Abstract:Soil potassium fertility could probably be changed by long-term fertilization management. Effect of 16-year different fertilization treatments on the change of potassium fertility in soil of a long-term field experiment located in Fengqiu County, Henan Province,China was studied in this paper. Five soil K testing methods (1 mol/L NH4OAC method, 1 mol/L boiling HNO3 method, total potassium method and two kinds of sodium tetraphenylboron (NaTPB) methods) were compared with regards to their ability to recover K change in soils. The results of soil K balance based on 16-year K input of fertilization and K removal by crop harvest showed that soil K depletion occurred in the soils without input of fertilizer K. K depletion rate of NP (apply nitrogen and phosphorus fertilizers) treatment was 5.2 time as that of the CK (without fertilizer) treatment which was K 28.5 kg/(hm2·a). For all the other 5 treatments with fertilizer K applied, soil K got positive balance because that the amount of K applied was higher than the amount K removed by crops. The higher soil K accumulation rates were found in the NK (apply nitrogen and potassium fertilizers) and PK (apply phosphorus and potassium fertilizers) treatment, that was K 207 and 190.5 kg/(hm2·a), respectively. While the K accumulation rates in soils treated with balanced fertilizers were comparatively lower, which ranged from K 31.5 to 63.0 kg/(hm2·a). The change of soil K measured by 5 soil testing methods differred from the estimated value calculated by soil K balance and the mass of the topsoil in the field. The average recovery rate (ratio of value determined to the value estimated) of change of soil K of 7 fertilizers treatment determined by 5 methods (1mol/L NH4OAC method, 1 mol/L boiling HNO3 method, two kinds of NaTPB methods and total potassium method ) were 39%, 51%, 66%, 71% and 176%, respectively. The variation coefficient of the recovery rate of the above 5 methods were 103%, 74%, 36%, 30% and 141%, respectively. The results indicated that the conventional 1 mol/L NH4OAC method, 1 mol/L boiling HNO3 method and total potassium method were not suitable for determination of K change in soils, while the NaTPB methods proposed were suitable for this purpose.