Can we account for the “missing” phosphorus in simulated low phosphorus agricultural systems?

Long term total phosphorus (P) concentration, inorganic P and / or organic P concentration in agricultural soils is not commonly measured. As a consequence, computer-based models, that have been developed to predict P responses to changing management practices, are typically tested against soil “agronomically available” P data (as measured by tests such as; Olsen-P, Morgan’s-P, Mehlich-3, etc.) and those that do test against total P are limited to a few agricultural experimental sites across the world. While there is some correlation with total soil P, the term “available P” is arguably a functional concept, influenced by a large number of biotic and abiotic factors, rather than a direct soil measurement. This highlights a developmental gap in P modelling which could help to further unlock our understanding of P biogeochemical cycling when used in conjunction with contemporary empirical P research. 

Investigating P cycling in agricultural systems using the computer-based model N14CP has demonstrated that the model can predict carbon and nitrogen cycling and crop yields well for systems receiving abundant fertiliser. However, in systems where there is no P applied, predicted yield responses are greatly underestimated, with “missing” P input concentrations equivalent to annual fertiliser application rates. To date, the testing of N14CP has not included the P pools due to a lack of soil total P and/or soil organic P data from long-term field trials. Using recent total, organic and inorganic phosphate concentrations in the topsoil and yield data from two contrasting long-term field trial sites in the UK and the USA, this research will test P outputs and modelled yields from N14CP. It is hypothesised that the model will underestimate soil P concentrations, and crop yield, in the absence of P fertiliser inputs. This study will then apply changes to the mode model inputs, outputs and control processes to investigate whether these are sufficient to supply the crops and soil with the “missing” P. 

Understanding this source of “missing” P in N14CP will not only be useful for developing our understanding of P processes in computer-based models but could also further understanding of P processes linked to P draw-down in agricultural systems that have a history of high legacy P concentrations.