Delineation of management zones and optimization of irrigation scheduling to improve irrigation water productivity and revenue in a farmland of Northwest China

Abstract

Precision agriculture has been increasingly practised in recent years. Under precision management, farmland is divided into several management zones to implement different strategies and improve irrigation water productivity and revenues. In this research, six soil properties (silt, sand, soil moisture content, available nitrogen, electrical conductivity and elevation) were selected, and fuzzy c-means clustering was used to delineate management zones. The field was divided into three zones. The differences in the mean of the soil properties among the zones were large and within a zone were small. The coefficient of variation of the properties and yield were also smaller than that before classification. The optimization was carried out by using a genetic algorithm based on the Jensen model. Three objective functions were set as maximum yield, maximum irrigation water productivity (IWP) and maximum revenue, and the weights were kept equal to 1/3. The WHCNS (soil water heat carbon nitrogen simulator) model was used to simulate the maize yield under optimized irrigation schedule for the three management zones and to calculate IWP and revenue. Compared with uniform management, IWP and revenue were increased by 0.6 kg m−3 and $61 ha−1, respectively. The optimized irrigation schedule can be used as a reference for the actual irrigation management. It can increase the IWP and revenue under the premise of achieving the target yield. The results show that the method can guide precision agricultural production and management in large-scale farmland.