Figure 5. Measured Sugarbeet root lengths under several water managements as function of the crop development stages (DVS). The continuous line shows the SWAP original function (after Utset et al. 2007b).
Likewise, Figure 6 depicts the Mediterranean-considered and the SWAP-original Kc and Ky coefficients as functions of the development stages. Unlike the original SWAP value, which sets the same Kc coefficient for all the sugarbeet crop development, we consider a variable Kc. Velicia (1998) showed that sugarbeet water requirements are lower than the maximum reference evapotranspiration during the initial growing period. However, sugarbeet water needs increase significantly after completing foliar development and starting the root-growing period (Velicia, 1998), which might be considered for a DVS of around 0.8. Crop water needs are reduced close to maturity. Furthermore, the relationships between sugarbeet transpiration and root yield should not be considered as linear, as originally in SWAP. A reduction of actual sugarbeet transpiration regarding its potential transpiration may imply no serious yield reduction during initial leaf development, but such a reduction has significant consequences on final yields at the beginning of the root growing period (Velicia, 1998). The Ky coefficient, as well as the relative importance of water shortage in Mediterranean sugarbeet yields, can be linearly reduced after the initial root-growing period (Velicia, 1998).
Figure 7 depicts the relationship between the measured yields and the simulated relative yields, as well as the corresponding regression line. The data in Figure 10 comprise all the pairs of simulated-actual yields, regardless of water managements or soil types. As shown in Figure 7, lower actual yields start notably at the 1:1 line, while yields of above 50 t/ha generally correlate well with the simulated relative yields. Indeed, SWAP only takes into account the yield reductions due to water shortage. However, final yields rely on many other issues besides water availability, such as weed infections, diseases or nutrient deficiency. Furthermore, relative yields are always estimated as one if irrigation water is sufficient, whereas actual yields are variable. That is why the dispersion shown in Figure 10 is higher around the highest simulated relative yields.
Was this article helpful?