Figure 7: Experimentally obtained sugarbeet yields vs. SWAP simulated relative yields. The dashed and the solid lines show the regressions considering all the data and fitting-reduced data, respectively (after Utset et al., 2007b).
The correlation coefficient between simulated relative yields and the actual sugarbeet yields was 0.71. According to the regression obtained, the average maximum yield to be used in equation  should be 87.5 t/ha. The regression is statistically significant at the 99% probability level and explains more than 50% of the yield variability. However, the simulated relative yields are not normally distributed, since they are skewed to their highest value. Therefore, statistical comparisons between simulated relative yields and actual yields are limited. Despite this, the relatively high correlation coefficient and the level of statistical regression confidence would allow this simple SWAP approach to simulate water management effects on sugarbeet yields.
The average sugarbeet yield corresponding to simulated relative yields higher than 0.95 was 90.8 t/ha. Taking into account both results, a potential maximum yield for sugarbeet of around 89 t/ha can be used in SWAP simulations to estimate the effects of water availability on the final yields during crop seasons.
The soil water contents at the average depth of 0-60 cm in the Cambisol and Fluvisol plots, measured during the sugarbeet irrigation season, are shown in Figure 8. The water contents are generally close to the field capacity, shown in the figure by a dashed line. Water contents much higher than the field capacity would indicate an excess of water in the root zone and eventual water percolation. Furthermore, water contents much lower than the field capacity would indicate potential water excess and inefficient irrigation management. The results shown in Figure 8 indicate that, despite their non-technical approach, farmers usually know how to manage irrigation in an acceptable way (Utset et al., 2006).
Was this article helpful?