Irrigation Scheduling for Cotton During the Flowering–Boll-Setting Stage Under Contrasting Temperature Scenarios in Northern Xinjiang

This study quantified irrigation–yield trade-offs in Northern Xinjiang under contrasting temperatures and evaluated the interannual performance of candidate irrigation schedules. We coupled APSIM-Cotton, calibrated and validated with field observations and published data, with multi-objective optimization to optimize schedules from initial flowering to early boll opening, maximizing mean seed cotton yield while minimizing irrigation input. We classified years from 1994 to 2025 into low-, moderate-, and high-temperature scenarios based on heat degree days above 35 °C (HDD35). Representative schedules required 243.75, 292.50, and 363.75 mm, respectively; under the high-temperature scenario, greater irrigation did not produce a commensurate yield gain. We therefore optimized irrigation scheduling under the high-temperature scenario with total irrigation fixed at 312 mm. The highest simulated mean yield increased with irrigation frequency, remained relatively stable between 7 and 19 events, and declined thereafter. K-medoids clustering identified four near-optimal schedules with 9, 10, 14, and 21 events. In five high-temperature years not used for fixed-quota optimization, the 9-event schedule had the highest mean simulated seed cotton yield (6878.7 kg/ha) and relative yields of at least 99%, using each year’s highest yield among the four candidates as the reference. This schedule performed well in the evaluated years but should be validated under field conditions.