A series of farming experiments was conducted between 2013 and 2017 in Range Creek Canyon, Utah, to better understand the opportunities and constraints faced by prehistoric farmers in the Southwest. The experiments were designed to collect data on the optimal amount of supplemental water that should be applied to maize fields given the costs in labor and benefits in greater yield. We investigate expected variation in water management strategies using an optimal irrigation model (OIM). The model makes clear that the payoff for farming is best understood as a continuum of relative success and that irrigation is one activity (probably of many) that may improve farming efficiency as well as increase harvest yields. The optimal harvest will always be less than the maximum harvest when there are significant operating costs associated with irrigation. Estimating the costs and benefits of irrigation in a specific area allows for an assessment of whether irrigation is expected, and if so, how much effort should be devoted to water management. A local dendroclimatological study is used to provide the prehistoric context for the Fremont who occupied Range Creek Canyon, and irrigation is expected even in periods of greater precipitation.