Abstract: 【Objective】The potassium (K) natural productivity from soil, K fertilizer use efficiency and apparent soil K balance were investigated to provide a theoretical basis for the sustainable management of K element in farmland. 【Method】The three fertilizer treatments used were control, NP and NPK, representing various combinations of N, P and K fertilizer applications. Long-term (over 20 years) experiments with barley-rice crop rotation were conducted to investigate the response of crop aboveground biomass, yield, the K content in grain and straw to different application treatments. 【Result】The results indicated that the grain yields on average were 8.24 and 9.22 t?hm-2 per year in barley-single crop rice and barley-early rice-late rice rotation cropping system without application fertilization for 20 years, which accounted for 75.6% and 71.9% of that in NPK treatments in the corresponding rotation system. The contribution of application fertilizers increased crop grain yields by 24.4% to 28.1%, and the contribution of K fertilizers increased crop grain yields by 11.8% to 14.0% in barley-single crop rice and barley-early rice-late rice rotation cropping system, respectively. The contribution of K fertilizers increased grain yields for barley (17.9%) was 36.1% higher than that for rice. The annual K uptakes from CK treatments accounted for 59.1% and 58.3% of that in NPK treatments. The K in straws accounted for 81.7% and 74.2% to 87.6% of total K uptake for barley and rice, respectively. The indigenous K supplying capacity in barley season was lower than that in rice season. The annual indigenous K supplying capacity for barley on average was 69.8%, which were 15.5%, 19.7% and 19.4% lower than that for early rice, late rice and single crop rice, respectively. The indigenous K supplying capacity in barley season decreased significantly over cultivation time at an average rate of 1.6% per year. But there was no significant change of the indigenous K supplying capacity in time during rice cropping. The indigenous K supplying capacity in rice season still was more than 80% after 20 years cropping. The internal K use efficiency was decreased because of K application. But there was no significant change of internal K use efficiency among treatments in time. Under balanced fertilization with fixed N, P and K inputs, the internal use efficiencies for barley, early rice, late rice and single-crop rice were 42.4, 44.7, 41.3 and 44.8 kg•kg-1, respectively. The partial factor productivity in NPK treatment was 83.5, 79.7, 71.6 and 69.7 kg•kg-1, respectively. The K use efficiency in barley season was higher than that in rice season. The apparent K recovery efficiency was 29.5% and 41.5% in barley-single crop rice and barley-early rice-late rice rotation system, respectively. Crop continuous uptake caused a net K lost of 148.2 and 182.7 kg•hm-2 per year in the soils without application of fertilizer in barley-single crop rice and barley-early rice-late rice rotation cropping system, respectively. Moreover, the application of nitrogen and phosphate fertilizer accelerated soil K depletion, the extra K amount of 57.8 and 65.0 kg•hm-2 was taken up by crop. Although K deficiency in soil could be relieved by K application, a net K lost of 146.5 and 85.5 kg•hm-2 per year in soils received K fertilizer from barley-single crop rice and barley-early rice-late rice rotation cropping system. 【Conclusion】There was a very strong capacity to provide K element from soils during rice growing season. But the K fertilizer use efficiency in barley season was higher than that in rice season. There was a still net K loss from soil with K application rate of 187.5 kg K2O•hm 2 per year. So a great attention should be paid to the apparent soil K balance.

Key words: long-term , rice , barley , potassium , efficiency , balance