Irrigation for crops in a sub-humid environment VII. Evaluation of irrigation strategies for cotton

Summary Empirical functions to predict the nitrogen uptake, increase in LAI and minimum leaf water potential (LWP) of cotton were incorporated into a water balance model for the Namoi Valley, N.S.W. A function was then developed to describe the lint yield of irrigated cotton as a function of water stress days at 4 stages of development, total nitrogen uptake and days of waterlogging. A water stress day was defined as predicted minimum leaf water potential less than -1.8 MPa up to 90 days after sowing and -2.4 MPa there-after; stress reduced yield by up to 40 kg lint ha^–1 d^–1 with greatest sensitivity at 81–140 days after sowing and when N uptake was highest. Nitrogen uptake was reduced by 0.98 kg per ha and yield reduced by 33.2 kg lint ha^–1 for each day of waterlogging. The model was used to evaluate various irrigation strategies by simulating production of cotton from historical rainfall data. With a water supply from off farm storage, net returns ($ M1^–1) were maximized by allocating 7 Ml ha^–1 of crop. The optimum practice was not to irrigate until 60 days from sowing and until the deficit in the root zone reached 50%. When the supply of water was less than 7 Ml ha^–1 there was no advantage in either delaying the start of irrigation or irrigating at a greater deficit; it was economically more rational to reduce the area shown or, if already sown, to irrigate part with 6 Ml ha^–1 and leave the rest as a raingrown crop. Irrigation decisions are compromises between reducing the risk of water stress and increasing the risk of waterlogging. The simulation showed that there is no single set of practices that is always best in every season; in a number of seasons practices other than those which on average are best, give better results.