Diagnostic du statut de nutrition phosphatée et potassique des prairies naturelles en présence de légumineuses

Diagnosis of the phosphate and the potash status of natural grasslands containing legumes

The nutrient index method based on nutrient concentrations in plant tissues relative to the degree of growth is of general use for estimating the nutrient status of pastures ; it relies on critical curves which serve as diagnostic tools ; for N, the critical curve gives the optimum N concentration for different levels of biomass accumulation in the swards ; for P and K, optimum concentrations are linear functions of the sward N concentration. However limitations in the use of the P nutrient index (Pi) were reported when the herbage contained a large proportion of white clover. Our aim was to confirm whether similar behaviour existed with other legumes, and for the K nutrition indices (Ki) as well. This work relies on an experimental field where different N and P supplies lead to a large differentiation in white clover contents, and on a network of 20 pastures in a Pyrenean valley. Results showed that for P, as well as for K, the average nutrition indices calculated on the non-legume fraction were always higher than those obtained for the total sward. For a given sward, the difference between the indices calculated on the non-legume fraction and those calculated on the whole sward increased with the legume content of the sward ; the relationship was similar to that obtained for clover in previous work. Pi and Ki calculated from mixed sward concentrations resulted in an underestimation of the sward nutrition status. Our results showed that there were limitations to the use of diagnostic systems of nutrient status based on relative concentrations when the sward contained a large proportion of legume of whatever genus. We concluded that mineral contents measured on mixed swards could lead to a substantial underestimation of the NPK status of the sward. When nutritient indices are used for fertilizer recommendations, in order to avoid excessive P and K applications, we suggest to base the diagnosis on the non-legume fraction when biomass sorting is possible ; if not, we propose a correction coefficient of the nutrition index that varies according to the legume content as estimated de visu.