Simulated and observed sugar beet and spring wheat yields and yield variability in a long-term field experiment

Potential and water-limited yields of sugar beet and spring wheat in the pe riod 1964-1993 were simulated with the WOFOST 6.0 crop growth model. Three different available water contents in the upper 100 cm of the soil at emerg ence, 20, 15 and 10 cm (20 WAC, 15 WAC and 10 WAC, resp.) were simulated. O utputs of the model were compared with observed yields from a long-term fie ld trial at Prague-Ruzyne, Czech Republic. The experiment was conducted in one field so each of the experimental crops was grown every other year. The re were four fertilizer treatments, nil, organic, inorganic and a combinati on of the last two. Our study aimed to estimate the difference between simu lated yields only limited by climatic factors and observed yields in the in tensive two-crop rotation and to estimate yield variability. For the 15 yea rs when beet was grown, the mean potential, water-limited (at 15 WAC) and a ctual yields (averaged over the four treatments) were 15.2, 12.3 and 10.9 t .ha(-1), resp. The equivalent figures for wheat were 8.3, 7.9 and 4.0 t.ha( -1). The inter-season variability of simulated water-limited beet yields re ached 20-38% (the increase being due to decrease of WAC) and 22-34% in obse rved yields with higher values in nil treatment. The observed variability i n wheat ranged from 16- 21% in fertilized and nil treatments, while the var iability of simulated yields ranged between 4-17%. The correlation between observed (fertilized treatments) and simulated water limited yields was bet ter in sugar beet (r = 0.60-0.75) than in wheat (r = 0.27-0.44). There was no correlation between potential and observed yields in both crops. The ana lysis of residuals showed possible ways for improvement of simulation resul ts.