This review focuses on the use of Interpretable Artificial Intelligence (IAI) and eXplainable Artificial Intelligence (XAI) models for data imputations and numerical or categorical hydroclimatic predictions from nonlinearly combined multidimensional predictors. The AI models considered in this paper involve Extreme Gradient Boosting, Light Gradient Boosting, Categorical Boosting, Extremely Randomized Trees, and Random Forest. These AI models can transform into XAI models when they are coupled with the explanatory methods such as the Shapley additive explanations and local interpretable model-agnostic explanations. The review highlights that the IAI models are capable of unveiling the rationale behind the predictions while XAI models are capable of discovering new knowledge and justifying AI-based results, which are critical for enhanced accountability of AI-driven predictions. The review also elaborates the importance of domain knowledge and interventional IAI modeling, potential advantages and disadvantages of hybrid IAI and non-IAI predictive modeling, unequivocal importance of balanced data in categorical decisions, and the choice and performance of IAI versus physics-based modeling. The review concludes with a proposed XAI framework to enhance the interpretability and explainability of AI models for hydroclimatic applications.

Restoring the hydrologic flow regime of urban areas by promoting infiltration, retention, and evapotranspiration on the site is one of the goals of low-impact development (LID). These goals can be achieved through the implementation of stormwater control measures (SCMs) such as green roofs and permeable pavements. The effectiveness of SCMs can be influenced not only by their design but also by their location. The present study applies multi-objective evolutionary optimization and Monte Carlo simulation approaches to help identify near-optimal locations of green roofs and permeable pavements in the catchment scale. The Nondominated Sorting Genetic Algorithm II was connected to the stormwater management model (SWMM) to identify the location of SCMs and characterize the tradeoffs between flow regime alteration and implementation costs. The impact of implementing SCMs is measured by peak flow, runoff volume, and the hydrologic footprint residence (HFR). The HFR is a new stormwater metric that represents dynamics of inundated areas and residence time of flood waves throughout downstream segments. The approach was tested in an illustrative case study of an 11.7-ha urban catchment divided into five subcatchments. The results indicate that locating SCMs in downstream subcatchments can reduce peak flow more effectively, whereas SCMs placed in upstream subcatchments better reduce the HFR. The proposed methodology can help stormwater managers to better assess the combined performance of LID-SCMs in different hydrologic scales and generate guidelines for prioritizing the implementation or retrofitting of urban areas with green infrastructure.

Urban water resources should be managed to meet conflicting demands for environmental health, economic prosperity, and social equity for present and future generations. While the sustainability of water resources can depend on dynamic interactions among natural, social, and infrastructure systems, typical water resource planning and management approaches are based on methodologies that ignore feedbacks and adaptations among these systems. This research develops and demonstrates a new complex adaptive systems approach to model the dynamic interactions among population growth, land-use change, the hydrologic cycle, residential water use, and interbasin transfers. Agent-based and cellular automaton models, representing consumers and policymakers who make land- and water-use decisions, are coupled with hydrologic models. The framework is applied for an illustrative case study to simulate urbanization and the water supply system over a long-term planning horizon. Results indicate that interactions amon...