Glacial Lake Outburst Floods (GLOFs) constitute one of the most notorious and destructive geohazards worldwide. They have occurred throughout recorded history and form a constant threat for local communities near glacierized regions. Although the recent deglaciation resulted in an increase in glacial lakes, both in size and number, worldwide, little is known about the possible relationship between climate variability and GLOF occurrence. Yet, it is generally assumed that GLOF frequency is currently increasing due to global glacier retreat. This is mainly due to a lack of continuous long-term flood records. Current records of GLOF occurrence, which are based on instrumental and historical data, are intrinsically limited to the last centuries. Consequently, the effect of climate change on GLOF occurrence and the likely evolution of GLOFs under future climate change scenarios remains unclear. However, a comprehensive understanding of the link between climate change, glacier variability, and GLOF occurrence is fundamental for future GLOF predictions and to improve flood hazard assessments. As in many other glacierized regions, GLOFs are a well-known phenomenon in the Patagonian Andes. They are particularly pronounced in the Baker region of Chilean Patagonia (47–48 °S), where repeated GLOFs occurred from the abrupt drainage of ice-dammed Cachet 2 Lake between April 2008 and November 2020. During these events, water from Cachet 2 Lake spills into Colonia River, a tributary of Baker River, and increases both river discharge and sediment suspended concentrations. Colonia GLOFs are able to block the regular Baker River flow and result in the inundation of large areas upstream of the Colonia-Baker confluence, such as the Valle Grande floodplain. Downstream, the Baker River triples in discharge and large amounts of sediment are transported, and ultimately deposited, in fjords. The repeated Baker River GLOFs during the 21st century and the location of the Baker River, which drains most of the eastern side of the Northern Patagonian Icefield (NPI) and therefore integrates meltwater from several lake-river systems, makes the Baker region ideally suited to investigate GLOFs and to study the impact of climate change on GLOF occurrence. To examine how GLOFs are recorded in fjord sediments, this study mapped the bathymetry of the head of Martínez Channel, i.e. the fjord in which the Baker River discharges, using multibeam echosounding. Results show that the subaquatic delta of Baker River is deeply incised by sinuous channels. The presence of sediment waves and coarser sediment within these channels suggest recent channel activity by turbidity currents. The latter is confirmed by sediment records collected at the head of the fjord, which reveal the presence of turbidites intercalated within silty background sediments, particularly on the delta plain in front of the main submarine channel. Although the turbidity currents are most likely generated by elevated river discharge and the associated relatively high suspended sediment loads, most turbidites are not related to GLOFs. Instead, they seem to represent other extreme discharge events, such as extreme precipitation or rain-on-snow events. By comparing geochemical and sedimentological results obtained on the sediment cores to the recent GLOF history of Baker River, we show that the recent 21st century Cachet 2 GLOF deposits can be distinguished from background sediments by their finer grain size and lower organic carbon content, reflecting the increased input of glacial sediments during GLOFs. In addition, the results obtained on the fjord sediment cores demonstrate that the 21st century GLOFs from Cachet 2 Lake, which occurred less than one year apart, are not recorded as individual layers but as units richer in sediment of glacial origin. This suggests that it is not possible to reconstruct GLOF frequency nor magnitude solely based on fjord sediments. Although 21 GLOFs from Cachet 2 Lake occurred between 2008 and 2017, the deposits with the clearest GLOF signature represent the initial events, implying that more glacial sediment was released during those first GLOFs, possibly due to lake-bed erosion. Consequently, it appears that sediment availability plays a more important role than flood magnitude in controlling GLOF deposit properties. Although GLOF frequency and magnitude cannot be accurately reconstructed using fjord sediments, high accumulation rates at the head of Martínez Channel highlight the potential of fjord sediment archives to establish pre-historical GLOF records at high temporal resolution. In addition, the bathymetric imagery and the sediment records obtained at the head of Martínez Channel show that site selection and multi-coring are fundamental to reconstruct the Baker River GLOF history, as fjord heads are dynamic sedimentary environments with rapidly migrating channels. Ideal locations to reconstruct GLOFs are found on the delta slope, away from any submarine channel influence. GLOF deposits are best identified close to the river mouth, as background sediments become progressively finer and less organic, thus more similar to GLOF deposits, with increasing distance from the lip of the Baker River delta. Given the unique context of the Baker River system, where a significant portion of the watershed is vegetated and where the fine and organic-poor signature of GLOF deposits clearly contrasts with the slightly coarser and organic background sediments, our results may only be applicable to fjord sediments from temperate regions. Distinguishing GLOF deposits from background sediments would likely be more challenging in high latitude fjords. Sediments deposited in floodplains constitute another faithful recorder of Baker River GLOFs. In the Valle Grande floodplain, which is located immediately upstream of the Colonia-Baker confluence, GLOFs are registered as organic-poor deposits intercalated within organic-rich background sediments. In contrast to marine archives, the sediments of the Valle Grande floodplain have lower accumulation rates, and can therefore be used to determine changes in GLOF occurrence on longer timescales (late Holocene). Based on four radiocarbon-dated sediment cores collected in the Valle Grande floodplain, our results show that high-magnitude GLOFs occurred intermittently in the upper Baker River watershed over the past 2.75 kyr. Two periods of increased flood activity occurred between approximately 2.57 and 2.17 cal kyr BP, and from 0.75 to 0 cal kyr BP. Comparison with independent proxy records of glacier variability reveals that these two periods of increased flood frequency match with Neoglacial advances. These advances seem to result from lower-than-average temperatures and wetter conditions. Based on these results, we suggest that there is a strong, yet indirect, link between climate variability and GLOF occurrence. We hypothesize that, on multi-millennial timescales, high-magnitude GLOFs from eastern NPI glaciers are more frequent at times when glaciers are larger and thicker, as such glaciers most likely form larger and stronger ice dams, which in turn are able to retain larger lakes. Our results therefore suggest that the probability that high-magnitude GLOFs occur decreases as glaciers thin and retreat. Conversely, the frequency of lower magnitude GLOFs tends to increase during glacier recession because of the rapid growth of glacial lakes and formation of new lakes. Although isolated cases of new lakes formed behind large glaciers could still produce large GLOFs locally, the likelihood of large lake drainage and therefore high-magnitude GLOF occurrence decreases. This study supports the use of sediment-based GLOF records in other GLOF-prone regions for proper flood hazard assessment. A broader knowledge of the impact of climate change on GLOF occurrence can help to prevent further development in flood-prone regions and will reduce the vulnerability of communities to floods. Long-term paleoflood records can be of great importance for integrating spatial planning and planned infrastructure projects, such as hydroelectric dams, particularly since electricity demand is increasing with economic growth.