Efficiency and risk of long-term fertilization of durum wheat

The objective of this study was to assess the effectiveness and risk of long-term (1990-2014) nitrogen-phosphorus fertilization on the productivity of durum wheat at a stationary fertilizer trial on soil type Pellic vertisols in the Institute of Field Crops - Chirpan, Bulgaria. The rates of N and P were 0, 40, 80,120 and 160 kg.ha(-1). The N was applied two times: one third at sowing, and the rest - as a top dressing at the tittering stage, and P - before sowing. The block design with four replicates was used. The agronomic efficiency (AE) was calculated using the formula: AE = (Y-Y-0)/F (kg.kg(-1)), where Y and Y-0 were grain yields with fertilization and unfertilized control, respectively, and F - amount of N fertilizer (kg.ha(-1)). The data and the regression dependencies between grain yield and N fertilization were analyzed by the SPSS program and the Duncan's multiple range test to find significant differences among means. The Pearson's correlation coefficient was determined. Comparison of the meteorological conditions in the studied years with the long-term 86-year period shows that there were significant deviations, especially in terms of humidity, and the deficit was primarily from vegetative rainfall. It was found that average for 25 years, the cultivar Progress realized a grain yield of 3.53 t.ha(-1). The yield was 2.34 t.ha(-1) without fertilization. At N fertilization the yield increased to N-120 - 4.02 t.ha(-1), exceeded the unfertilized by 17.2%. The yield reached up to 2.61 t.ha(-1) at P-120, with 11.8% above unfertilized. With the increase of P rate there was a tendency for yield decrease due to the established good supply of phosphates. The grain yield increased at N120-160P80-120-4.06-4.33 t.ha(-1). The highest value was at N120P120. The grain yield without fertilization was lowest in 1999 - 1.03 t.ha(-1), and highest - 3.96 t.ha(-1) in 2001. The AE increased to N-80 reaching 173 kg grain/kg N fertilizer. At low applied P-40 AE was the best - 35 kg grain/kg fertilizer. The regression model showed that without N can be expected grain yields of up to 1.50 t.ha(-1) in 4 of the years; 1.50-2.00 t.ha(-1) in 5 of the years; 2.00-2.50 t.ha(-1) in 6 of them; 2.50-3.00 t.ha(-1) in 8 of the experimental years and over 3.00 t.ha(-1) in 2 of the years.