Integrated crop-livestock systems (ICLS) can be an alternative to increase the productivity of agroecosystems by enhancing nutrient cycling via grazing animals. Despite the holistic approach that bears the designing of ICLS, fertilization practices are proceeded in a conventional crop basis, disregarding nutrient fluxes at the appropriate spatial and temporal dynamics. We argue that fertilization practices in ICLS must follow the same integrated approach. To test this, we compared a conventional crop fertilization strategy versus a system fertilization approach applied to two production systems being a conventional cropping system and ICLS. The conventional cropping system consisted of a soybean crop succeeded by a non-grazed Italian ryegrass cover crop. The ICLS model consisted of a soybean-Italian ryegrass rotation grazed by sheep. In the conventional crop fertilization strategy phosphorus and potassium were applied at soybean sowing and nitrogen at the Italian ryegrass establishment. The system fertilization consisted of the application of all nutrients during the Italian ryegrass establishment. Accordingly, treatments were fertilization strategies in a factorial framework with production systems randomly distributed in a complete block design with four replicates. Results indicated for the first time greater daily herbage accumulation rate (24%; P < 0.01) and total herbage production (18%; P < 0.05) in the system fertilization compared with conventional crop fertilization. Consequently, system fertilization allowed for greater stocking rates in the pasture phase (17%; P < 0.05). The ICLS presented greater equivalent soybean yield (P < 0.001), energy production (P < 0.01), and system productivity (P < 0.05) compared with the cropping system, regardless of fertilization strategies. Soybean yield was not affected by fertilization strategies or grazing. In conclusion, the adoption of system fertilization strategy and crop-livestock integration enhance the production without jeopardizing soybean grain yields, so that land use is optimized by a greater energy production per unit of nutrient applied.

集成的牲畜－牲畜系统（ICLS）可以通过增强放牧动物的养分循环来提高农业生态系统的生产力。尽管采用整体方法进行ICLS的设计，但在常规作物的基础上进行了施肥实践，忽略了在适当的时空动态下的养分通量。我们认为ICLS中的施肥方法必须遵循相同的综合方法。为了测试这一点，我们将常规作物施肥策略与系统施肥方法进行了比较，该方法应用于两个生产系统，即常规作物种植系统和ICLS。传统的种植系统包括大豆作物，再加上未播种的意大利黑麦草覆盖作物。ICLS模型由绵羊放牧的大豆-意大利黑麦草轮换组成。在常规的作物施肥策略中，在意大利黑麦草生​​产基地，大豆播种时施用了磷和钾，而大豆则施用了氮。系统施肥包括在意大利黑麦草种植期间施用所有养分。因此，处理是在析因框架中的施肥策略，生产系统随机分布在一个完整的块设计中，具有四个重复。结果首次表明每日草本植物蓄积率更高（24％；处理方法是在析因框架中采用受精策略，生产系统随机分布在一个完整的区块设计中，一式四份。结果首次表明每日草本植物蓄积率更高（24％；处理方法是在析因框架中采用受精策略，生产系统随机分布在一个完整的区块设计中，一式四份。结果首次表明每日草本植物蓄积率更高（24％；与常规作物施肥相比，系统施肥的P <0.01）和总牧草产量（18％；P <0.05）。因此，系统施肥可以提高牧场阶段的放牧率（17％；P <0.05）。ICLS表现出更高的当量大豆产量（P <0.001），能源生产（P <0.01）和系统生产率（P<0.05），与施肥系统相比，无论施肥策略如何。大豆产量不受施肥策略或放牧的影响。总之，采用系统施肥策略和作物－畜牧业一体化提高了产量而又不损害大豆籽粒的产量，因此，通过增加每单位养分的能量产量，可以优化土地利用。