Long-term simulations of site-specific irrigation management for Arizona cotton production

Abstract

Engineering technologies for site-specific irrigation management (SSIM) have already been developed for applications in precision irrigation. However, further studies are needed to identify scenarios where SSIM leads to better agronomic outcomes than conventional uniform irrigation management (CUIM). The objective was to conduct a long-term simulation study to compare SSIM and CUIM given spatial soil variability at the Maricopa Agricultural Center (MAC) in Arizona. More than 500 surface soil samples were collected across a 730-ha area of the MAC from 1984 to 1987. A more detailed soil data set was more recently obtained across a 5.9-ha area at a MAC location designated for SSIM studies. Ordinary kriging was used for spatial interpolation of soil hydraulic properties within $$10\,\hbox {m} \times 10\,\hbox {m}$$ zones across the MAC, and 11 field parcels with an area of approximately 60 ha were delineated on the MAC quarter sections. Using an agroecosystem model, simulations of cotton production at the zone level with a 30-year weather record were conducted using a field-tested algorithm to optimize irrigation schedules for SSIM and CUIM. Long-term seed cotton yield, irrigation requirements, water use efficiency, and marginal net return for SSIM and CUIM strategies were often not different ($$p>0.05$$). Differences in seed cotton yield and irrigation requirements among the tested irrigation strategies were less than 11% and 6%, respectively, and within the typical range of model error. Most soils on the MAC have enough available water holding capacity to sustain cotton production at full potential with weekly CUIM, and advantages of SSIM were not consistently demonstrated by the simulations.