We present a synchronization of the NGRIP, GRIP and GISP2 ice cores onto a master chronology extending back to 104ka before present, providing a consistent chronological framework for these three Greenland records. The synchronization aligns distinct peaks in volcanic proxy records and other impurity records (chemo-stratigraphic matching) and assumes that these layers of elevated impurity content represent the same, instantaneous event in the past at all three sites. More than 900 marker horizons between the three cores have been identified and our matching is independently confirmed by 24 new and previously identified volcanic ash (tephra) tie-points. Using the reference horizons, we transfer the widely used Greenland ice-core chronology, GICC05modelext, to the two Summit cores, GRIP and GISP2. Furthermore, we provide gas chronologies for the Summit cores that are consistent with the GICC05modelext timescale by utilizing both existing and new gas data (CH4concentration and δ15N of N2). We infer that the accumulation contrast between the stadial and interstadial phases of the glacial period was ~10% greater at Summit compared to at NGRIP. The δ18O temperature-proxy records from NGRIP, GRIP, and GISP2 are generally very similar and display synchronous behaviour at climate transitions. The δ18O differences between Summit and NGRIP, however, changed slowly over the Last Glacial-Interglacial cycle and also underwent abrupt millennial-to-centennial-scale variations. We suggest that this observed latitudinal δ18O gradient in Greenland during the glacial period is the result of 1) relatively higher degree of precipitation with a Pacific signature at NGRIP, 2) increased summer bias in precipitation at Summit, and 3) enhanced Rayleigh distillation due to an increased source-to-site distance and a potentially larger source-to-site temperature gradient. We propose that these processes are governed by changes in the North American Ice Sheet (NAIS) volume and North Atlantic sea-ice extent and/or sea-surface temperatures (SST) on orbital timescales, and that changing sea-ice extent and SSTs are the driving mechanisms on shorter timescales. Finally, we observe that maxima in the Summit-NGRIP δ18O difference are roughly coincident with prominent Heinrich events. This suggests that the climatic reorganization that takes place during stadials with Heinrich events, possibly driven by a southward expansion of sea ice and low SSTs in the North Atlantic, are recorded in the ice-core records.