Objectives Potassium (K) fertilizer is effective on increasing yields of both sweet potato and wheat in sweet potato/wheat rotation. The paper studied the suitable K management for the system.

Methods A local field experiment was conducted for three consecutive years in Liuhe Experimental Base of Jiangsu Academy of Agricultural Sciences, Nanjing City. At the same annual K fertilizer application rate of K₂O 270 kg/hm², five K allocation rates in sweet potato (S) and wheat (W) were setup as S0W270, S90W180, S135W135, S180W90 and S270W0. The yield and yield components of sweet potatoes and wheat were investigated, the uptakes and physiological efficiency of K fertilizer were calculated, and the apparent balance of soil K was estimated.

Results In sweet potato, K fertilizer allocation rates obviously affected the tuber yield, tuber number per plant, fresh weight per tuber, ratio of top to root, the K uptake and physiological efficiency. In wheat, the effects of K allocation rates on grain yield, ear number per hectare, grain number per ear, 1000-grain weight, ratio of straw to grain, the uptake and physiological efficiency, and K harvest index were not significant. The highest tuber yield and total rotation yield were obtained in treatment of S270W0, and followed by S0W270. Compared to the treatments of S90W180, S135W135 and S180W90, the tuber yield and total rotation yield under the S270W0 were increased by 20.7%‒24.5% (P < 0.05) and 17.8%‒20.9% (P < 0.01), respectively, and those under the S0W270 were increased by 9.9%‒13.4% (P > 0.05) and 8.2%‒11.0% (P > 0.05), respectively. Under the treatment of S270W0, the tuber number per plant was 5.2%‒10.4% and the fresh weight per tuber was 8.5%‒30.6% higher than the three treatments (P < 0.01), which were the two main factors leading to high yield. While under the S0W270, the tuber number per plant was 18.9%‒24.8% higher than those of the three treatments (P < 0.001), which was the main reason for tuber yield increase, although the single tuber fresh weight was the lowest. The whole plant K uptakes in treatments of S180W90 and S270W0 were 9.7%‒16.1% higher than those in treatments of S0W270 and S90W180 (P < 0.05). With the increase of K fertilizer rate on sweet potato, the top to root ratio was first increased and then decreased (P < 0.001), K physiological efficiency was in the reverse trend (P < 0.01), K harvest index and commodity rate were not changed (P > 0.05). The average K uptake was K₂O 228.0 kg/hm² and K harvest index of sweet potato was 0.55, which were 2.3 and 5.5 times of those of wheat. On average 46.6% of total plant K was stored in sweet potato vines, returning sweet potato vine to fields could meet the K demand of wheat. Averaged 91.0% of the whole plant K in wheat was stored in straw, returning wheat straw to fields could only meet 39.5% of the K demand of sweet potato. Without returning sweet potato vines and wheat straws to fields, K₂O 327.9 kg/hm² was taken out from field within one rotation cycle, the soil readily available K was in status of rapid depletion at the end of rotation. In this study, the soil readily available K content was decreased by 49.2% after three rotations under the K₂O application rate of 270 kg/hm². Under the condition of returning all sweet potato vine and wheat straw to fields, K balance could be maintained with annual K₂O application rate of 134.3 kg/hm².

Conclusions In sweet potato/wheat rotation, returning sweet potato vines to soil could meet the K requirement of wheat, applying all the potash on sweet potato could produce the highest rotation yields and fertilizer efficiency. In case of total vines and wheat straws were returned to soil, the potash rate could be minimized from 270 kg/hm² to 135 kg/hm² for keeping the soil K balance. For the experiment area and rotation system, the appropriate management of K fertilizer is “all straw and vine returning + proper K fertilizer rate + all K fertilizer applied on sweet potato”.