The pyfao56 automatic irrigation scheduling algorithm

The pyfao56 Python package provides an open-source, evapotranspiration-based water balance model that is useful for irrigation management and water resource planning. Originally, pyfao56 required users to input explicit information on the timing and depth of irrigation; however, more recent applications called for an automated method to determine irrigation schedules in the model. The objectives of this research were to develop a new pyfao56 methodology for computation of irrigation schedules based on 25 user-specified parameters and to demonstrate the approach using data from a 2018 cotton field study at Maricopa, Arizona. The pyfao56 autoirrigation methodology involves a new software object (“AutoIrrigate”) for specifying associated parameters and an algorithm that determines whether or not irrigation is indicated on each daily timestep. Parameterization of the pyfao56 model for the conditions of the 2018 cotton field study demonstrated that the model could respond to irrigation management variability, with root mean square errors of 8–15 mm between measured and modeled root zone soil water depletion. A variety of autoirrigation scenarios were devised to evaluate pyfao56 outcomes for the 2018 Arizona cotton season with irrigation scheduling based on different management criteria related to soil water status, plant water stress, evapotranspiration replacement, practical constraints, and precipitation forecasts. Additional scenarios tested the ability of the autoirrigation algorithm to replicate the actual 2018 field schedule and to mimic schedules for production-scale Arizona cotton production with furrow flood, overhead sprinkler, and subsurface drip irrigation systems. The resulting irrigation schedules were widely variable with number of irrigation events ranging 8–103, total seasonal irrigation ranging 642–1110 mm, and maximum daily irrigation ranging 10–108 mm. The pyfao56 autoirrigation methodology provides a flexible tool for simulating realistic irrigation management schedules that have practical relevance for real field applications with different irrigation methods, system constraints, and management philosophies. • A method to automatically compute irrigation schedules in pyfao56 was developed. • Up to 25 parameters are available to control irrigation scheduling computations. • The method was demonstrated to schedule irrigation for Arizona cotton production. • Alternative parameterizations led to broad variability in autoirrigation outcomes. • The autoirrigation methodology improved realism of simulated irrigation schedules.