Article
Computer Science, Interdisciplinary Applications
Shicheng Li, James Yang
Summary: This study develops an improved water-level prediction framework by coupling machine learning models with ensemble algorithms. The ensemble models generally enhance the prediction efficiency and accurately capture extreme values. The models based on the boosting algorithm perform the best and also show good performance in multi-step-ahead forecasts.
ENGINEERING WITH COMPUTERS
(2023)
Article
Engineering, Civil
Farshid Rezaei, Rezvane Ghorbani, Najmeh Mahjouri
Summary: This paper improves the accuracy of daily and monthly runoff discharge forecasts by developing an ensemble model based on Bayesian maximum entropy (BME), and evaluates its performance by comparing it with 12 other ensemble models. The results show that the ensemble models, especially the BME and ANN-based ensemble models, significantly improve both monthly and daily river discharge forecasts.
WATER RESOURCES MANAGEMENT
(2022)
Article
Geosciences, Multidisciplinary
Kieran M. R. Hunt, Gwyneth R. Matthews, Florian Pappenberger, Christel Prudhomme
Summary: Accurate river streamflow forecasts are crucial for water security, flood preparedness, agriculture, and industry. Traditional physics-based models have improved over time, but are limited by empirical relationships and lack of data. Artificial neural networks, particularly long short-term memory (LSTM) networks, have shown promise in simulating non-linear systems. Hybrid forecasting systems combining physics-based approaches and statistical techniques have been explored for hydrological applications. This study evaluates the performance of LSTM networks in predicting streamflow at 10 river gauge stations in the western United States and compares it with physics-based models. The results demonstrate the potential of LSTM networks in improving hydrological forecasting.
HYDROLOGY AND EARTH SYSTEM SCIENCES
(2022)
Article
Engineering, Civil
Xin Liu, Liping Zhang, Dunxian She, Jie Chen, Jun Xia, Xinchi Chen, Tongtiegang Zhao
Summary: This paper develops an integrated postprocessing framework for hydrometeorological ensemble forecasts, which aims to address the data correction issue for hydrological models. By using a reasonable design of canonical events and employing postprocessed ensemble precipitation forecasts, the performance of hydrological forecasts can be improved in terms of lead times and accuracy. The Bayesian model averaging (BMA) scheme further enhances the forecast effect by generating more skillful and reliable probabilistic hydrological forecasts.
JOURNAL OF HYDROLOGY
(2022)
Article
Environmental Sciences
Rui Yang, Hui Liu, Yanfei Li
Summary: This study introduces a new method of uncertainty quantification driven by point prediction to solve the engineering problem of water quality forecasting under the influence of complex environmental factors. By constructing a multi-factor correlation analysis system and utilizing singular spectrum analysis, the model improves data fusion interpretability and reduces volatility of water quality data. The model also adopts a multi-resolution-multi-objective optimization ensemble method to mine deeper potential information. Experimental results show that the model outperforms existing models in quantifying the uncertainty of water quality prediction.
Article
Geosciences, Multidisciplinary
Yuxue Guo, Xinting Yu, Yue-Ping Xu, Hao Chen, Haiting Gu, Jingkai Xie
Summary: This study developed an AI-based management methodology integrating multi-step streamflow forecasts and multi-objective reservoir operation optimization for water resource allocation. The study found that both GRU and LSTM performed equally well on streamflow forecasts, with GRU potentially being the preferred method due to its simpler structure and less modeling time. Higher forecast performance could lead to improved reservoir operation, while uncertain forecasts were more valuable than deterministic forecasts.
HYDROLOGY AND EARTH SYSTEM SCIENCES
(2021)
Article
Engineering, Civil
Saumitra Rai, Rallapalli Srinivas, Joe Magner
Summary: Riparian corridors are crucial for maintaining downstream ecosystem services, such as water quantity and quality. This study proposes a comprehensive decision support framework for managing riparian zones in the Minnesota River Basin. The framework integrates fuzzy analytical hierarchy process and fuzzy inference system to suggest suitable best management practices based on stream features, riparian characteristics, and riparian functions. The results highlight the importance of riparian revegetation, buffer strips, wetland management, the addition of woody debris, and an increase in tree cover as effective BMPs.
JOURNAL OF HYDROLOGY
(2022)
Article
Meteorology & Atmospheric Sciences
Cecile Penland, Prashant D. Sardeshmukh
Summary: Forecast ensembles are often skewed due to the state dependence of chaotic system dynamics. Even in the simplest systems with state-dependent noise, skewness can be generated even when initial and climatological forecast distributions are symmetric. Systems with state-dependent noise typically have skewed and heavy-tailed forecast distributions, which has implications for forecasting extreme anomaly risks. Misrepresenting state-dependent noise in ensemble forecast systems leads to state-dependent errors in forecast probability distributions, which cannot be corrected by simple bias corrections. The ensemble standard deviation is commonly used as a metric for ensemble spread, and the ensemble skewness can be used as a metric for the difference between ensemble-mean and most likely forecast, as well as the risk of extreme deviations from the ensemble-mean forecast.
QUARTERLY JOURNAL OF THE ROYAL METEOROLOGICAL SOCIETY
(2022)
Article
Environmental Sciences
C. Dionisio Perez-Blanco, Francesco Sapino
Summary: With water scarcity becoming more severe, irrigators are adopting modern irrigation systems to increase water consumption for crops and mitigate negative impacts on production. This study uses a multi-model ensemble approach to assess the responses of irrigators to pecuniary compensations designed to sustain irrigation-dependent ecosystem services. The findings suggest that a conservationist strategy has superior economic performance and robustness compared to an autonomous adaptation strategy. However, sensible incentives are needed to prevent irrigators from adopting irrigation technologies despite the compensations. In severe climate change scenarios, additional payments for watershed services will be necessary to ensure the sustainability of irrigation-dependent ecosystem services.
WATER RESOURCES RESEARCH
(2022)
Article
Environmental Sciences
J. Eli Asarian, Crystal Robinson, Laurel Genzoli
Summary: Low streamflows can increase vulnerability to warming, impacting coldwater fish. Water managers need tools to quantify these impacts and predict future water temperatures. Contrary to most statistical models' assumptions, many seasonally changing factors (e.g., water sources and solar radiation) cause relationships between flow and water temperature to vary throughout the year.
WATER RESOURCES RESEARCH
(2023)
Article
Geosciences, Multidisciplinary
Manuel C. Almeida, Pedro S. Coelho
Summary: The prediction of river water temperature is crucial in environmental science, especially for low-order rivers with limited temperature datasets. In this study, five models were used to predict the water temperature of 83 rivers with a large scarcity of forcing datasets. The results emphasize the importance of hyperparameter optimization and suggest that all the models considered in this study are crucial when the number of predictor variables and observed river water temperature values are limited.
GEOSCIENTIFIC MODEL DEVELOPMENT
(2023)
Article
Engineering, Marine
R. Vettor, C. Guedes Soares
Summary: This study investigates different methodologies for predicting the uncertainties in ship fuel consumption, including ensemble weather forecasts, probabilistic approaches, and a classical first-order second-moment method. By comparing the results from different methods, it is found that the actual fuel consumption falls within the predicted range with reasonable agreement.
Article
Environmental Sciences
Matteo Giuliani, Jonathan R. Lamontagne, Mohamad Hejazi, Patrick M. Reed, Andrea Castelletti
Summary: The implementation of global climate change mitigation policies at the local scale can have negative impacts on interconnected water-energy-food systems. Research on the Zambezi Watercourse in southern Africa demonstrates that mitigation policies related to land-use change emissions can increase local water demands and pose risks to the stability of the entire system. Similar vulnerabilities may exist in other river basins in Africa.
NATURE CLIMATE CHANGE
(2022)
Article
Engineering, Civil
Velpuri Manikanta, Jew Das, K. Nikhil Teja, N. V. Umamahesh
Summary: This study examines the capability of ensemble precipitation forecasts obtained from two NWP models, and post-processes the raw ensemble members using two methods. The findings suggest that QRF post-processed forecasts are superior to other forecasts in terms of all verification measures at shorter lead times. However, the skill of both raw and post-processed forecasts declines at higher lead times.
JOURNAL OF HYDROLOGY
(2023)
Article
Environmental Sciences
Christa A. Kelleher, Heather E. Golden, Stacey A. Archfield
Summary: The study found that while most sites showed annual warming trends, these trends obscured sub-annual cooling trends at many sites. Monthly trend anomalies were spatially organized, showing persistent regional patterns at both reference and human-impacted sites.
ENVIRONMENTAL RESEARCH LETTERS
(2021)
Article
Environmental Sciences
Al Mahdi Saadi, Amina Msilini, Christian Charron, Andre St-Hilaire, Taha B. M. J. Ouarda
Summary: This study used GAM and MARS models to estimate the areas of potential thermal refuges in rivers, finding that MARS outperformed GAM in forecasting and estimating the variability of thermal refuge areas. The results highlight the importance of thermal refuges in rivers for fish and offer a relatively simple approach for water resources managers to protect these habitats.
RIVER RESEARCH AND APPLICATIONS
(2022)
Article
Ecology
Olfa Abidi, Andre St-Hilaire, Taha B. M. J. Ouarda, Christian Charron, Claudine Boyer, Anik Daigle
Summary: This study focuses on modeling water temperature indices for Atlantic salmon and finds that the non-linear generalized additive model performs better than the linear multiple linear regression model in estimating these indices. Additionally, the use of hierarchical clustering analysis for delineating homogeneous regions is considered to be more flexible and leads to better performance compared to neighborhood-based approaches.
ECOLOGICAL INFORMATICS
(2022)
Article
Ecology
Jeremie Boudreault, Andre St-Hilaire, Fateh Chebana
Summary: Habitat suitability curves (HSC) are important for fish habitat models, but existing methods do not consider differences at small scales. This study uses functional data analysis (FDA) to develop more accurate site-specific HSC. The results show that FDA is an innovative framework that can be used to predict more representative site-specific HSC.
ECOLOGICAL MODELLING
(2022)
Article
Environmental Sciences
Carolyn J. M. Brown, R. Allen Curry, Michelle A. Gray, Jennifer Lento, Deborah L. MacLatchy, Wendy A. Monk, Scott A. Pavey, Andre St-Hilaire, Bernhard Wegscheider, Kelly R. Munkittrick
Summary: In most countries, major development projects must undergo an Environmental Impact Assessment (EIA) process. However, many assessments fail to link pre- and post-development monitoring effectively. Fish are important components of EIA evaluations, and the concept of Ecosystem Services (ES) provides a framework centered around the needs and benefits of fish. Focusing on the critical needs of fish in an environmental monitoring framework can better align risk, development, and monitoring assessment processes.
ENVIRONMENTAL MANAGEMENT
(2022)
Article
Environmental Sciences
Habiba Ferchichi, Andre St-Hilaire, Taha B. M. J. Ouarda, Benoit Levesque
Summary: Water temperature is important for the equilibrium of aquatic systems and the health of aquaculture biota. Machine learning models and regression model were developed to estimate coastal sea surface temperatures in Eastern Canada, with the artificial neural network model showing the highest accuracy in SST prediction.
ESTUARIES AND COASTS
(2022)
Article
Water Resources
Siavash P. Markhali, Annie Poulin, Marie-Amelie Boucher
Summary: The main objective of this paper is to quantify the uncertainty associated with the spatio-temporal variability of catchment descriptors in distributed hydrology models, and its impact on simulating flooding events. The researchers use an ensemble approach to characterize the uncertainties of spatio-temporal variations. They calibrate two hydrological models and simulate six catchments with different sizes and characteristics in southern Quebec. The results show that the spatial resolution of the models has a significant effect on the uncertainty, while catchment size and temporal scale have a minor role.
HYDROLOGICAL PROCESSES
(2022)
Article
Water Resources
Philippe Gatien, Richard Arsenault, Jean-Luc Martel, Andre St-Hilaire
Summary: It is important to regulate river water temperature in order to protect aquatic organisms and their habitat. This study uses a hydraulic model combined with meteorological data to simulate water temperatures in a river system in British Columbia, Canada, and achieves accurate results. The study also finds that certain boundary conditions and data density have limitations on the downstream water temperatures.
CANADIAN WATER RESOURCES JOURNAL
(2023)
Article
Water Resources
Mostafa Khorsandi, Andre St-Hilaire, Richard Arsenault
Summary: This study compares the single-site and multisite calibration methods for thermal modelling and finds that multisite calibration and parameter upscaling provide better performance and shorter computing time compared to single-site calibration.
HYDROLOGICAL SCIENCES JOURNAL
(2022)
Article
Engineering, Civil
Veronique Dubos, Ilias Hani, Taha B. M. J. Ouarda, Andre St-Hilaire
Summary: In cold boreal regions, the risk of spring flooding is determined by the intensity of peak flow, and short-term forecasting of peak flow intensity is uncertain and depends on ongoing snowmelt conditions. This study proposes a simple operational model based on the Generalized Additive Model (GAM) for short-term spring freshet peak flow forecasting. The model uses publicly available daily hydrological and meteorological data and was tested on five rivers in Quebec, Canada.
JOURNAL OF HYDROLOGY
(2022)
Article
Engineering, Civil
Adoubi Vincent De Paul Adombi, Romain Chesnaux, Marie-Amelie Boucher
Summary: This study proposes and develops a theory-guided machine learning model for modeling the behavior of a real aquifer and compares its performance with a numerical model and a traditional MLP model. The results show that the theory-guided model outperforms the other two models in capturing the spatiotemporal dynamics of groundwater flow.
JOURNAL OF HYDROLOGY
(2022)
Article
Water Resources
Valerie Jean, Marie-Amelie Boucher, Anissa Frini, Dominic Roussel
Summary: Flood forecasts are important for decision-making, but technical improvements in the forecasting systems do not always lead to a reduction in damages. Streamflow forecasts, especially ensemble and probabilistic forecasts, provide more information but are more difficult to interpret. Contextualizing the forecasts, representing the time evolution of flood events accurately, and improving communication between forecasters and decision-makers are highlighted as important aspects by participants.
CANADIAN WATER RESOURCES JOURNAL
(2023)
Article
Water Resources
A. F. Nolin, M. P. Girardin, J. F. Adamowski, R. Barzegar, M. - A. Boucher, J. C. Tardif, Y. Bergeron
Summary: This study examines the changes in spring mean discharge and flood drivers in the Upper Harricana River in northwestern Quebec over the last 250 years. The results show that although the decline in snow cover and increase in temperature may offset each other, two models project an increase in high mean spring discharge for the region. The study highlights the importance of estimating future flood risks based on multi-model ensembles.
JOURNAL OF HYDROLOGY-REGIONAL STUDIES
(2023)
Review
Geosciences, Multidisciplinary
Louise J. Slater, Louise Arnal, Marie-Amelie Boucher, Annie Y. -Y. Chang, Simon Moulds, Conor Murphy, Grey Nearing, Guy Shalev, Chaopeng Shen, Linda Speight, Gabriele Villarini, Robert L. Wilby, Andrew Wood, Massimiliano Zappa
Summary: Hybrid hydroclimatic forecasting systems integrate predictions from physics-based models and data-driven methods to improve meteorological and hydroclimatic variables and events predictions. These systems have gained attention due to advances in prediction systems, the strengths of machine learning, and increased access to computational resources. Key challenges and opportunities for further research include obtaining explainable results, assimilating human influences, improving predictive skill through ensemble techniques, creating seamless prediction schemes, incorporating initial conditions, acknowledging spatial variability, and increasing operational uptake.
HYDROLOGY AND EARTH SYSTEM SCIENCES
(2023)
Article
Environmental Sciences
Pierre Masselot, Taha B. M. J. Ouarda, Christian Charron, Celine Campagna, Eric Lavigne, Andre St-Hilaire, Fateh Chebana, Pierre Valois, Pierre Gosselin
Summary: This study models summer heat-related mortality in two metropolitan areas of Quebec, Canada using seven climate indices. The results show that the Atlantic Multidecadal Oscillation is the best predictor of heat-related mortality and can predict up to 20% of the interannual variability.
ENVIRONMENTAL EPIDEMIOLOGY
(2022)
Article
Geosciences, Multidisciplinary
Jing Xu, Francois Anctil, Marie-Amelie Boucher
Summary: Forecast uncertainties are inevitable in deterministic analysis of dynamical systems. Ensemble forecasting is an effective tool to represent error growth and capture uncertainties. This study compares the performance of evolutionary multi-objective optimization with a conventional state-of-the-art post-processor in eliminating forecast biases and maintaining proper dispersion. The evolutionary multi-objective optimization method demonstrated superiority in communicating with end-users for performance improvement.
HYDROLOGY AND EARTH SYSTEM SCIENCES
(2022)