Machine learning (ML) applications in Earth and environmental sciences (EES) have gained incredible momentum in recent years. However, these ML applications have largely evolved in ‘isolation’ from the mechanistic, process-based modelling (PBM) paradigms, which have historically been the cornerstone of scientific discovery and policy support. In this perspective, we assert that the cultural barriers between the ML and PBM communities limit the potential of ML, and even its ‘hybridization’ with PBM, for EES applications. Fundamental, but often ignored, differences between ML and PBM are discussed as well as their strengths and weaknesses in light of three overarching modelling objectives in EES, (1) nowcasting and prediction, (2) scenario analysis, and (3) diagnostic learning. The paper ponders over a ‘coevolutionary’ approach to model building, shifting away from a borrowing to a co-creation culture, to develop a generation of models that leverage the unique strengths of ML such as scalability to big data and high-dimensional mapping, while remaining faithful to process-based knowledge base and principles of model explainability and interpretability, and therefore, falsifiability.

Socioeconomic and climatic changes and limited water resources pose various challenges to water, energy, and food sectors across the globe. The inevitable interactions between water, energy, and food systems bring about trade-offs but also synergies under different decisions and policies. To gain insights into these issues, we developed a water-energy-food (WEF) nexus model that incorporates both production (supply) and demands sides of WEF systems into a single system-of-systems model using the system dynamics (SD) approach. The model is applied to Saskatchewan, Canada, and so is named WEF-Sask. The model results reveal the various levels of sensitivities of water, energy, and food (and feed) sectors to the socioeconomic and climatic drivers. The analysis of trade-offs and synergies shows that the proposed large irrigation expansion (400%) boosts food production by 1.6% while reducing hydropower production by 2.7% in Saskatchewan. Wind expansion strategy (from 5% to 30% of total capacity) makes synergies that not only contribute to electricity supply but also reduce greenhouse gas emissions, industrial water demand, and groundwater use by 2.0, 5.7, and 3.8%, respectively. Biofuel use (blending mandate: 10% ethanol and 5% biodiesel) in transportation cuts GHG emissions by 1.2% but reduces the potential food export (food surplus) by 5.0%. The WEF-Sask model allows for scenario analysis toward integrated resources management, and its generic model structure can be expanded to other regions.

We explore how to address the challenges of adaptation of water resources systems under changing conditions by supporting flexible, resilient and low-regret solutions, coupled with on-going monitoring and evaluation. This will require improved understanding of the linkages between biophysical and social aspects in order to better anticipate the possible future co-evolution of water systems and society. We also present a call to enhance the dialogue and foster the actions of governments, the international scientific community, research funding agencies and additional stakeholders in order to develop effective solutions to support water resources systems adaptation. Finally, we call the scientific community to a renewed and unified effort to deliver an innovative message to stakeholders. Water science is essential to resolve the water crisis, but the effectiveness of solutions depends, inter alia, on the capability of scientists to deliver a new, coherent and technical vision for the future development of water systems.