This study investigates the capability of sequence-to-sequence machine learning (ML) architectures in an effort to develop streamflow forecasting tools for Canadian watersheds. Such tools are useful to inform local and region-specific water management and flood forecasting related activities. Two powerful deep-learning variants of the Recurrent Neural Network were investigated, namely the standard and attention-based encoder-decoder long short-term memory (LSTM) models. Both models were forced with past hydro-meteorological states and daily meteorological data with a look-back time window of several days. These models were tested for 10 different watersheds from the Ottawa River watershed, located within the Great Lakes Saint-Lawrence region of Canada, an economic powerhouse of the country. The results of training and testing phases suggest that both models are able to simulate overall hydrograph patterns well when compared to observational records. Between the two models, the attention model significantly outperforms the standard model in all watersheds, suggesting the importance and usefulness of the attention mechanism in ML architectures, not well explored for hydrological applications. The mean performance accuracy of the attention model on unseen data, when assessed in terms of mean Nash–Sutcliffe Efficiency and Kling-Gupta Efficiency is, respectively, found to be 0.985 and 0.954 for these watersheds. Streamflow forecasts with lead times of up to 5 days with the attention model demonstrate overall skillful performance with well above the benchmark accuracy of 70%. The results of the study suggest that the encoder–decoder LSTM, with attention mechanism, is a powerful modelling choice for developing streamflow forecasting systems for Canadian watersheds.
Gradients of conductivity and major ions in the coastal zone of the eastern Georgian Bay of Lake Huron appear to limit the spatial distribution of invasive dreissenid mussels. Rivers flowing into Georgian Bay from the Canadian Shield have relatively low conductivity compared to the main body of Lake Huron, which creates a gradient of solutes near the river mouths. Field observations show a strong positive correlation between conductivity and calcium concentration. Thus, we use conductivity to infer the calcium concentrations required for the successful growth of dreissenid mussels. Most dreissenid mussels were observed in regions where specific conductivities were greater than 140 μS/cm. Field observations were used to examine how the calcium poor river water mixes within the coastal zone, resulting in solute gradients that determine mussel distribution. When river flows are low in late summer, there is only a weak solute gradient across the coastal zone, implying an intrusion of open bay waters into the shallow embayments, that favor the growth of dreissenid mussels. In contrast, during spring when river flows are as much as 10 times higher, there is a strong solute gradient that extends further into the lake, and the low calcium appears to limit the growth of dreissenid mussels. Thus, the seasonal character of solute gradients helps describe the spatial distribution of dreissenid mussels and explains the localized absence of a species that is otherwise prevalent in much of the Laurentian Great Lakes.
The Sweepstakes, in Fathom Five National Marine Park, is Ontario’s most iconic shipwreck with over 100,000 visitors each summer. Continued exposure to water currents has directly and indirectly affected the integrity of the wreck and resulted in management interventions including efforts to stabilize the wreck and control vessel activity (both duration and speed). Despite these efforts, a scour ring is present in the sediment around the Sweepstakes, raising concerns regarding the prolonged stability of the wreck. An extensive series of field measurements were made during the summer of 2015 with the aim of differentiating between natural hydrological processes present at this site and human-derived water movements during the summer visitor season. There is a high-degree of natural current variability from processes as diverse as wind-induced surface gravity waves, internal gravity waves, and diurnal flows due to differential heating. Our results show that summer circulation driven by internal gravity waves derived from upwelling, surface waves, and differential heating was insignificant with respect to sediment resuspension and thus unlikely to produce the observed scour around the shipwreck. Scour is most likely caused by energetic winter storms, which should be a focus of future studies. While vessel induced currents were detectable at the shipwreck, they were no larger than the normal summer hydrodynamic variability, thus suggesting that management efforts continue to protect the site generally.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.