Perhaps the most important application of Earth System Modeling in the coming decades will be to inform climate impact mitigation efforts. Most climate impacts will be experienced on land, and these impacts will be responded to by humans living in specific locations with locally defined natural and sociopolitical constraints. Whether these decisions are farmers altering their planting and fertilization schedules under increasing flash-drought conditions, or Arctic communities changing their supply strategies based on rapidly changing river and permafrost conditions, climate impact mitigation decisions are frequently made at region to local scales. Decision makers will need to understand and reason about, holistically, the various tradeoffs of specific actions, and they will do so through scenario discovery and models exploring those scenarios.

Providing resources for decision-makers through modeling requires regional to local modeling capabilities, driven by global climate models. The required models cannot be one-size-fits-all; models must be purpose-built with application appropriate detail to capture the right processes at the right scales for a given question. The challenge then is to develop natural and human system modeling frameworks that allow the rapid development of purpose-built models at regional-to-local scales, driven by the global climate context. Amanzi–ATS is a watershed-to-river basin scale hydrologic modeling framework built on a foundation of process-based physics kernels to ensure generality and written to meet this need. It was designed from the ground up to be flexibly configurable, building on a graph-based coupling strategy that combines those fine-grained process models at runtime, and solves them across multiple views of the domain. Currently being ported to Kokkos, Amanzi–ATS leverages leadership class machines, and is capable of scaling to large river basins. ATS focuses on hydrologic flow and transport and, in various combinations for various applications, is coupled with ELM for vegetation and surface processes, PFLOTRAN for belowground biogeochemistry, and MOSART for river routing.

In this work, I describe three case studies in major river basins across the nation: the Saganavirktok River on Alaska’s North Slope, the Portage River in the Great Lakes Region, and the Neches River on the Gulf Coast. Each model, built out of common components, is being used to address climate impact mitigation efforts in EESSD research projects.