Article
Computer Science, Artificial Intelligence
Arun M. George, Sounak Dey, Dighanchal Banerjee, Arijit Mukherjee, Manan Suri
Summary: IoT-based automated systems require efficient online time-series analysis and forecasting, which is challenging to achieve on low-cost constrained edge devices. This study proposes a novel spiking reservoir based network that relies on temporal spike encoding and feedback-driven online learning mechanism for online time series forecasting. The network outperforms conventional methods like SARIMA, Online ARIMA, Stacked LSTM, achieving up to 8% higher R2 score while using negligible buffer memory.
Article
Multidisciplinary Sciences
Vitor Hugo Serravalle Reis Rodrigues, Paulo Roberto de Melo Barros Junior, Euler Bentes dos Santos Marinho, Jose Luis Lima de Jesus Silva
Summary: Developing accurate models for groundwater control is crucial for managing and planning water resources from aquifer reservoirs. The proposed Wavelet Gated Multiformer combines the strengths of a vanilla Transformer and a Wavelet Crossformer to improve the model's predictive capabilities by computing the relationships between time-series points and finding trending periodic patterns. This model outperforms other transformer-like models in terms of Mean Absolute Error reduction.
SCIENTIFIC REPORTS
(2023)
Article
Engineering, Civil
Julien Monteil, Anton Dekusar, Claudio Gambella, Yassine Lassoued, Martin Mevissen
Summary: This work investigates the use of deep learning models for long-term large-scale traffic prediction tasks, focusing on scalability. By analyzing 14 weeks of speed observations from over 1000 segments in downtown Los Angeles, different machine learning and deep learning predictors were studied, along with their scalability to larger areas. The study shows that modeling temporal and spatial features into deep learning predictors can be beneficial for long-term predictions, while simpler predictors achieve satisfactory performance for link-based and short-term forecasting, with a trade-off in prediction accuracy, horizon, training time, and model sizing discussed.
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
(2022)
Article
Computer Science, Artificial Intelligence
Rakshitha Godahewa, Kasun Bandara, Geoffrey Webb, Slawek Smyl, Christoph Bergmeir
Summary: Ensembling techniques are used to improve the performance of Global Forecasting Models (GFM) and univariate models in heterogeneous datasets. A new clustered ensembles methodology is proposed to train multiple GFMs on different clusters of series, achieving higher accuracy than baseline models.
KNOWLEDGE-BASED SYSTEMS
(2021)
Article
Computer Science, Information Systems
Giambattista Gruosso, Giancarlo Storti Gajani
Summary: The availability of reliable photovoltaic power forecasting tools is crucial for the dissemination of this technology. Edge computing can localize and make predictions feasible with the use of small, low power, and inexpensive devices. This article explores prediction methods based on Artificial Neural Networks (ANNs) models, and investigates techniques to reduce their cost. The aging effects of solar panels are also considered.
Article
Engineering, Civil
Xingsheng Shu, Yong Peng, Wei Ding, Ziru Wang, Jian Wu
Summary: In this study, two innovative models, DirCNN and DRCNN, are proposed for multi-step-ahead monthly streamflow forecasting. Compared to traditional models, DirCNN and DRCNN outperform the comparison models and demonstrate good consistency in forecasting accuracy with an increase in lead time. The stacking order of candidate sequences has little effect on the forecasting accuracy of DirCNN and DRCNN.
WATER RESOURCES MANAGEMENT
(2022)
Article
Computer Science, Theory & Methods
Konstantinos Benidis, Syama Sundar Rangapuram, Valentin Flunkert, Yuyang Wang, Danielle Maddix, Caner Turkmen, Jan Gasthaus, Michael Bohlke-Schneider, David Salinas, Lorenzo Stella, Francois-Xavier Aubet, Laurent Callot, Tim Januschowski
Summary: Deep learning based forecasting methods have achieved remarkable success in time series prediction and have become widely used in industrial applications and forecasting competitions. This article provides an introduction to deep forecasting, discussing important building blocks and summarizing recent literature.
ACM COMPUTING SURVEYS
(2023)
Article
Thermodynamics
Mushrafi Munim Sushmit, Islam Mohammed Mahbubul
Summary: This study explores the application of quantum-enhanced deep feedforward neural networks (FFNs) and fully connected quantum neural networks (QNNs) in solar irradiance forecasting, showing their superior performance compared to traditional models. The quantum-enhanced FFNs demonstrate robust and competitive performance, highlighting the promising prospects of quantum-integrated techniques in improving the accuracy of solar irradiance prediction models.
ENERGY CONVERSION AND MANAGEMENT
(2023)
Article
Computer Science, Artificial Intelligence
Zonghan Wu, Da Zheng, Shirui Pan, Quan Gan, Guodong Long, George Karypis
Summary: This article introduces a novel spatial-temporal graph neural network called TraverseNet for capturing the spatial-temporal dependencies in traffic data. Compared to other spatial-temporal neural networks, TraverseNet views space and time as an inseparable whole and utilizes message traverse mechanisms to explore the dependencies in the spatial-temporal graph.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2022)
Article
Computer Science, Artificial Intelligence
Ming Jin, Yu Zheng, Yuan-Fang Li, Siheng Chen, Bin Yang, Shirui Pan
Summary: This paper proposes a continuous model, MTGODE, to forecast multivariate time series by overcoming the limitations of discrete neural architectures, high complexity, and reliance on graph priors. MTGODE utilizes dynamic graph neural ordinary differential equations to unify spatial and temporal message passing, resulting in superior forecasting performance on benchmark datasets.
IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING
(2023)
Article
Geochemistry & Geophysics
Andrew P. Barnes, Thomas R. Kjeldsen, Nick McCullen
Summary: This study proposes a new methodology for improving regional precipitation forecasts using video-based convolutional neural networks (CNNs). Comparing the CNN models with GloSEA5 forecasts, it is found that CNN models outperform GloSEA5 in predicting extreme events, and treating the forecasts as an ensemble can further improve prediction accuracy.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2022)
Article
Energy & Fuels
Haolin Yang, Kristen R. Schell
Summary: The study investigates the impact of transfer learning on price prediction DNN representations, showing that it improves accuracy. The GRU-TL architecture, pre-trained on a hybrid dataset, outperforms statistical and deep learning benchmarks. Transfer learning enables the pre-trained DNN representation to learn target dataset features more accurately.
Article
Engineering, Electrical & Electronic
Di Wu, Weixuan Lin
Summary: Accurate short-term electric load forecasting is crucial for the safety and efficient operation of modern electric power systems. The graph neural network has been successfully used in STLF by leveraging the spatial dependency between residential households. However, training GNN models requires a large amount of data, which may be lacking in newly built residential areas.
IEEE TRANSACTIONS ON SMART GRID
(2023)
Article
Multidisciplinary Sciences
Rusul L. Abduljabbar, Hussein Dia, Pei-Wei Tsai
Summary: This paper develops and evaluates BiLSTM short term traffic forecasting models using data from a congested freeway in Melbourne, Australia. The results show superior performance of BiLSTM for multiple prediction horizons under both base year and future year scenarios.
SCIENTIFIC REPORTS
(2021)
Article
Energy & Fuels
Salih Gunduz, Umut Ugurlu, Ilkay Oksuz
Summary: This paper proposes using transfer learning to utilize information from other electricity price markets for forecasting. The experiments show that transfer learning significantly improves the electricity price forecasting performance, and the method outperforms state-of-the-art algorithms.
SUSTAINABLE ENERGY GRIDS & NETWORKS
(2023)
Article
Meteorology & Atmospheric Sciences
Hal Ritchie, Stephane Belair, Natacha B. Bernier, Mark Buehner, Martin Charron, Vincent Fortin, Louis Garand, Pieter Houtekamer, Syed Husain, Stephane Laroche, Jean-Francois Lemieux, Hai Lin, Ron McTaggart-Cowan, Jason Milbrandt, Herschel Mitchell, Pierre Pellerin, Janusz Pudykiewicz, Leo Separovic, Gregory C. Smith, Monique Tanguay, Paul A. Vaillancourt
Summary: Contributions of Recherche en Prevision Numerique (RPN) to Numerical Weather Prediction (NWP): A review article invited by Atmosphere-Ocean
Article
Meteorology & Atmospheric Sciences
C. Garnaud, M. MacKay, V Fortin
Summary: This study evaluates the impact of the Canadian Small Lake Model (CSLM) on surface water temperature, ice phenology, and near-surface atmospheric conditions. The results show that CSLM improves lake surface temperature and ice cover simulation, leading to better initialization of lake conditions in operational systems.
JOURNAL OF ADVANCES IN MODELING EARTH SYSTEMS
(2022)
Article
Engineering, Civil
Etienne Gaborit, Vincent Fortin, Dorothy Durnford
Summary: This study investigates the benefit of the dynamically zoned target release (DZTR) reservoir model to improve upon the natural lake model for storage and outflow simulations of regulated reservoirs. Results show that DZTR brings significant improvements upon the natural lake model, with better outflow simulations and more realistic storage simulations even without observed lake level data. However, automatic calibration of the DZTR model parameters did not significantly improve results, and it is considered risky to implement DZTR without any flow observations downstream of a reservoir.
CANADIAN JOURNAL OF CIVIL ENGINEERING
(2022)
Article
Environmental Sciences
V. Vionnet, M. Verville, V. Fortin, M. Brugman, M. Abrahamowicz, F. Lemay, J. M. Theriault, M. Lafaysse, J. A. Milbrandt
Summary: The phase of precipitation in mountains greatly affects the snow cover evolution. This study tests the impact of atmospheric-based precipitation-phase partitioning methods (PPMs) on snowpack simulations using the Crocus snowpack scheme. The results show that the snow-level based PPM improves the estimation of snowfall occurrence and snow water equivalent (SWE) simulation compared to other methods. The study highlights the importance of detailed evaluation of precipitation phase and the benefit of post-processed snow level in mountain snow hydrology.
WATER RESOURCES RESEARCH
(2022)
Article
Environmental Sciences
Marinela del Carmen Valencia Giraldo, Simon Ricard, Francois Anctil
Summary: There is ongoing debate about whether probabilistic (top-down) or possibilistic (bottom-up) approaches are more suitable for estimating potential future climate impacts. In the context of deep uncertainty, bottom-up approaches that assess the sensitivity and vulnerability of systems to climate changes have become more popular. This study proposes a refined framework that combines the scenario-neutral method of the bottom-up approach with elements of the top-down approach. The results reveal regional and differential behaviors of hydroclimatology and low flows under different climate scenarios.
Article
Meteorology & Atmospheric Sciences
Barbara Casati, Tom Robinson, Francois Lemay, Morten Koltzow, Thomas Haiden, Eva Mekis, Franck Lespinas, Vincent Fortin, Gabrielle Gascon, Jason Milbrandt, Greg Smith
Summary: As part of the Year of Polar Prediction (YOPP), Environment and Climate Change Canada (ECCC) developed the Canadian Arctic Prediction System (CAPS), which outperformed other operational systems in predicting surface variables in the Arctic. A verification exercise during YOPP identified strengths, weaknesses, and systematic behaviors of the Canadian deterministic prediction systems at high latitudes. This exercise also led to the development of better verification practices for surface variables in polar regions.
Article
Water Resources
Adrien Pierre, Daniel F. Nadeau, Antoine Thiboult, Alain N. Rousseau, Alain Tremblay, Pierre-Erik Isabelle, Francois Anctil
Summary: Water bodies such as lakes and reservoirs influence the regional climate through the evaporation of water. This study analyzed in-situ observations of a reservoir in a subarctic environment to understand its impact. The results showed that the annual evaporation rate was 590 +/- 66 mm, accounting for approximately 51% of the annual precipitation. The study also revealed the opposite diurnal cycles of sensible and latent heat fluxes during the open water period.
HYDROLOGICAL PROCESSES
(2023)
Article
Engineering, Civil
Michael Osina Torres, Amaury Tilmant, Emixi Valdez Medina, Francois Anctil, Maria-Helena Ramos
Summary: Improving the operational effectiveness of hydropower systems is crucial due to the shift to renewable energy sources and increasing costs associated with new hydro facilities. This study focuses on the relationship between short-term streamflow forecasts and hydropower generation, as well as the impact of uncertainties on energy output. A numerical experiment using hydrologic ensemble forecasts and reservoir optimization models was conducted in Canada. The results show that forecast quality affects energy production, but it is not a one-to-one causal relationship. Additionally, the diversity of hydrological models contributes to energy production, suggesting the value of model structure diversity.
JOURNAL OF WATER RESOURCES PLANNING AND MANAGEMENT
(2023)
Article
Water Resources
Nicolas Fontaine, Marie-Amelie Boucher, Francois Anctil, Jean Odry, Simon Lachance-Cloutier, Vincent Fortin, Richard Turcotte
Summary: This study explores the possibility of improving regional hydrological forecasts by combining forecasts from large-scale hydrological models and regional systems using simple methods. The outputs from the two systems are combined using various methods, and the performance is evaluated using multiple well-known metrics. The results show that even though the regional system performs better, simple weighted combinations can improve regional hydrological forecasts, with similar performance improvement observed for ungauged basins.
CANADIAN WATER RESOURCES JOURNAL
(2023)
Article
Engineering, Civil
F. Sergeant, R. Therrien, F. Anctil, Laura Gatel
Summary: In cold regions, climate warming causes permafrost thaw and changes the groundwater flow dynamics from local to regional systems. The recession slope of arctic catchment hydrograph is linearly related to permafrost thawing depth, making recession analysis a valuable method to study permafrost thawing dynamics in areas with limited permafrost observations. However, the linear relationship is influenced by permafrost extent, landscape topography, and aquifer properties.
JOURNAL OF HYDROLOGY
(2023)
Article
Geosciences, Multidisciplinary
Pierre Valois, Francois Anctil, Genevieve Cloutier, Maxime Tessier, Naomie Herpin-Saunier
Summary: The frequency and severity of flooding events are expected to increase with climate change in Quebec. A longitudinal study conducted in the province examined the adaptive behaviors of residents in high flood risk zones, finding that there has been no significant increase in adaptive behavior between 2015 and 2019. However, households that have experienced a flood or flood alert in the past are more likely to adapt. The study also identified income, flood experience, and perception of living in a flood-prone zone as important predictors of behavior adoption rates.
INTERNATIONAL JOURNAL OF DISASTER RISK REDUCTION
(2023)
Article
Geosciences, Multidisciplinary
Simon Ricard, Philippe Lucas-Picher, Antoine Thiboult, Francois Anctil
Summary: A simplified hydroclimatic modelling workflow is proposed to quantify the impact of climate change on water discharge without resorting to meteorological observations. The method combines asynchronous hydroclimatic modelling and quantile perturbation applied to streamflow observations. The results show that the proposed workflow produces useful and reliable hydrologic scenarios, which can predict seasonal mean flows similar to a conventional hydroclimatic modelling approach.
HYDROLOGY AND EARTH SYSTEM SCIENCES
(2023)
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)
Article
Geosciences, Multidisciplinary
Emixi Sthefany Valdez, Francois Anctil, Maria-Helena Ramos
Summary: This study investigates the interactions between a precipitation post-processor and other uncertainty quantification tools in a hydrometeorological forecasting chain. The results show that the post-processor significantly improves the quality of precipitation forecasts, but its effectiveness in improving hydrological forecasts depends on various factors such as the configuration of the forecasting system, forecast attribute, lead time, and catchment size. Therefore, the combined effect of the precipitation post-processor and other uncertainty quantification methods should be considered when designing or enhancing hydrometeorological ensemble forecasting systems.
HYDROLOGY AND EARTH SYSTEM SCIENCES
(2022)
Article
Geography, Physical
Georg Lackner, Florent Domine, Daniel F. Nadeau, Annie-Claude Parent, Francois Anctil, Matthieu Lafaysse, Marie Dumont
Summary: Arctic landscapes are covered in snow for at least 6 months a year, and the energy balance of the snow cover plays a key role in influencing various factors. The study aimed to quantify major heat fluxes above, within, and below a low-Arctic snowpack. Results showed that radiative losses are counterbalanced by sensible heat flux, with minimal latent heat flux. The model reproduced the observed energy balance well, but had deficiencies in simulating turbulent heat fluxes at an hourly timescale due to atmospheric stratification effects.