Modeling of phosphorus dynamics in contrasting agroecosystems using long-term field experiments

Long-term field experiments on phosphorus (P) fertilization were originally designed to study crop needs in different soil types by analyzing the effects of several rates of P fertilization on yields, their P concentrations and dynamics of plant-available soil P. The objective of this study was to test a computer-based model to simulate the P dynamics at the field scale using plant database and analyzing for plant-available P by a hierarchical process-based approach. It predicts both the concentration (C-P) of phosphate ions (Pi) in soil solution and the associated Pi amounts that in time equilibrate with Pi in solution. Five experiments, representative of contrasting soil types, land-use, and climates were selected. Our model equilibrates the change in plant-available P in the upper soil layer to the P budget between annual P inputs and outputs. Rates of P fertilization affected simulations following the same expected pattern across sites. Field-observed and simulated values are in good agreements in all sites. The field-observed variations of C-P per unit of P budget ranged from 0.007 to 2.49 (mu g P L-1) (kg P ha(-1))(-1). The predictions are of the same order of magnitude. Predictions were compared with empirical long-term data and mismatches were discussed. This investigation highlights the scientific interest of long-term field P experiments to test and validate models describing P dynamics at the scale of the agricultural fields under different agricultural management practices.