Can long-term fertilization accelerate pedogenesis? Depicting soil processes boosted by annual NPK-inputs since 1928 on bare loess Luvisol (INRAE-Versailles)

International audience

英语. Human activities worldwide menace beneficial soil ecosystem services, but long-term anthropogenic impacts on soil properties and processes are often difficult to assess in field conditions. Here we exploited INRAE's patrimonial '42-plots' bare-fallow experiment, an unique long term experiment in the world created in 1928 in Versailles (France), to emphasize long-term impacts of annual inputs of NPK fertilizers (ammonium, phosphate, potassium salts) and basic amendments (lime, basic slag) on loess Luvisols. We selected plots receiving monovalent (Na+, K+), acid (NH4+), basic (Ca2+) and non-amended (reference) plots, thus embracing the today widely diverging physicochemical surface soils conditions. Temporal changes of soil characteristics were studied on historical archived topsoil samples, whereas soil-depth impacts were studied on samples from subsurface horizons collected in 2015 until 120-cm depth. Bare-fallow management caused a rapid organic matter (OM) decay, soil acidification, CEC reduction and lixiviation of cations. With reduced OM-buffering capacities, specific fertilizer-induced physicochemical conditions enhanced the development of several soil-forming processes. NH(4)fertilizers amplified soil acidification (pH < 4), lixiviation (i.e. 1.5 kg m 2 of Ca), aluminization of the exchange complex, and weathering of ferromagnesian minerals and plagioclase feldspars. Under (NH4)(2)HPO4 fertilization, a quasi-total dissolution of chlorite and hornblende occurred, pointing to an acidocomplexolysis process in which the PO43- anion likely plays the role of complexing organic acids in Podzols. Ammonium fertilizers also affected E, B and C horizons. In reference plots, similar but lower effects remained restricted to the surface horizon. Na/Kfertilizers favoured substantial clay translocation (i.e. 10-15 kg m(-2)) from the Ap to underlying E and E/Bt horizons. Liming amendments counteracted acidifying effects of OM-depletion, and raised the pH to 8-8.5 and exchangeable Ca to >95%. It may be clear that the initial design specifications do not allow a direct comparison with current conditions of agricultural soils. However, in the view of global climate change, foreseeing a lowering of organic carbon contents in soils, the 42-plots trial acts like an "alert launcher", forecasting risks of soil degradation with respect to mineral soil phases, parameters and processes, generally buffered and masked by the presence of organic matter. The 42-plots experiment forms a high-valued playing field for experimental research, offering a unique centennial time-span of differing physicochemical properties in a soil context with close initial pedogenetic connexion.