For 1000 years, human populations in dryland regions of the North American Southwest (NAS) extensively constructed diverse forms of agricultural infrastructure, including canals, linear rock alignments, check dams, stock ponds, and other earthworks and rock structures. The long-term hydrological impacts of these and the demographic and socio-political drivers of construction and maintenance have yet to be fully documented or vetted. This paper summarizes existing knowledge attained from the United Stated portion of the NAS, but a lot is still unknown about Northwest Mexico. There remain outstanding questions related to understanding how ancient agriculture might improve modern adaptability and resilience. The detailed ecological and topographical variability of this arid landscape illustrates the essential need for infrastructure in maintaining water and managing the impacts of climate change on the hydrological cycle. We describe pros and cons of different types of infrastructure and examine socio-environmental trade-offs between robustness and vulnerability produced by reliance on infrastructure, drawing from existing literature to examine timescales longer than a human lifespan. The development of historically-informed management approaches to increase dryland climate resilience benefits from incorporating constraints and opportunities mediated by past landscape modifications. We present a plan for leveraging existing knowledge, available science, and potential, to extend our knowledge base and further explore causal relationships.