Abstract
#E-mail: asfrid@mail.ru
Crop yields of 4-pole crop rotation for plots of long-term (42 years) field experiment on sod-podzolic sandy loamy soil of Lyubertsy experimental field were analyzed to study the effect of various forms of potash fertilizers. In order to identify the features of long-term yield dynamics and reduce the role of changes from year to year associated with crop differences, it is proposed to smooth out a long-term series of yields using the moving average method with a time window equal to the duration of rotation. As a result, clear yield changes with a cycle of 7–12 years were revealed, probably related to the cyclical nature of weather conditions. The long-term yield dynamics for plots corresponding to the control variant were similar to each other; the same was revealed for plots of the NP-variant. On the contrary, long-term dynamics for plots of all NPK variants were grouped by field repetitions, not by variants. Statistical differences between plots in terms of their yield were estimated using the t-criterion and the Wilcoxon criterion for conjugate samples. According to the series of differences between annual yields and their trend values, the errors of the average long-term yield indicators for plots were estimated, which made it possible to compare them with each other. The same was used for a 10-year period of studying the aftereffect of potash fertilizers. Long-term yield trends in plots in all cases turned out to be insignificant, but for plots of control and NP-variants, there was a tendency to decrease yields over time.