Cold-Water Corals (CWC) are filter-feeding heterotrophic cnidarians that dwell in relatively cold and deep marine environments and support rich ecosystems. They exhibit a global distribution but have been particularly well-studied in the NE Atlantic Ocean where framework building species like Lophelia pertusa have constructed impressive CWC “mounds” at a wide range of spatial scales. Historically, CWC mounds were believed to rely on primary productivity associated with seafloor hydrocarbon seepage as early discoveries by oil and gas exploration companies often found them near seepage related structures like pockmarks. However, isotopic evidence has since revealed that CWC primarily feed on surface derived phyto-detritus and zooplankton. In the NE Atlantic region, the CWC mounds appear to cluster in mound provinces which are predominantly located below the winter mixed layer, at the depth of the permanent thermocline (500-1000 m). The strong vertical density gradient (pycnocline) at this depth interval amplifies the bottom currents generated by the interaction of oceanic motions (e.g. tides and currents) with the steep continental slopes. This hydrodynamic regime benefits the CWC mounds by increasing the surface derived organic matter flux to the coral polyps. As this bottom current regime also influences sedimentary processes, CWC mounds are often associated with characteristic sedimentary bedforms and contourite drifts. The global distribution, the good preservation potential and the ability to derive radiometric ages and geochemical proxies from the aragonite skeleton of reef-forming CWC, makes them a promising palaeo-environmental archive for studying climate-driven changes in the ocean dynamics. On the Bay of Biscay shelf-break and upper slope, several provinces of small (1-5 m high and 50-250 m in diameter) CWC “mini-mounds” have been discovered. These mound provinces appear to lack modern coral growth and occur at relatively shallow (250-500 m) depths above the contemporary permanent thermocline. Their position on the shelf-break and upper slope also puts them at risk of destruction by trawl fishing. These enigmatic fossil mini-mounds thus offer great potential to investigate the influence of both climatedriven changes in ocean dynamics as well as anthropogenic impact on the initiation, development and demise of CWC mounds. Furthermore, a newly discovered mini-mound province on the NW Iberian margin occurs in close proximity of a pockmark field which allows to shed more light on the controversial connection between CWC mounds and hydrocarbon seepage processes. A combination of geophysical (multibeam bathymetry, backscatter and seismic reflection profiling), oceanographic (CTD) and video data were used to investigate the geomorphological features and oceanographic processes in and around the mini-mound province located on the outer Ortegal Spur and around the Ferrol Canyon head (NW Iberian margin). The Ferrol Canyon head is characterised by erosional (erosional and abraded surfaces, contourite channels and furrows), depositional (contourite drifts and sediment waves) and mixed (contourite terrace) sedimentary features, indicating a dominant control of bottom currents on the sedimentary processes. The bottom currents are related to the interaction of the canyon head topography with both the Mediterranean Sea Water (MSW) contour current and with M2 internal tides associated to the pycnocline at the interface between the Eastern North Atlantic Central Water (ENACW) and the MSW. Using a semi-automated mapping approach, over 171 mini-mounds with a diameter of 60-190 m and a mound height of 1.5-3.1 m were mapped on the Ortegal spur contourite terrace at 400-560 m depth. Preliminary video observations indicate that the mounded features are associated with CWC (mainly L. pertusa) rubble fields, identifying them as fossil CWC mounds. These mounds predominantly occur above the ENACW-MSW interface and the associated erosional bottom current regime. As this regime is thought to favour CWC growth, the existence of the mini-mounds suggests this interface may have been up to 200 m shallower in the past. This hypothesis is corroborated by the presence of the Ortegal Spur contourite terrace, a sedimentary feature often associated with water mass interfaces. Using semi-automated mapping and multivariate morphometric analysis (PCA and PERMANOVA), a significant morphological difference was found between a group of relatively large and scattered mini-mounds, adjacent to the Ortegal Spur pockmark field, and a group of smaller, clustered mounds on the other side of the Ferrol Canyon head. In contrast, no significant morphological difference, other than elevation( mounds)/depression(pockmarks), exists between the adjacent mini-mounds and pockmarks. This suggests that pre-existing pockmarks may have been colonised by CWC as methane-derived authigenic carbonate cementation, often associated with these seepage features, offered a preferential firm settling ground to coral larvae. Seismic facies signatures indicative of fluid flow occur below both the pockmarks and some of these mini-mounds supporting this hypothesis. The smaller, clustered mounds are similar in morphology to CWC mini-mounds in other provinces where seafloor seepage is absent and may have resulted from CWC colonization of smaller firm features like dropstones. The timeframe of CWC mini-mound development on the Bay of Biscay shelf-edge and upper slope was determined by U/Th dating of L. pertusa samples. The most complete record was recovered from the Celtic margin where an entire mound sequence was sampled (30 samples) in addition to 7 samples collected from core tops distributed throughout the mini-mound province. The 5 existing L. pertusa ages from seabed samples on the Armorican margin mini-mounds were supplemented with 7 additional ages. From the NW Iberian margin, only 3 polyps from a single L. pertusa coral branch were dated. The main phase of CWC mound growth on the Celtic margin occurred in the early to mid- Holocene (9-6 ka BP) followed by an apparent CWC demise and recolonization during the late Holocene (1.17±0.02 ka BP up to 0.15±0.02 ka BP). The CWC mini-mounds on the Armorican margin appear to have developed simultaneously since most samples date back to the early to mid-Holocene (9.6-7 ka BP). Two samples from 1.4±0.2 ka BP and 0.026±0.018 ka BP (1924±18 AD) suggest late Holocene recolonization also occurred in this setting. The L. pertusa branch recovered from the NW Iberian margin indicates that it grew during the early-Holocene (9.3-9.6±0.03 ka BP) but further sampling is necessary to better frame the CWC mound development in this province. In order to investigate the palaeoceanographic processes active during the early to mid- Holocene episode of CWC mound growth on the Celtic margin, multi-proxy time series of εNd, bottom water temperature and Δ14C were determined on the U/Th dated L. pertusa samples. The record displays significant sub-millennial scale variability in the εNd (- 11.70±0.7 to -14.70±0.4) and Δ14C (23[13:33] to -89[-96:-82] ‰) signatures of the NE Atlantic winter mixed layer, accompanied by mild shifts in temperature (9.8±0.9 to 11.4±0.9 °C). These fluctuations most likely represent episodes of intensified canyon-driven upwelling of poorly ventilated water to the CWC mini-mounds on the shelf-edge. Analogous to the modern oceanographic processes along this margin, the upwelling episodes appear to be driven by dynamic changes in the slope current direction under persistent positive North Atlantic Oscillation (NAO) atmospheric forcing. This interpretation is in good agreement with early to mid-Holocene proxy reconstructions of SPG circulation and continental precipitation patterns, both driven by the NAO. However, upwelling alone cannot explain the mid-Holocene occurrence of a 14C depleted and unradiogenic (in εNd) water mass which has no modern analogue. The unradiogenic εNd signature excludes a southern origin of the poorly ventilated water (e.g. Antarctic Intermediate or Bottom Water), suggesting a North Atlantic source instead. In order to reconcile this observation with the well-documented mid-Holocene intensification of the Atlantic Meridional Overturning Circulation (AMOC), this work proposes the hypothesis that the marked mid-Holocene increase in Nordic Sea convection depth flushed the remnants of an extremely 14C depleted abyssal Arctic Ocean reservoir and propagated its depleted signal throughout the mid-depth North Atlantic. The temporal distribution of dated L. pertusa samples from the Celtic and Armorican margin mini-mounds also indicates sub-millennial scale variability in coral abundance, suggesting an important control of natural processes on CWC growth. Episodes of peak coral abundance occur at 9.0, 8.5, 7.3, 6.9, 6.5 and at 0.30 ka BP which appear to coincide with periods of positive NAO conditions as demonstrated by continental rainfall proxy records. The demise of the sampled CWC mini-mound on the Celtic margin also coincides with a mid-Holocene atmospheric reorganisation. While our limited dataset can only provide a first indication of the true temporal distribution of these corals and while such a correlation does not necessarily imply a causation, the large-scale impact of the NAO atmospheric forcing on primary productivity and mixed layer depth (MLD) merits the hypothesis that the NAO may drive variability in CWC abundance on the Bay of Biscay mini-mounds. Analogous to the modern Bay of Biscay, positive NAO conditions may have induced a shallower MLD (and permanent thermocline) and more persistent upwelling which would have favoured CWC growth on the mini-mounds by driving an earlier and more persistent seasonal phytoplankton bloom over the shelf-break. The occurrence of upwelling during early to mid-Holocene episodes of positive NAO is corroborated by the shifts in water ventilation recorded in the coral Δ14C signature. Persistent positive NAO conditions may also have intensified the rate of deep water formation in the AMOC which would cause a further shoaling of the permanent thermocline and the associated hydrodynamic regime, benefitting the relatively shallow CWC mini-mounds. The early to mid-Holocene phase of CWC mound development indeed seems to coincide with a period of overall enhanced AMOC. Finally, canyon-driven upwelling processes may also be important in enhancing the suspended particle concentration near the shelf-break. The discovery of L. pertusa samples from the last 400 years implies pre-industrial environmental conditions (up to 1924±18 AD) in the Celtic and Armorican margin mini-mound provinces were able to sustain CWC growth. As these provinces lie on the shelf break, a region which has been subjected to significant fishing activity, the present-day lack of living CWC may be attributed to habitat destruction by trawl fishing. Fortunately, the Celtic margin mini-mound province has been designated as a Marine Conservation Zone in 2013 and also the Armorican margin mini-mounds lie within a region proposed as a Special Area of Conservation since 2014. In contrast, no conservation efforts have been undertaken in the newly discovered NW Iberian mini-mound province where fishing pressure is exceptionally high. However, even if conservation efforts are successful in mitigating additional damage, evidence from other regions suggests CWC reefs take more than a decade to recover and renewed coral colonization may be absent in sub-optimal environmental conditions. Given that the NAO and AMOC are projected to weaken in the coming decades as a result of anthropogenic climate change and given that rising ocean temperatures may push the shallow mini-mound provinces outside of the temperature tolerance range of L. pertusa, renewed coral colonization appears unlikely. Nevertheless, protection of the mini-mound provinces are warranted as the rubble fields are a habitat to other marine species and sediment resuspension by trawling on the shelf-break may adversely affect the last remaining live CWC reefs within the canyons of the Bay of Biscay.