Journal of Water Resources Planning and Management

This paper provides a comprehensive review of tree-based models and their application in condition assessment and prediction of water, wastewater, and sewer pipe failures. Tree-based models have gained significant attention in recent years due to their effectiveness in capturing complex relationships between parameters of systems and their ability in handling large data sets. This study explores a range of tree-based models, including decision trees and ensemble trees utilizing bagging, boosting, and stacking strategies. The paper thoroughly examines the strengths and limitations of these models, specifically in the context of assessing the pipes’ condition and predicting their failures. In most cases, tree-based algorithms outperformed other prevalent models. Random forest was found to be the most frequently used approach in this field. Moreover, the models successfully predicted the failures when augmented with a richer failure data set. Finally, it was identified that existing evaluation metrics might not be necessarily suitable for assessing the prediction models in the water and sewer networks.

Sustainable urban stormwater management presents challenges due to uncertainties and conflicting stakeholder preferences. In this regard, our research introduces a comprehensive framework that integrates self-organization map (SOM) and many-objective robust decision-making (MORDM) methods to bridge the knowledge gap between deep uncertainty and conflict resolution. To do so, first, MORDM generates optimal strategies for the implementation of low-impact development (LID) practices and quantifies their robustness under different uncertain scenarios. Then, the results are integrated into the SOM framework, wherein multi-dimensional decision space can be classified and stakeholders are enabled to navigate their satisfaction to reach a subset of robust strategies. Finally, to determine the most suitable management strategy, a novel game theory-based group multicriteria decision-making (GT-GMCDM) method is employed, which can effectively consider stakeholders’ conflicts in the selection process. Results from applying the proposed methodology in District 10 of Tehran municipality, Iran, demonstrate that the selected alternative can reduce TSS contamination load and runoff volume (RV) by 30% and 27%, respectively. Indeed, findings prove that in addition to improving stormwater quality and quantity, the suggested methodology can simultaneously consider stakeholders’ preferences and find a reliable strategy. This study also offers valuable insights into robust stormwater management, with a focus on ensuring the sustainability of urban environments.

Climatic phenomena, particularly hydrological droughts, have led to significant changes in reservoir operation strategies. Multipurpose reservoir operations are essential for effectively managing stored water resources for various activities like electricity generation and agricultural irrigation. Despite considerable efforts to support decision making for each economic activity, there remains a weak integration across these sectors in joint analyses. To address this, an integrated approach combining a model of a large power system and a model at the basin scale is proposed to analyze the operation of both power and agricultural systems. This approach allows evaluation of the operating policies of a multipurpose reservoir and its performance at both the local and regional scales under different hydrological scenarios. A modification is implemented whereby the priority of water extraction to agricultural users is increased. Its effects are assessed for different hydrological trajectories in a case study in the Laja Lake basin in southern Chile, the biggest Chilean basin with a capacity of up to $\mathrm{5,500}{\mathrm{Hm}}^3$. The Laja Lake, a multipurpose reservoir with substantial hydroelectric generation capacity and extensive agricultural areas plays a crucial role in the operation of the national power system. Based on an analysis of 2025, it is demonstrated that hydrological changes directly impact electrical and agricultural performance. Drought conditions increase thermal generation, costs, emission intensity, and water deficits. Furthermore, the policy modification reveals tradeoffs between the power sector’s emissions and agricultural water deficits. For drier scenarios, increasing agricultural extraction priority results in low additional operational costs and emissions from the power system, which supports adopting a policy aligned with netzero objectives.

Changes in demand, various hydrological inputs, and environmental stressors are among the issues that reservoir managers and policymakers face on a regular basis. These concerns have sparked interest in applying different techniques to determine reservoir operation policy decisions. As the resolution of the analysis increases, it becomes more difficult to effectively represent a real-world system using traditional methods such as dynamic programming and stochastic dynamic programming for determining the best reservoir operation policy. One of the challenges is the “curse of dimensionality,” which means the number of samples needed to estimate an arbitrary function with a given level of accuracy grows exponentially with respect to the number of input variables (i.e., dimensionality) of the function. Deep reinforcement learning (DRL) is an intelligent approach to overcome the curses of stochastic optimization problems for reservoir operation policy decisions. To our knowledge, this study is the first attempt that examines various novel DRL continuous-action policy gradient methods, including deep deterministic policy gradients, twin delayed DDPG (TD3), and two different versions of Soft Actor-Critic (SAC18 and SAC19) for optimizing reservoir operation policy. In this study, multiple DRL techniques were implemented to find an optimal operation policy for Folsom Reservoir in California. The reservoir system supplies agricultural, municipal, hydropower, and environmental flow demands and flood control operations to the City of Sacramento. Analysis suggests that the TD3 and SAC are robust to meet the Folsom Reservoir’s demands and optimize reservoir operation policies. Experiments on continuous-action spaces of reservoir policy decisions demonstrated that the DRL techniques can efficiently learn strategic policies in spaces and can overcome the curse of dimensionality and modeling.

Abstract

This work had the purpose to provoke discussion about more adaptive reservoir operations. This work created online collaborative modeling environments by using web spreadsheets (Google Sheets) during a video conference session. In each session, up to 6 collaborators immersed in water user roles. Then the collaborators consumed, saved, and traded water in response to their available water, other choices, and real-time discussion of choices. The collaboration was an alternative to prior offline or high-performance computing efforts that programmed water allocation rules to try to satisfy forecasted water demands across hydrologic scenarios. The collaboration also differed from prior efforts that excluded stakeholders, extracted data from participants, had a lead modeler or facilitation team mediate participant interactions with a model, or built a model then presented findings at the project end. In model sessions, collaborators improved more adaptive operations rather than separately developed and tested competing alternatives. 26 Colorado River managers and experts demonstrated use for a combined Lake Powell-Lake Mead water bank. The author used discussion and feedback to synthesize 10 lessons. For example, model to provoke discussion and insights rather than propose a solution, solicit feedback early, allow trades to increase flexibility, and recognize limits of model acceptability and adoption. To generate more actionable insights, next steps are engage multiple groups within the same model session and explore more management alternatives.

Practical Applications

Researchers, consultants, facilitators, and project leaders can build their own online collaborative model environments for their study system(s). Leaders can invite basin managers, stakeholders, colleagues, students, and the public to collaborate during video conference or in-person sessions. Leaders can also use the collaborative model environment(s) to prompt discussion of future basin operations, solicit feedback to improve operations, and/or make the model environments more user friendly. 26 Colorado River Basin managers and experts demonstrated use for a combined Lake Powell-Lake Mead water bank.