Subsurface flow and phosphorus dynamics in beech forest hillslopes during sprinkling experiments: how fast is phosphorus replenished?

The phosphorus (P) concentration of soil solution is of key importance for plant nutrition. During large rainfall events, the P concentration is altered by lateral and vertical subsurface storm flow (SSF) that facilitates P mobilization, redistribution within the soil profile and potential P export from the ecosystem. These processes are not well studied under field conditions. Important factors of the replenishment of P concentrations in soil solutions are the rate of P replenishment (by biotic and abiotic processes) and the P buffering capacity of soils. Lab experiments have shown that replenishment times can vary between minutes and months. The question remains of how P concentrations in lateral and vertical SSF vary under natural field conditions. We present results of large-scale sprinkling experiments simulating 150 mm throughfall at 200 m2 plots on hillslopes at three beech forests in Germany. We aimed at quantifying lateral and vertical SSF and associated P concentrations on the forest floor, in the mineral soil and in the saprolite during sprinkling experiments in spring and summer. The sites differed mainly in terms of soil depth, skeleton content and soil P stock (between 189 and 624 g/m2 in the top 1 m soil depth). Vertical SSF in the mineral soil and in the saprolite was at least 2 orders of magnitude larger than lateral SSF at the same depth. Vertical and lateral SSF consisted mainly of pre-event water that was replaced by sprinkling water. Higher P concentrations in SSF in the first 1 to 2 h after the onset of SSF indicated nutrient flushing, but P concentrations in the mineral soil and saprolite were nearly constant thereafter for most of the experiment despite a strong increase in SSF. This suggests that P in the soil solution at all three sites was replenished fast by mineral or organic sources. If chemostatic transport conditions would dominate in SSF, annual P losses at the lateral and vertical boundary of a forest plot could be approximated by knowing the average P concentration and the water fluxes in forest soils. A rough estimation of the annual P loss based on this simplified assumption for one of our sites with longer SSF data resulted in an annual P loss of 3.16 mg/m2/a. This P loss is similar to estimates from a previous study at the same site using bi-weekly groundwater samples. Our approximated annual P loss in SSF was in a similar order of magnitude as P input by dry and wet deposition and by mineral weathering. Despite the f