The impact of the variability of natural precipitation and NPK fertilisation on maize yield was examined in the 21 year of the long-term experiment ’OMTK A-17’ (Nagyhörcsök, Experiment site of the Institute for Soil Sciences and Agricultural Chemistry, Centre for Agricultural Research, Hungarian Academy of Sciences). The main characteristics of the calcareous chernozem soil of the production site are as follows: pH (KCl): 7.3, CaCO3: 5%; humus content: 3%; clay content: 20–22%; AL-soluble P2O5 and K2O: 60–80 and 180–200; KCl-Mg: 150–180; KCl+EDTA-soluble Mn, Cu and Zn content: 80–150, 2–3 and 1–2 mg kg-1. The experiment had 20 treatments, 4 replications and a split-split-plot design on a total of 80 plots. The impact of nitrogen and phosphorus could be examined on 3–3 levels, while that of potassium on 2–2 levels in all of the possible 3×3×2=18 combinations. These combinations are supplemented by a nontreated control treatment and high NPK dose treatment which was not listed in the factorial system. The main findigs are described as follows:

Of the 21 experimental years of maize, the weather was normal in 9.5%, dry in 9.5%, drought in 42.9% and rainy in 38.1% of the time. Water deficient times (dry and drought) represented 52.4% of the experimental period.

In normal years, yields were characterised by balanced fertiliser impacts in comparison with control soils. N fertilisation, deficient NP and NK fertilisation resulted in extra yields of 2.7 t ha-1, 2.8 t ha-1 and 3.4 t ha-1, respectively. Yield could hardly be increased by total NPK (8.3 t ha-1) treatment.

In dry conditions, N, NP and NK fertilisation resulted in extra yields of 1.1 t ha-1, 2.8 t ha-1 and 3.1 t ha-1, respectively, which could not be significantly increased with the total NPK (3.2 t ha-1) treatments. The average of treatments in dry crop years resulted in 6.0 t ha-1, which is 18% less than in normal years.

During drought, similar impacts of N, NP, NK and NPK fertilisation was obtained as in drought. The average of treatments was 6.4 t ha-1, which is 12% less than in normal years.

In times of abundant precipitation, one-sided N and deficient NP and NK fertilisations resulted in a more positive effect of favourable water supply (N: 43%, NP: 49%, NK: 59%). NPK treatments could not further increase yield. The average of experiments in years with abundant precipitation was 7.5 t ha-1, which is 3% more than in normal years.

As regards the period that is determinant in terms of maize yield, the weather during summer periods was of chief importance (R2=0.5670***) as opposed to winter periods (R2=0.1232*). The R2 value of the whole year was 0.6822***. R2 was 0.6784*** in each month of the growing years. Also, the precipitation circumstances of the different vegetation periods were determinant (R2=0.5670***). The yield of maize during the vegetation period was mostly determined by precipitation in June (R2= 0.4730**). Weather during harvesting was also significant (R2=0.4797**). The frequency of precipitation anomalies determined yield (R2=0.5253***). The whole coefficient of determination of the conducted experiments (R2=0.6822***) showed that the yearly precipitation and fertilisation had a nearly 70% impact on maize yield.

Considering the above described changes in climate (increasing frequency of drought), the author of this paper came to a conclusion that precipitation optimums and the possibility of obtaining optimal maize yields will decrease in the future; therefore, the role of irrigation could be increasingly dominant. The results obtained with this model, which can be regarded novel even on the international scale, show the impact of normal, dry, drought and rainy years, as well as NPK fertilisation quantified in t ha-1. These results could be a favourable basis for the specific scientific description of the yield-affecting physical and chemical impacts of drought processes.

Keywords: climate change, precipitation variability, NPK fertilisation, maize, yield