In regions where rainfall may be significant, the effective rainfall must be considered to reduce irrigation requirements and minimize the negative effects of overirrigation. These impacts include deep percolation, transport of nutrients and solutes into groundwater, surface runoff, and soil erosion. The scheduling of irrigation therefore requires data on potential future rainfall. This information can be in the form of a short-term weather forecast, or using information relative to previous rainfall records covering a significant number of years or using random generated rainfall events .
Short-term forecasts have been used successfully in wet regions, where there is a high rainfall probability, its quantity being the only unknown factor. In these cases, the net irrigation quantity is reduced to allow for rainwater storage in the soil without generating percolation. This type of strategy is highly adaptable to crops with a deep-root system, on soils with moderate to high water-retention capacity, and for systems applying light and frequent irrigations. In semiarid regions, the rainfall during irrigation periods is usually from local storms, and both their frequency and quantity are highly variable. In such cases, irrigators must rely on actual rainfall data and delay irrigation dates by a number of days as a function of the quantity fallen and the actual evapotranspiration rate.
There are several irrigation scheduling models that program irrigation by using short-term rainfall forecasts . The EPICPHASE model  is an example of a mechanistic crop growth model, capable of simulating different irrigation strategies using the weather forecast. The application of models having such capabilities may help to achieve water savings, as well as reduce the risk of percolation and the resulting leaching of nutrients, while not affecting yields.
Was this article helpful?