Search

The lower Athabasca River (Canada) has experienced notable declines in streamflow and increasing oil sands development since the 1970s. This study investigates the potential impacts of climate change on navigability using both observed historical and projected future flows derived via hydrological simulations driven by an ensemble of statistically downscaled general circulation model climate data. Our use of proposed indices that form the Aboriginal Navigation Index (ANI) and a new index based on percentage over threshold (POT) occurrences yielded novel insights into anticipated changes to the flow regime. Comparisons of near (2041–2070) and far (2071–2100) future periods with the historical baseline (1981–2010) yielded results that project significant reductions in the 500 m3 s−1 POT during the fall navigability period spanning weeks 34 to 43, as well as reductions in the integrated ANIFall. These results indicate that challenging navigational conditions may become more frequent in the second half of the 21st century, not only during this fall period but also earlier into the summer, due to a shift in the flow regime, with potentially severe impacts on the users of the river channels. Our assessment approach is transferable to other regional study areas and should be considered in water management and environmental flow frameworks. [ABSTRACT FROM AUTHOR]

Recent developments in the field of computational intelligence (CI) techniques are helping to solve various problems of water resources modeling and management. These techniques have also shown their potential as an alternative approach to conventional controllers. In this paper, artificial neural networks (ANN) and fuzzy systems (FS) are shown to be efficient alternatives to using optimal control algorithms in real-time control of the polder water system of Overwaard in The Netherlands. The relation between the optimal decision or action and the influencing parameters are learned by ANN and FS and then used to derive the decisions and control actions in real-lime. It was possible to reproduce the centralized behavior (in terms of water levels and corresponding discharges) of optimal control action by using easily measurable local information. Moreover, it is demonstrated that model simulation with external intelligent controllers is ten times faster than that with the optimal control. DOI: 10.1061/(ASCE)0733-9496(2005)131:2(135) CE Database subject headings: Water management: Neural networks: Fuzzy sets; Computation: Netherlands.

River ice is a common occurrence in cold climate hydrological systems. The annual cycle of river ice formation, growth, decay and clearance can include low flows and ice jams, as well as mid-winter and spring break-up events. Reports and associated data on river ice occurrence are often limited to site and season-specific studies. Within Canada, the National Hydrometric Program (NHP) operates a network of gauging stations with water level as the primary measured variable to derive discharge. In the late 1990s, the Water Science and Technology Directorate of Environment and Climate Change Canada initiated a long-term effort to compile, archive and extract river ice related information from NHP hydrometric records. This data article describes the original research data set produced by this near 20-year effort: the Canadian River Ice Database (CRID). The CRID holds almost 73,000 variables from a network of 196 NHP stations throughout Canada that were in operation within the period 1894 to 2015. Over 100,000 paper and digital files were reviewed representing 10,378 station-years of active operation. The task of compiling this database involved manual extraction and input of more than 460,000 data entries on water level, discharge, date, time and data quality rating. Guidelines on the data extraction, rating procedure and challenges are provided. At each location, a time series of up to 15 variables specific to the occurrence of freeze-up and winter-low events, mid-winter break-up, ice thickness, spring break-up and maximum open-water level were compiled. This database follows up on several earlier efforts to compile information on river ice, which are summarized herein, and expands the scope and detail for use in Canadian river ice research and applications. Following the Government of Canada Open Data initiative, this original river ice data set is available at: https://doi.org/10.18164/c21e1852-ba8e-44af-bc13-48eeedfcf2f4 (de Rham et al., 2020). [ABSTRACT FROM AUTHOR]

The rapid advance in information processing systems in recent decades had directed engineering research towards the development of intelligent systems that can evolve models of natural phenomena automatically--'by themselves,' so to speak. In this respect, a wide range of machine learning techniques like decision trees, artificial neural networks (ANNs), Bayesian methods, fuzzy-rule based systems, and evolutionary algorithms have been successfully applied to model different civil engineering systems. In this study, the possibility of using yet another machine learning paradigm that is firmly based on the theory of statistical learning, namely that of the support vector machine (SVM), is investigated. An interesting property of this approach is that it is an approximate implementation of a structural risk minimization (SRM) induction principle that aims at minimizing a bound on the generalization error of a model, rather than minimizing only the mean square error over the data set. In this paper, the basic ideas underlying statistical learning theory and SVM are reviewed, and the potential of the SVM for feature classification and multiple regression (modeling) problems is demonstrated by applying the method to two different cases of model induction from empirical data. The relative performance of the SVM is then analyzed by comparing its results with that of ANNs on the same data sets.