Description: The Baltic Ice Lake, Yoldia Sea, Ancylus Lake and Littorina Sea stages of regression and transgression controlled the history of the Baltic Sea since the last glacial period. Many details regarding their development remain unknown, including the question whether the regression of the Ancylus Lake (between 10,700 and 10,200 cal yr BP) took place west of the Darss Sill, or elsewhere. This study addresses whether a drowned river system in Fehmarn Belt (SW Baltic Sea) can be related to the drainage of the Ancylus Lake. The river channel is cut into glacial till in the western part of Fehmarn Belt. Here, the channel reaches an incision depth of up to 12 m at a base level of 40 m b.s.l. (below sea level). Near Mecklenburg Bay, the buried channel is incised down to 60 m b.s.l. and widens rapidly from several hundred meters to more than 1 km, fading towards east. It was mainly shaped as part of a glacial meltwater system at a maximum water level of 30 m b.s.l., and is therefore not related to the Ancylus Lake drainage. During the lowstand of the Baltic Ice Lake, local shallow water bodies covered the study area. A previously reported westward directed drainage of a lake in the eastern Fehmarn Belt could be restricted to a time interval between the highstand of the Ancylus Lake and prior to the Littorina Transgression. Timing, water level and potential water discharge of this event suggest a connection to a partial drainage of the Ancylus Lake. It is concluded that the threshold separating the Ancylus Lake from the North Sea needs to be located west of Fehmarn Belt. Highlights ► A paleo-river channel is incised into the seafloor of Fehmarn Belt, SW Baltic Sea. ► This channel is unrelated to the drainage of the Ancylus Lake through the Dana River. ► Widespread erosion took place after the Ancylus Lake phase of the Baltic Sea. ► This event relates to a partial drainage of the Ancylus Lake. ► A new threshold of the Ancylus Lake west of Fehmarn Belt is suggested.

Description: Highlights: • A first standardized and publicly available Holocene relative sea-level database for the Baltic Sea is presented. • The database holds 1099 revised data points with an estimation of vertical and chronological uncertainties. • Negative RSL tendencies prevail over the positive and complex tendencies in the Baltic Sea Basin. • Mid-Holocene RSL highstand occurred around 7.5–6.5 ka BP being consistent with the end of the final melting of the LIS. • The contribution of ice loading in the eastern Baltic Sea Basin is likely overestimated in the ICE-5G and ICE-6G_C models. Abstract: We present a compilation and analysis of 1099 Holocene relative shore-level (RSL) indicators located around the Baltic Sea including 867 relative sea-level data points and 232 data points from the Ancylus Lake and the following transitional phase. The spatial distribution covers the Baltic Sea and near-coastal areas fairly well, but some gaps remain mainly in Sweden. RSL data follow the standardized HOLSEA format and, thus, are ready for spatially comprehensive applications in, e.g., glacial isostatic adjustment (GIA) modelling. We apply a SQL database system to store the nationally provided data sets in their individual form and to map the different input into the HOLSEA format as the information content of the individual data sets from the Baltic Sea area differs. About 80% of the RSL data is related to the last marine stage in Baltic Sea history after 8.5 ka BP (thousand years before present). These samples are grouped according to their dominant RSL tendencies into three clusters: regions with negative, positive and complex (transitional) RSL tendencies. Overall, regions with isostatic uplift driven negative tendencies dominate and show regression in the Baltic Sea basin during the last marine stage. Shifts from positive to negative tendencies in RSL data from transitional regions show a mid-Holocene highstand around 7.5–6.5 ka BP which is consistent with the end of the final melting of the Laurentide Ice Sheet. Comparisons of RSL data with GIA predictions including global ICE-5G and ICE-6G_C ice histories show good fit with RSL data from the regions with negative tendencies, whereas in the transitional areas in the eastern Baltic, predictions for the mid-Holocene clearly overestimate the RSL and fail to recover the mid-Holocene RSL highstand derived from the proxy reconstructions. These results motivate improvements of ice-sheet and Earth-structure models and show the potential and benefits of the new compilation for future studies.