Water-balance measured ETc (mm)
Figure 10B: Absolute differences vs. Water-balance measured ETc
Figure 10: Comparison between water-balance measured and SWAP-simulated ETc (A) as well as absolute differences between water-balance measured and SWAP-simulated as functions of waterbalance measured evapotranspirations (B) (after Utset et al., 2007b).
Despite these differences, the regression line between simulated and water-balance measured ETc has a gradient of 1.14. The regression is statistically significant at the 99% confidence level. Simulated ETc values can explain up to 56.8% of the variability of waterbalance ETc estimates. Furthermore, Table 6 shows the summary statistics of the waterbalance measured ETc, the SWAP simulated ETc and the paired differences. As shown in the Table, both ETc data can be considered as distributed normally. According to a t-test for difference between means, considering non-equal variances, there is no statistical difference between the water-balance measured and the SWAP simulated sugarbeet ETc at the 95% confidence level. The same result was achieved by comparing the probability distributions of the water-balance measured and the SWAP simulated ETc values through a Kolmogorov-Smirnov test. However, according to a variance-comparison F-test, the water-balance measured ETc is significantly more variable than the SWAP simulated ETc at the 95% confidence level.
Was this article helpful?