Phosphorus analysis of floodplain sediments to reconstruct human impact and pristine conditions in a lowland river

The derivation of nutrient limits in freshwater needs information on the natural background of these nutrients. This study uses floodplain sediment P concentrations to identify human impact over the past 7,500 years and to infer the pristine P concentrations in sediment and water. Floodplain sediment cores from three sites in the lowland Dijle river catchment (BE) were analysed and radiocarbon dated. The sediment degree of P saturation (DPS) and oxalate extractable P (Pox) were correlated with two proxies of human activity: archaeological site count index (SCI) and pollen signal-based human impact (HI). In addition, sediment DPS was related to total dissolved P (TDP) concentrations in the water, calibrated with present-day data. Significant positive correlations (p < 0.001) were found among the three proxies across the sites between 4,500–1,500 Before Present (BP). After 1,500 BP the SCI increased exponentially, the DPS increased as well, but starting from 500 BP, it decreased towards 150 BP, to peak in most recent times. The impact of pre-and early complex societies on sediment DPS was small until 2,000 BP. During the Roman Period (ca. 2000–1500 BP), an increase in DPS indicated cultural eutrophication, coinciding with vegetation change and a rise in archaeological site density. The estimated pristine TDP concentration in the river was 28 µg TDP/L [95 % CI: 24; 42], an order below current concentrations. Upland peats in the headwaters likely had TDP concentrations < 10 µg TDP/L. This study illustrated the potential of floodplain sediment P analysis to estimate human impact and natural background P concentrations at a catchment scale. However, DPS should not be used as a proxy for human activity in organic layers where organic P dominates extractable P in the sediment.