A soil phosphorus dynamics (SPD) model

The dynamics of soil phosphorus (P) control its bioavailability. Yet, it remains a challenge to quantify soil P dynamics. Here, we developed a soil P dynamics (SPD) model. We then assimilated eight datasets of 426-day changes in Hedley P fractions into the SPD model, to quantify the dynamics of six major P pools in eight soil samples that are representative of a wide type of soils. The performance of our SPD model was better for labile P, secondary mineral P, and occluded P than for non-occluded organic P (Po) and primary mineral P. All parameters describing soil P dynamics were approximately constrained by the datasets. The average turnover rates were labile P 0.040 g g^-1 d^-1, non-occluded Po 0.051 g g^-1 d^-1, secondary mineral P 0.023 g g^-1 d^-1, primary mineral P 0.00088 g g^-1 d^-1, occluded Po 0.0066 g g^-1 d^-1, and occluded inorganic P 0.0065 g g^-1 d^-1, in the greenhouse environment studied. Labile P was transferred on average more to non-occluded Po (transfer coefficient of 0.42) and secondary mineral P (0.38) than to plants (0.20). Soil pH and organic C concentration were the key soil properties regulating the competition for P between plants and soil secondary minerals. The turnover rate of labile P was positively correlated with that of non-occluded Po and secondary mineral P. The pool size of labile P was most sensitive to its turnover rate. Overall, we suggest data assimilation can contribute significantly to an improved understanding of soil P dynamics.