Diversified farming systems is a challenge in lowland soils around the world. In Southern Brazil, integrated crop-livestock systems (ICLS) have been used for grain crops and livestock production under no-till. However, their impacts on soil K dynamics are still poorly explored. The aim of this study was to quantify the K forms in the soil and the K budget after 66 months of different flooded rice production systems under no-till ICLS in a subtropical paddy field. Four production systems were established in 2013: (i) flooded rice monocropping under conventional tillage (CT) and winter fallow (R-CT); (ii) no-tillage (NT) flooded rice monocropping with ryegrass pasture in winter (R-ICLS); (iii) NT flooded rice and soybean rotation in the summer and ryegrass pasture in winter (RS-ICLS), and (iv) NT flooded rice, soybean, maize and stapf grass rotation in the summer period and ryegrass and clover pasture in the winter (RSM-ICLS). In 2018, soil samples were taken at the 0–10, 10–20 and 20–30 cm soil layers. Exchangeable potassium (K), non-exchangeable K, structural K, and total K were evaluated. Exchangeable K was not affected by the treatments. The higher the difference between inputs and outputs of K in the systems evaluated, the higher was the content of structural and total K in the soil. However, the K budget in the system was not sufficient to fully explain the variation in the non-exchangeable K content in the soil. Comparing the systems with the same frequency of flooded rice cultivation, the difference between inputs and outputs of K was about 2.3 times higher in the R-ICLS system (830 kg ha⁻¹ of K) compared to R-CT (366 kg ha⁻¹), which increase in 70 % the concentration of non-exchangeable K in the 0–30 cm soil layer (622 and 361 mg kg⁻¹ in the R-ICLS and R-CT, respectively). In the systems where flooded rice rotates with rainfed crops (RS-ICLS and RSM-ICLS), the drying cycles resulted in K fixation with higher energy, turning the surplus of K added via fertilization into structural forms. These results indicate that lowland soils cultivated with K fertilization rates greater than the export of K can result in a fixation of K, decreasing its availability in the short-medium term. Moreover, it strongly emphasizes that adequate K fertilization in lowland soils with 2:1 clayminerals should focus on the reposition of exported K in order to prevent temporary fixation of K by fixation on structural forms.

多样化的农业系统是世界各地低地土壤面临的挑战。在巴西南部，综合作物畜牧系统 (ICLS) 已用于免耕粮食作物和畜牧生产。然而，它们对土壤 K 动态的影响仍然缺乏探索。本研究的目的是量化亚热带稻田免耕 ICLS 下不同水淹水稻生产系统 66 个月后土壤中的钾形态和钾平衡。2013 年建立了四个生产系统：(i) 常规耕作 (CT) 和冬季休耕 (R-CT) 下的水稻单作；(ii) 冬季免耕 (NT) 用黑麦草牧场单作淹水水稻 (R-ICLS)；(iii) NT 淹没水稻和大豆在夏季和冬季黑麦草牧场 (RS-ICLS)，以及 (iv) NT 淹没水稻、大豆、夏季种植玉米和 stapf 草，冬季种植黑麦草和三叶草（RSM-ICLS）。2018 年在 0-10、10-20 和 20-30 cm 土层采集土壤样品。评估了可交换钾 (K)、不可交换 K、结构 K 和总 K。可交换钾不受处理的影响。评估系统中钾的输入和输出之间的差异越大，土壤中结构和总钾的含量就越高。然而，系统中的 K 收支不足以完全解释土壤中不可交换 K 含量的变化。与相同频率的水淹水稻种植系统相比，R-ICLS 系统（830 kg ha 土壤样品取自 0-10、10-20 和 20-30 cm 土层。评估了可交换钾 (K)、不可交换 K、结构 K 和总 K。可交换钾不受处理的影响。评估系统中钾的输入和输出之间的差异越大，土壤中结构和总钾的含量就越高。然而，系统中的 K 收支不足以完全解释土壤中不可交换 K 含量的变化。与相同频率的水淹水稻种植系统相比，R-ICLS 系统（830 kg ha 土壤样品取自 0-10、10-20 和 20-30 cm 土层。评估了可交换钾 (K)、不可交换 K、结构 K 和总 K。可交换钾不受处理的影响。评估系统中钾的输入和输出之间的差异越大，土壤中结构和总钾的含量就越高。然而，系统中的 K 收支不足以完全解释土壤中不可交换 K 含量的变化。与相同频率的水淹水稻种植系统相比，R-ICLS 系统（830 kg ha 评估系统中钾的输入和输出之间的差异越大，土壤中结构和总钾的含量就越高。然而，系统中的 K 收支不足以完全解释土壤中不可交换 K 含量的变化。与相同频率的水淹水稻种植系统相比，R-ICLS 系统（830 kg ha 评估系统中钾的输入和输出之间的差异越大，土壤中结构和总钾的含量就越高。然而，系统中的 K 收支不足以完全解释土壤中不可交换 K 含量的变化。与相同频率的水淹水稻种植系统相比，R-ICLS 系统（830 kg ha^-1 of K) 与 R-CT (366 kg ha ^-1 ) 相比，0-30 cm 土壤层中不可交换 K 的浓度增加了 70%（622 和 361 mg kg ^-1分别在 R-ICLS 和 R-CT 中）。在淹水水稻与雨养作物轮作的系统中（RS-ICLS 和 RSM-ICLS），干燥循环导致以更高的能量固定钾，将通过施肥添加的过剩钾转化为结构形式。这些结果表明，低地土壤中钾肥的施肥率大于钾的输出量，会导致钾的固定，在中短期内降低其可用性。此外，它强烈强调在含有 2:1 粘土矿物的低地土壤中适当施钾应侧重于输出钾的重新定位，以防止通过固定结构形式而临时固定钾。