea system, including the development of coupled land-sea-atmosphere models, has improved. Despite marked progress in Baltic Sea research over the last decade, several gaps remain in our knowledge and understanding. The current understanding of salinity changes is limited, and future projections of salinity evolution are uncertain. In addition, modelling of the hydrological cycle in atmospheric climate models is severely biased. More detailed investigations of regional precipitation and evaporation patterns (including runoff), atmospheric variability, highly saline water inflows, exchange between sub-basins, circulation, and especially turbulent mixing are still needed. Furthermore, more highly resolved oceanographic models are necessary. In addition, models that incorporate more advanced carbon cycle and ecosystem descriptions and improved description of water-sediment interactions are needed. There is also a need for new climate projections and simulations with improved atmospheric and oceanographic coupled model systems. These and other research challenges are addressed by the recently formed Baltic Earth research programme, the successor of the BALTEX programme, which ended in 2013. Baltic Earth will treat anthropogenic changes and impacts together with their natural drivers. Baltic Earth will serve as a network for earth system sciences in the region, following in the BALTEX tradition but in a wider context. (C) 2014 The Authors. Published by Elsevier Ltd.