LAUSR

Abstract Phosphorus (P) levels in eutrophic lakes with restricted external P inputs often show hysteresis resulting from sediment P diagenetic recycling. Hard water oxygenated lakes are often thought to be less susceptible to P release from sediment due to precipitation of P with calcium carbonate and P sorption on iron (Fe) oxyhydroxides in oxidized surface sediments. However, perpetuated eutrophic conditions in many hard water oxygenated lakes persist despite drastic reduction of external P inputs. What limits P sediment binding capacity and drives its release from sediments in such lakes is often less understood. Here we characterize sediment P diagenetic processes driving persistent internal loading in the Bay of Quinte, a hard water, polymictic embayment of Lake Ontario, Canada. We used a multifaceted approach, combining a comprehensive three-year study of sediment P binding forms, recent sediment stratigraphy and accumulation, high spatial resolution measurement of dissolved oxygen, redox potential and pH across the sediment-water interface and depth profiles of nutrient and metals concentrations in pore water. Despite oxygenated bottom water conditions P diffusive fluxes during summer are substantial (1–6.5 mg P/m2/day), representing as much as 50% of external P inputs. Estimated rates of benthic organic carbon degradation indicate that anaerobic organic P mineralization in anoxic sediments is the main driver of P release, which is strongly influenced by Fe reduction. Estimated rates of benthic organic carbon degradation indicate that anaerobic organic P mineralization in anoxic sediments is the main driver of P release. In surface sediment the redox sensitive P is the prevailing diagenetically reactive P phase, emphasizing strong coupling with ferric Fe. While this indicates substantial P sorption on Fe oxyhydroxides, their spare binding capacity is limited, as reducible Fe: P ratios are close to sorption limit. The calcium carbonate bound P increased with sediment depth at the expense of redox sensitive P, suggesting diagenetic sequestration into apatite with sediment burial. However, only 40% of diagenetically reactive P forms, which account for ~2/3 of the total P in surface sediment, are permanently retained during burial. Importantly, our data show that in fact significant P-flux occurs in shallow areas, with an apparent discordance with concurrent measures of long-term P release from the sediments. These results also demonstrate that short-term measures cannot be extrapolated to long-term estimates and vice versa.