The potential of a coordinated system of gates for flood irrigation management in paddy rice farm

Rice is one of the most important staple foods in the world. In Europe, Italy is the main producer of rice, with almost all production concentrated in the northeast of the country. Traditionally, rice is grown in fields that are flooded from before planting until just before harvest. This water management technique requires a great deal of labour for farmers who have to manually adjust the inlet and outlet gates to maintain a constant ponding water level in the fields, especially when there is fluctuation of water supply at the farm inlet, for example as a result of rainfall. In addition, the practice of flood irrigation is very water-intensive. New technologies based on remotely and automatically controlled gates are being studied to increase the efficiency of this irrigation method. The objective of this work is to explore the potential of a coordinated and intelligent system of gates for efficient farm irrigation management and ponding water level maintenance. Based on information and measurements from a real case study consisting of a 40-hectare paddy rice farm located in northern Italy, where automatic gates and water level sensors were placed at strategic points of the farm canals and fields, respectively, a proportional-integral (PI) and a non-linear model predictive control (NMPC) of water levels were implemented and compared through modelling and simulation experiments. The results show that the proportional-integral control reproduces the actions that the farmer uses when faced with situations of surplus of water in the fields or a shortage of water in the farm canal. In particular, the general coordination of the gates is lost, and the individual binomial field-gate prevails as an independent system in the farmer's operation. Conversely, non-linear predictive control coordinates the gate operation to obtain a uniform ponding water level in the fields when there is a shortage of water, or significant water conservation when there is an excess of water as a result of rainfall. In conclusion, a nonlinear predictive control model seems to be a suitable strategy to advance irrigation management in rice farms, allowing rice farmers to continue the tradition of flooding while increasing its performance.