How can long-term experimental plots can help us to understand the sustainability of different phosphorus inputs ?

During the last twenty years, we observed a constant reduction of mineral fertilizer use, due to prices increase and environmental awareness, and an increase of crop removal, leading to a phosphorus (P) budget decrease. These changes are feared for a decrease of soil P content, which is already observed in some regions in Wallonia. However, P being an essential element for plant growth, is a such management compatible with yield maintaining? Are the current cropping systems sustainable? To answer to the questions, different studies are made. However, long-term data are rarely available to understand the influence of cropping systems on the soil behavior, leaching risks or to choose adequate indicators of P. To answer to these questions in our soils, 2 experimental plots of the Walloon Agricultural Research Center. These experimental plots were established in 1967 and 1959 in order to evaluate the effect of, respectively, 3 P and K input levels and different organic inputs on the production. Soils samples were taken in plots and analyzed in laboratory. So, different P indicators and edaphic parameters were determined. This study showed that all indicators are coherent with P levels and correlated with yields but no many differences can be shown between fertilizer types. Meanly, zero P-input engenders a decrease of yield of 7%, while a double input increases yield of 2% in comparison to plots with an input corresponding to crop export. So, financially, the zero P-input option is rarely profitable in the long-term and double input of P removed is never financially sustainable. Leaching into deeper soils levels was studied with analysis of deeper horizons which indicated any leaching , even in plots with double inputs. Indeed, soil P contents in depth were similar in these plots than those with no P-inputs or soils under forest cover. So, to conclude, these plots help to study the sustainability of cropping systems in real situations and to determine appropriate management of P.