The Holocene's climate variability, influenced by solar activity, the North Atlantic Oscillation (NAO), and North Atlantic gyre dynamics, is well-documented. However, disentangling natural variability from anthropogenic impacts (deforestation and land use changes) remains a challenge. This study focuses on the last 2500 years in north-western France, investigating the interplay between climate and human activity using palynological proxies (pollen and dinoflagellate cysts). The research question centers on identifying environmental changes and exploring their implications for coastal societies during this period, aiming to provide insights relevant not only to the specific region but also to broader understanding of long-term socio-ecosystemic trajectories. The study's purpose is to analyze high-resolution coastal records near continental sources to understand climate-society interactions during the pre-industrial period, enhancing understanding of contemporary changes by placing them within a longer-term context of natural and anthropogenic forcings. The macro-estuarine sedimentary environment of the Bay of Brest offers a unique archive for investigating this complex interplay.

Existing literature highlights the Holocene's millennial-scale climate variability, characterized by solar-related (2500-year) and ocean-related (1500-year) cycles. Atlantic subpolar gyre (SPG) dynamics are identified as a key driver, modulating heat transport and influencing North Atlantic sea surface temperatures. The NAO's role in shaping regional climate variability is also emphasized. Previous palynological studies have revealed human-environmental interactions but often lack the resolution to capture the nuances of pre-industrial societies' responses to environmental change. Existing archaeological evidence hints at social and economic instability during the late Roman period in the region, but the interplay between climate and societal collapse remains unclear. This study aims to address this gap by integrating detailed palynological analysis with existing climate and archaeological data.

Two high-resolution sediment cores (G and KS-02) from the Bay of Brest, spanning the last 2500 years, were analyzed. The composite core, with a temporal resolution of approximately 35 years, was generated using radiocarbon dating (14C-AMS) of gastropods, calibrated using the Marine13 curve and accounting for regional reservoir effects (ΔR estimated at 40 ± 23 years). Palynological analyses involved processing sediment samples (<150 µm fraction) using chemical and physical methods to extract and concentrate pollen and dinocyst assemblages. A minimum of 300 pollen grains and 150 dinocysts were counted per sample to ensure statistical robustness. Pollen and dinocyst percentages were calculated, and concentrations were determined using the Lycopodium marker-grain method. The results were then compared with existing paleoclimatic data from the North Atlantic, including storm activity records, SPG strength indicators (water density differences and benthic foraminifera δ18O), precipitation quantifications from speleothems, and Loire River discharge proxies (Ti-XRF).

The analysis revealed a striking '1.7-1.4 ka AP event' – a 300-year period (250-550 AD) of significantly increased tree pollen percentages (exceeding even Mesolithic levels), indicative of extensive forest regrowth. This occurred during a period of increasing landscape opening, making it an atypical phenomenon. The increase in tree pollen, especially *Corylus* (hazel), and riparian species (*Alnus* and *Salix*), along with the dinocyst *Lingulodinium machaerophorum*, points to wetter conditions in north-western France. This contrasts with evidence from the Bay of Biscay and northern Spain indicating drier conditions during the same period. These contrasting trends suggest a differential north-south climate pattern, akin to the modern NAO. Interestingly, the 1.7-1.4 ka AP event coincided with a sharp decrease in *Cerealia*-type pollen (cultivated plants) and anthropogenic pollen indicators (API), suggesting agricultural abandonment and a potential collapse of the agrarian system. The increased fluvial input during this period, as evidenced by higher pollen fluxes and *L. machaerophorum* percentages, may be linked to increased storminess associated with SPG strengthening and a northward shift of the westerly wind belt. After 800 years BP, the pattern reversed, with decreased tree pollen and increased Ti-XRF signals (indicative of soil erosion from deforestation) suggesting increased human activity following the Medieval Climatic Optimum.

The findings suggest that the '1.7-1.4 ka AP event' was likely driven by a combination of factors. Strengthened SPG and NAC, leading to increased northward heat transport and modified atmospheric circulation (Northward positioning of westerly wind belt), produced warmer, wetter winters. This favorable climate supported significant forest regrowth. The simultaneous decline in agricultural indicators, however, suggests that this forest regrowth may have been partly due to agricultural abandonment, perhaps exacerbated by climatic deterioration (e.g., increased storminess and flooding) and the existing socio-political instability of the late Roman Period. The contrasting humidity patterns between northwestern and southwestern Europe during this period could be explained by a NAO-like pattern, influencing precipitation regimes differentially across the region. The study highlights a dynamic interplay between climate, human activity, and societal resilience.

This study demonstrates a significant and unexpected forest revival in northwestern Europe during the late Roman period. This '1.7-1.4 ka AP event' likely resulted from a combination of climatic changes (increased humidity associated with SPG strengthening and altered atmospheric circulation) and a concomitant collapse of the coastal agrarian system, potentially due to the intensification of storminess and other socio-political factors. Further research could focus on more detailed archaeological investigations to refine the understanding of societal responses to this climate-driven change and on extending similar palynological analyses across a wider geographical area to better understand the spatial extent and variability of this phenomenon.

The study focuses on a limited geographical area in northwestern France. While the comparison with other regional paleoclimate data adds context, it would be beneficial to expand the study area to enhance the generalizability of the findings. The reliance on palynological data, while effective for vegetation changes, might not fully capture the complexities of societal responses to climate variability. Further integration with archaeological and historical data would strengthen the interpretations. The precision of the dating method (14C-AMS on marine gastropods) has some limitations related to regional reservoir effect uncertainty that might lead to minor temporal inaccuracies.