Agronomy, modelling and economics of reactive phosphate rocks as slow-release phosphate fertilizers for grasslands

Abstract

Reactive phosphate rocks (RPRs) from Sechura, Peru (SPR) and North Carolina, USA (NCPR) were compared with triple superphosphate (TSP) as phosphate (P) fertilizers for permanent grass/clover pastures in four field trials in New Zealand. Trial sites ranged in initial pH (in water) from 5.7 to 6.3 and in rainfall from 712 to 1338 mm yr−1. SPR and NCPR were used in the unground ‘as-received’ state. Fertilizers were applied annually for six years. Pasture was harvested by frequent mowing, and herbage dry matter (DM) yields were measured at each cut. Herbage P concentrations were measured at each cut in two trials and on most cuts in the other two. For all sites combined, DM production from RPRs was initially significantly less than from TSP but it improved relative to TSP with time. Substitution values of RPR relative to TSP, denoted by S.V. (TSP/RPR) and defined as the ratio of P in TSP to P in RPR required to produce the same plant response during a specified period of time, were estimated by relating yields from RPR treatments to the yield response curve for different application rates of TSP. For the four trials combined, S.V. (TSP/SPR) increased from 0.32 in year 2 to 0.85 in year 6. S.V. values for NCPR were similar. The site which had the lowest S.V. values (average 0.20) for total production over six years was the site with highest pH (6.3) and lowest rainfall (712 mm). Corresponding S.V. values for the other sites were 0.50 to 0.78. Herbage P concentrations showed a similar pattern of RPR performance relative to TSP to that shown by DM production except at the highest application rate where TSP always supported much higher herbage P concentrations than RPR. The pattern of DM production from RPR relative to TSP was explained on the basis of a model involving soil P pools of undissolved fertilizer P and plant-available P respectively, with the hypothesis that P dissolved from RPR entered the plant-available P pool and was used with the same efficiency as P entering by dissolution of TSP. Model predictions of substitution values using directly measured RPR dissolution rates agreed well with observed substitution values. The advantage of RPRs in comparison to soluble P fertilizers for permanent pastures was considered to lie in their lower price and not in greater nutrient efficiency. Economic advantage was calculated in terms of the return on investment from establishing and maintaining a pool of RPR in the soil large enough to release the required annual amount of plant-available P compared with the cost of annual applications of soluble P fertilizer.