This study presents a procedure for simultaneously addressing stochastic input demands and resource supplies for irrigated agriculture within a linear modeling framework. Specifically, the effect of stochastic crop net irrigation requirements and streamflow supplies on irrigation water management is examined. Irrigators pay a self-protection cost, in terms of water management decisions, to increase the probability that stochastic crop water demand is satisfied and anticipated water supply is available. Self-protection cost is lower when increasing the probability that anticipated water supplies are delivered, ceteris paribus, than when increasing the probability that the crop receives full net irrigation requirement in the study region.