On the challenges of global entity-aware deep learning models for groundwater level prediction

Hydrology and earth system sciences, Journal Year: 2024, Volume and Issue: 28(3), P. 525 - 543

Published: Feb. 8, 2024

Abstract. The application of machine learning (ML) including deep models in hydrogeology to model and predict groundwater level monitoring wells has gained some traction recent years. Currently, the dominant class is so-called single-well model, where one trained for each well separately. However, developments neighbouring disciplines hydrology (rainfall–runoff modelling) have shown that global models, being able incorporate data several wells, may advantages. These are often called “entity-aware models“, as they usually rely on static differentiate entities, i.e. or catchments surface hydrology. We test two kinds information characterize a global, entity-aware set-up: first, environmental features continuously available thus theoretically enable spatial generalization (regionalization), second, time-series derived from past time series at respective well. Moreover, we random integer entity comparison. use published dataset 108 Germany, evaluate performance terms Nash–Sutcliffe efficiency (NSE) an in-sample out-of-sample setting, representing temporal generalization. Our results show work with mean NSE >0.8 comparable to, even outperforming, models. do not generalize spatially setting (mean <0.7, lower than without information). Strikingly, all variants, regardless type used, basically perform equally both in- out-of-sample. conclusion fact does awareness, but uses merely unique identifiers, raising research question how properly establish awareness Potential future avenues lie bigger datasets, relatively small number might be enough take full advantage Also, more needed find meaningful ML hydrogeology.

Language: Английский

---

Artificial intelligence in groundwater management: Innovations, challenges, and future prospects

International Journal of Science and Research Archive, Journal Year: 2024, Volume and Issue: 11(1), P. 502 - 512

Published: Jan. 26, 2024

The integration of Artificial Intelligence (AI) in groundwater management is a transformative stage, characterized by innovation and challenges. This research paper explores the multilayered application AI this field, dividing its contributions, addressing associated challenges, revealing prospects future potential. AI-driven innovations are designed to revolutionize management, providing precise predictive modeling, real-time monitoring, data integration. However, these face challenges such as interpretability issues, specialized technical expertise requirements, limited quality quantity for effective model performance. In future, holds significant promise management. Advanced models can yield improved predictions behavior, identify vulnerable areas prone pollution depletion, prompt proactive interventions, foster collaborative platforms among scientists, policymakers, local communities. Collaborative driven offer potential synergistic engagement communities, collectively guiding resource Embracing AI's while remains pivotal sustainable resilient practices. By embracing landscape will continue evolve.

Language: Английский

---

Heavy metals prediction in coastal marine sediments using hybridized machine learning models with metaheuristic optimization algorithm

Chemosphere, Journal Year: 2024, Volume and Issue: 352, P. 141329 - 141329

Published: Jan. 29, 2024

Language: Английский

---

Groundwater Level Forecasting Using Machine Learning: A Case Study of the Baekje Weir in Four Major Rivers Project, South Korea

Water Resources Research, Journal Year: 2024, Volume and Issue: 60(5)

Published: May 1, 2024

Abstract Understanding the impact of human‐made structures on groundwater levels is essential, with like dams or weirs presenting unique challenges and opportunities for study. The Baekje weir in South Korea presents an interesting case as has undergone full gate opening, which generally not reservoirs, providing valuable opportunity simulating removal conditions. main objectives are investigation level fluctuations under various operations, distances from weir, seasonal variations. study utilizes observed data that simulates conditions without including scenarios opening. Multiple machine learning algorithms—Random Forest (RF), Artificial Neural Network, Support Vector Regression (SVR), Gradient Boosting, Extreme Boosting (XGBoost)—are used to develop accurate prediction models. models' performance assessed using coefficient determination, Root mean square error (RMSE), Mean Absolute Error (MAE) indices, visualized through Taylor diagrams. Results indicate XGBoost outperforms other models all three groups during both training testing phases. Specifically, surpasses RF by 2.09% ( R 2 ), 5.66% 10.1% training, SVR 11.2% 42.0% 129.2% testing. Additionally, generates maps, a practical tool managing systems informing decision‐making operations. This only sheds light dynamic relationship between operations but also provides actionable insights effective water management similar hydrological settings.

Language: Английский

---

Agent-based models of groundwater systems: A review of an emerging approach to simulate the interactions between groundwater and society

Environmental Modelling & Software, Journal Year: 2024, Volume and Issue: 175, P. 105980 - 105980

Published: Feb. 17, 2024

Understanding how society can address and mitigate threats to groundwater sustainability remains a pressing challenge in the Anthropocene era. This article presents first comprehensive critical review of coupling Groundwater Models Agent-Based (GW-ABMs) four key challenges: (1) adequately representing human behaviour, (2) capturing spatial temporal variations, (3) integrating two-way feedback loops between social physical systems, (4) incorporating water governance structures. Our findings indicate growing effort model bounded rationality behaviour (Challenge 1 or C1) dominant focus on policy applications (C4). Future research should data scarcity issues through Epstein's Backward approach (C2), capture feedbacks via tele-coupled GW-ABMs, explore other modelling techniques like Analytic Elements (C3). We conclude with recommendations thrust future GW-ABMs highest standards, aiming enhance their acceptance impact decision-making formulation for sustainable management.

Language: Английский

---