Long-term mineralization of organic phosphorus in agricultural soils : evaluation and control factors

Total phosphorus (P) in the plough layer of cropped soils is at 30-40% in the form of organic compounds (SOP). Little is known about SOP mineralization, even though this process may significantly contribute to the soil plant-available P and therefore to the P nutrition of cropped plants.The objective of this thesis is to quantify the long-term mineralization of SOP under field conditions and to identify its driving factors (SOP speciation, soil properties, climatic conditions).To answer these questions, four long-term field experiment (LTFE) were selected, two belonging to the INRAE LTFE network on mineral fertilization and the other two to the Observatory of Environmental Research on Organic Waste Products, i.e. OWP (SOERE-PRO). They were selected to cover a diversity of agropedoclimatic situations. The underlying hypothesis is that the nature (superphosphate, urban composts, urban sewage sludge, cattle manure) and the fertilization P dose (between 0 and 112 kg P ha-1 yr-1), crop residue management, soil classification, climatic conditions, and crop succession, affect the SOP mineralization rate. In addition, the time-series databases of crop and soil measurements from these trials are complete, and the soil were regularly sampled and stored. This allowed determining changes in soil SOP and inorganic P (SIP) levels over several decades, as well as the plot cumulative P budget (Bcum = ∑(P supplied − P exported in crops)). The SOP mineralization rate was calculated using the bi-compartmental model of Hénin and Dupuis (1945). This model describes the annual evolution of the SOP stock as a function of its mineralization and the incorporation of P into the SOP through the decomposition of crop residues and OWP. In two of the four LTFE, the speciation of SOP and SIP was determined by nuclear magnetic resonance of P in NaOH-EDTA solutions (P-NMR).The initial SOP stock varied considerably among the trials from 368 to 1145 kg ha-1. On average across trials and treatments, 3.2 kg P ha-1 year-1 of post-harvest aerial and root residues were incorporated into the SOP stock for 10.9 kg P ha-1 year-1 released into the soil solution. For OWP, P input varied considerably from 24 to 112 kg P ha-1 yr-1 (mean 32.4) depending on the reasoning (i.e. 2 t C-OWP ha-1 yr-1 or 170 kg N-OWP ha-1 every two years) and the C/P and N/P ratio. On average, 5.5 kg P ha-1 year-1 was incorporated into the SOP and 26.9 kg P ha-1 year-1 was released into the soil solution.The SOP mineralization rates ranged from 1.7 to 11.2 kg P ha-1 yr-1 (5.1 on average or 0.8% of the SOP stock) with a residence time ranging from 56 to 227 yrs. The P-NMR results indicate that the SOP is predominantly in the form of orthophosphate monoesters, in which inositolhexakisphosphates (IHP) are the main compounds. The low SOP mineralization values could be explained by the presence of IHP, which are known strongly interact with the soil solid phase, and by a high concentration of phosphate ions in soil solution, which could inhibit the enzymatic catalysis of IHP. Despite very different Bcum values (between -724 and +1830 kg P ha-1 for one of the trials), the SOP stock is constant in all LTFE while, on the contrary, the SIP stock changes strictly according to Bcum values.In conclusion, this thesis indicates that SOP mineralization is low, although variable among the four LTFE. It varies proportionally to the SOP stock but the drivers of this stock and of the mineralization coefficient could not be clearly identified, probably due to the scarcity of analyzed situations. This mineralization flux contributes marginally to the soil plant-available P, compared to the diffusion mechanism of phosphate ions at the solid-to-solution interface.