Process-based mass-balance modeling of soil phosphorus availability in a grassland fertilized with N and P

International audienceModeling changes in plant-available soil P in relation to P budgets should integrate the isotopic kinetic approach, which describes the dynamics of P ion transfer at the solid-to-solution interface. We tested a process-based mass-balance model that uses the quantity of P ions in solution, the diffusive P ions (Pr) in the solid phase, and the annual P budget to describe the soil P availability of a timothy (Phleum pratense L.) grassland that received additions of annual P and N fertilizer. An experiment was established on a gravely-sandy loam soil in 1998, with combinations of P (0, 15, 30, and 45 kg ha−1) and N (0, 60, 120, and 180 kg ha−1) applied annually from 1999 to 2006. An isotopic dilution analysis was performed on soils sampled in 2006 to calibrate the Freundlich kinetic equation which describes the dynamics of Pr transfer at the solid-to-solution interface as a function of time (t) and concentration of P ions in solution (Cp). Model simulations were performed over 9 years (1999–2007). Measurements of Cp from soils sampled between 2001 and 2007 were compared with simulated values to evaluate model performance. The amount of Pr was estimated for two transfer periods, of 2 and 3 months, to evaluate the extent and contribution of slow P ion reactions. The Freundlich kinetic equation was defined as: Pr = 7.78 × Cp0.41 × t 0.36 (with Pr < PrLIMIT, 192 observations, Adj. R 2 = 1.0, P < 0.001). Simulated Pr values were 74 and 84% of total inorganic soil P for the transfer periods of 2 and 3 months, respectively. For the two transfer periods, and for each combination of N and P additions, simulations accurately reflected the long-term effects of P and N fertilization on the trends of measured Cp with root mean square deviation (RMSD) between measured and simulated values of less than 0.17. Across P applications, the simulations were slightly improved with a 2-month transfer period for limiting N conditions (0 and 60 kg N ha−1; Y = 0.95X + 0.06, R2 = 0.76, RMSD = 0.08) and a 3-month transfer period for non-limiting conditions (120 and 180 kg N ha−1; Y = 0.86X + 0.04, R2 = 0.78, RMSD = 0.06). This approach needs to be tested in various soil types and diverse cropping systems because the estimation of Pr value can be quite sensitive to the extent of rapid and slow reactions, hence the transfer periods. For this gravely-sandy loam soil, the proposed approach accurately describes the functioning of P cycling and confirms the agronomic importance of solution and solid phase P ions in managed grasslands