Agrohydrological Models

Agrohydrological or water-oriented models were significantly developed during the last years (Bastiaansen et al., 2004). The models SWAP (Van Dam et al., 1997), DRAINMOD (Kandil et al., 1995), WAVE (Vanclooster et al., 1994), ISAREG (Teixeira and Pereira, 1992) and HYDRUS (Simunek et al., 1998) can be considered as agrohydrological models, among others.

The unsaturated zone, i.e. the zone between the soil surface and the groundwater, is a complicated system governed by highly non-linear processes and interactions. Flow processes can alternatively be described by means of physical-mathematical models. According to Bastiaansen et al. (2004), unsaturated-zone models can be used to simulate the timing of irrigations and irrigation depths, drain spacing and drain depth, and system behaviour and response. The models have increased our understanding of irrigation and drainage processes in the context of soil-plant-atmosphere systems. Progress in modelling can be attributed to merging separated theories of infiltration, plant growth, evapotranspiration and flow to drain pipes into a single numerical code (Bastiaansen et al., 2004).

According to Van Ittersum et al. (2003) agrohydrological models are more suitable to irrigation and water-use assessments than crop-growth oriented models, although both approaches have been used.

3.2.1 Constraints of Crop-Oriented Models Regarding the Simulation of Soil Water-Movement and Crop Water-Use

Actually, water moves not only down within the soil, but also lateral and even upward, depending on the potential gradients (Kutilek and Nielsen, 1994). Transport phenomena, as water movement into the soil, are driven by potential gradients that depend on gravity, water extraction by roots and water that enters or leaves the profile from top or bottom, causing different soil water suctions in the different layers.

The cascade approach could be appropriate on sandy soils or if the objective is to calculate the amount of water available to the crop over longer periods of time. However, this approach could fail in soils with significant clay and silt content and if the objective is to calculate daily soil-water profiles, as needed in irrigation assessments. A Richards-based approach might be more appropriate in those cases (Van Ittersum et al., 2003).

Several researchers have pointed out the DSSAT limitations regarding soil-water simulations, due to the use of the cascade approach in such simulations (Gabrielle et al., 1995; Maraux et al., 1998; Mastrorilli et al., 2003). According to Ritchie (1998), there is definitely a need to have better DSSAT simulations of the water balance in very poorly drained conditions where oxygen stresses will impact plant growth. Some attempts to improve DSSAT in that concern have been made already (Yang et al., 2004).

Particularly, since the cascade approach is unable to simulate upward water movement due to capillary rising, DSSAT yield predictions significantly depart from actual yields under heavy rain conditions (Rosenzweig et al., 2002). Particularly, Utset et al. (2006) showed that capillary rising can be an important component of maize water balance, when maize is cropped nearby river and channels, where shallower water tables can be found.

Extreme rainfall conditions will be more frequent in the near future, due to climate change (IPCC, 2007). As pointed out by Rosenzweig et al. (2002), yield loses in the US due to heavy rainfall could be very important in the future and the modelling-based climate-risk assessments should take them into account.

Utset et al. (2006) results also showed that water excess could significantly affect crop production. Figure 3 depicts the maize relative transpiration (ratio between actual and maximum transpiration) as simulated during dry years through SWAP by Utset et al. (2006). RT is depicted in Figure 3 as a function of Total Water Supply (TWS), considering a water table at a 2-m depth. The line shows the obtained regression. As can be seen in the figure, higher TWS gives raise to a reduction in RT rather than to further increments. Since the Feddes et al. (1978) root water-uptake function, included in SWAP, is able to account on water-excess effects on crop water use, the Utset et al. (2006) simulations can estimate the extreme rainfall effect also, whereas other models are unable to evaluate that consequence. Utset et al. (2006) simulation results indicate that precipitation excess could bring a negative effect in flooded irrigated maize, if relatively shallow water tables are found.

0 0

Post a comment