The motivation for this study was to create lunar surface exploration scenarios that would support current science needs, as captured in the Lunar Exploration Analysis Group (LEAG) Roadmap for Lunar Exploration. A science-driven capability to meet those needs required enhanced capability, relative to the Apollo J missions, to provide a broader field context for (1) improved interpretation of samples and measurements; (2) greater flexibility in the selection and nature of activities at field stations; as well as (3) greater potential for breakthrough science. Here, we offer advanced regional-scale (hundreds of kilometers) surface exploration scenarios, essentially design reference missions, for three high-priority targets representing the broadest differences in the nature and distribution of geological features. South Pole–Aitken Basin is the largest and oldest confirmed lunar impact basin. Covering most of the farside southern hemisphere and >2000 km in diameter, it contains extraordinarily diverse features and geochemical anomalies that are widely scattered and thus would require several regional-scale missions. Tsiolkovsky is an anomaly among farside craters: It is mare-filled in the thickest portion of farside crust, young, and has well-preserved impact structures, yet it is surrounded by the ancient Tsiolkovsky-Stark Basin. Aristarchus Plateau is a tectonically uplifted plateau associated with the formation of Imbrium Basin, and it is found on a concentric ring of basin. Features encompassing the entire range of mare volcanism activity in style and age are found either on the relatively compact plateau or within hundreds of kilometers in surrounding western Oceanus Procellarum. Our regional-scale architecture would allow science objectives for study of Aristarchus Plateau or Tsiolkovsky to be addressed from one landing site conveniently located on the target, while South Pole–Aitken Basin would require several missions to achieve such objectives. We describe the geological context and resulting investigations, as well as the required tools, instruments, and activities. We assumed, as initially instructed, science need–driven capabilities at least a generation beyond the Apollo J missions, i.e., the availability of a minimum of two pressurized rovers capable of hundreds of kilometers driving range at average speeds of 10–15 kph (without recharge), four crew, and 700 kg of science payload. The implications of such science-conducive architecture in the context of other architectures under consideration are discussed.