Dauphin River Research Articles

Ice roughness estimation via remotely piloted aircraft and photogrammetry

Abstract. The monitoring of fluvial ice covers can be time intensive, dangerous, and costly if detailed data are required. Ice covers on a river surface cause resistance to water flow, which increases upstream water levels. Ice with a higher degree of roughness causes increased flow resistance and therefore even higher upstream water levels. Aerial images collected via remotely piloted aircraft (RPA) were processed with structure from motion photogrammetry to create a digital elevation model (DEM) and then produce quantitative measurements of surface ice roughness. Images and surface ice roughness values were collected over 2 years on the Dauphin River in Manitoba, Canada. It was hypothesized that surface ice roughness would be indicative of subsurface ice roughness. This hypothesis was tested by comparing RPA-measured surface ice roughness values to those predicted by the Nezhikhovskiy equation, wherein subsurface ice roughness is proportional to ice thickness. Various statistical metrics were used to represent the roughness height of the DEMs. Strong trends were identified in the comparison of RPA-measured ice surface roughness to subsurface ice roughness values predicted by the Nezhikhovskiy equation, as well as with comparisons to ice thickness. The standard deviation and interquartile range of roughness heights were determined to be the most representative statistical metrics and several properties of the DEMs of fluvial ice covers were calculated and observed. No DEMs were found to be normally distributed. This first attempt at using RPA-derived measurements of surface ice roughness to estimate river ice flow resistance is shown to have considerable potential and will hopefully be verified and improved upon by subsequent measurements on a wide variety of rivers and ice covers.

IceMaskNet: River ice detection and characterization using deep learning algorithms applied to aerial photography

Dynamic ice processes can significantly affect various river characteristics such as hydraulics, sediment transport, water quality and morphology. River ice can also impede ship navigation and can induce flood hazard. Study of ice processes is thus crucial for understanding rivers in cold regions. These processes vary according to the four different phases of river ice development: formation, progression, recession and breakup. Monitoring and observation of river ice by remote sensing and close-range photogrammetry have recently attracted significant attention from river ice researchers, and the emergence of remotely piloted aircraft systems with onboard cameras has facilitated safe surveying of river ice. Despite all the developments in this field, fast and accurate data acquisition is still very demanding. One of the challenging tasks in data acquisition from aerial imagery is ice detection and classification. This study presents a novel algorithm called IceMaskNet for automatic river ice detection and characterization from aerial imagery. IceMaskNet utilizes an improved version of the Mask R-CNN, a novel Region-based Convolutional Neural Network with an additional mask. The presented deep learning algorithm is able to detect river ice from aerial imagery and characterize it as belonging to one of six different classes: broken ice, frazil slush, ice cover, open water, border ice, or frazil pan. The developed algorithm is tested using data collected on the Dauphin River, in Manitoba, Canada. Aerial photography from several sections of the river with various slopes and bend scales were used to train IceMaskNet to detect, classify and characterize river ice. The presented algorithm detected and classified river ice with average accuracies of 95% and 91%, respectively.

A detailed energy budget analysis of river supercooling and the importance of accurately quantifying net radiation to predict ice formation

AbstractRiver supercooling and ice formation is a regular occurrence throughout the winter in northern countries. The resulting frazil ice production can obstruct the flow through intakes along the river, causing major problems for hydropower and water treatment facilities, among others. Therefore, river ice modellers attempt to calculate the river energy budget and predict when supercooling will occur in order to anticipate and mitigate the effects of potential intake blockages. Despite this, very few energy budget studies have taken place during freeze‐up, and none have specifically analysed individual supercooling events. To improve our understanding of the freeze‐up energy budget detailed measurements of air temperature, relative humidity, barometric pressure, wind speed and direction, short‐ and longwave radiation, and water temperature were made on the Dauphin River in Manitoba. During the river freeze‐up period of late October to early November 2019, a total of six supercooling events were recorded. Analysis of the energy budget throughout the supercooling period revealed that the most significant heat source was net shortwave radiation, reaching up to 298 W/m2, while the most significant heat loss was net longwave radiation, accounting for losses of up to 135 W/m2. Longwave radiation was also the most significant heat flux overall during the individual supercooling events, accounting for up to 84% of the total heat flux irrespective of flux direction, highlighting the importance of properly quantifying this flux during energy budget calculations. Five different sensible (Qh) and latent (Qe) heat flux calculations were also compared, using the bulk aerodynamic method as the baseline. It was found that the Priestley and Taylor method most‐closely matched the bulk aerodynamic method on a daily timescale with an average offset of 8.5 W/m2 for Qh and 10.1 W/m2 for Qe, while a Dalton‐type equation provided by Webb and Zhang was the most similar on a sub‐daily timescale with average offsets of 20.0 and 14.7 W/m2 for Qh and Qe, respectively.

Morphology and blood metabolites reflect recent spatial and temporal differences among Lake Winnipeg walleye, Sander vitreus

The invasive rainbow smelt (Osmerus mordax) was an abundant food source for Lake Winnipeg walleye (Sander vitreus), especially in the north basin of the lake, until the smelt’s collapse in approximately 2013. We quantified changing length-at-age (≈ growth rates) and relative mass (≈ body condition) in Lake Winnipeg walleye caught for a gillnet index data set. Here, walleye showed smaller length-at-age, particularly young fish in the north basin, over time. This approach to assessing growth suggests a constraint in the north basin fish, possibly a nutritional limitation between 2017 and 2018, that was not present in the south. We then analyzed a separate group of walleye (≥452 mm in fork length) sampled in 2017 as part of a large-scale tracking study, which had a similar slope in length-mass relationship to large walleye caught in that year for the gillnet index data. A panel of metabolites in whole blood samples associated with amino acid metabolism and protein turnover was compared. These metabolites revealed elevated essential amino acids in fish caught in the Dauphin River, and suggest that protein degradation may be elevated in north basin walleye. Therefore, based on both growth estimates and metabolites associated with protein balance, we suggest there were spatially distinct separations affecting Lake Winnipeg walleye with decreased nutritional status of walleye in the north basin of Lake Winnipeg being of particular concern.

Effects of freeze-up consolidation event surges on river hydraulics and ice dynamics on the Lower Dauphin River

Ice cover formation in fast-moving rivers during freeze-up involves progression of the leading edge with intermittent collapses that cause rapid consolidation (mechanical thickening) of the ice cover and withdrawal of the leading edge downstream. The release of water from storage during these consolidation events causes a surge wave to travel downstream, resulting in an unsteady flow condition that briefly increases water levels and other concomitant hydraulic parameters that are difficult to measure directly such as discharge, velocity, and shear stress. In this study, the previously published Rising Limb Analysis Method (RLAM) was used to deduce the transient hydraulic properties at several monitoring sites on the Lower Dauphin River during the release of surge waves caused by freeze-up consolidation events in 2016–2017. It was found that parameters were most greatly amplified near the toe of the consolidated reach, and decayed in the downstream direction. On one occasion, computed discharge was found to be nearly three times the steady state discharge during the passage of a surge wave that caused water levels to rise by 0.93 m in less than 10 min at one gauge location. The calculated consolidation stress applied to the ice cover during this event increased by 99 Pa; the peak stress was approximately 2.5 times the unperturbed value. Results of the RLAM application highlight the interdependent relationship between ice dynamics and unsteady flow hydraulics during freeze-up, and the importance of considering the effects of consolidation events when predicting the final ice thickness in fast-moving rivers.

Field monitoring of secondary consolidation events and ice cover progression during freeze-up on the Lower Dauphin River, Manitoba

Dynamic freeze-up processes on fast moving rivers determine the ultimate thickness of the ice cover and the associated ice-affected water levels. Secondary consolidation events, the shoving and mechanical thickening of a progressing ice cover at freeze-up, can cause rapid water level rises that can flood land unexpectedly. Additionally, they can lead to over-thickening of an ice cover beyond that calculated with a steady discharge. The lack of field data and observations of secondary consolidations in the literature results in a poor understanding of the conditions that affect the timing and severity of these events. In this study, data pertinent to further the understanding of secondary consolidation events were collected over two freeze-up monitoring seasons on the Lower Dauphin River. The high spatial resolution of pressure transducers and trail cameras, along with first-hand observations of consolidation events, provide a dataset that was used to analyze the extent of several secondary consolidation events and their effect on water levels throughout the 12 km reach. The most significant event caused a water level increase of 2.5 m over a span of 8 min when approximately 7.5 km of ice cover consolidated to a length of 4.7 km. An unmanned aerial vehicle was used to obtain aerial imagery of the ice front, which advanced through a series of localized shoves as incoming ice packed against the stationary cover. Other data collected includes top-of-ice elevations, ice thickness estimations, and meteorological data. Collectively, the dataset provides a relatively complete account of two freeze-up seasons on the Lower Dauphin River, including the processes of ice cover advancement and consolidation that play an important role in the peak ice-affected water levels and flooding potential. The dataset can be used by ice researchers to assess the ability of current state-of-the-art numerical models in simulating dynamic freeze-up processes, and highlight areas for future research and model development.

Nutrient sequestration in the Lake Winnipeg watershed

Abstract Lake Winnipeg, Manitoba, has been subjected to significant increases in nutrient loading over the last few decades, and consequently has experienced significant and widespread algal blooms. The objective of our study was to identify sources of nutrients in the Lake Winnipeg basin, and quantify their removal (sequestration) into 28 of the largest lakes and reservoirs located in the Saskatchewan, Dauphin, Red, and Winnipeg river sub-basins, thus preventing their transport downstream to Lake Winnipeg. Discharges were determined daily, and nutrient parameter concentrations determined once or twice each month upstream and downstream from each of the lakes and reservoirs for three years. Concentrations of P and N in source waters of the Lake Winnipeg basin varied substantially, with the lowest concentrations occurring in pristine headwaters of the Saskatchewan River (mean TP = 14 μg/L; mean TN = 217 μg/L) and some of the highest concentrations occurring in small streams that originated within agricultural landscapes in the headwaters of the Dauphin River sub-basin (mean TP = 133 μg/L; mean TN = 1313 μg/L). Twelve reservoirs in the Saskatchewan River sub-basin collectively sequestered 92% of the TP inputs and 68% of the TN inputs to the sub-basin. In P-rich lakes, relatively more N was sequestered than P compared with nutrient impoverished lakes. A total 13,215 t/yr TP was discharged into Lake Winnipeg while 8234 t TP/yr, was sequestered into the lakes and reservoirs. The Red River sub-basin was the principal source of nutrients to Lake Winnipeg and should be the focus of nutrient management in the Lake Winnipeg basin.

Geospatial modelling to determine the behaviour of ice cover formation during freeze-up of the Dauphin River in Manitoba

During the formation of a frazil-generated ice cover, ice bridging can occur along sections of rivers that are geomorphologically suited (e.g. very sinuous, narrowing channel width, low gradient bed slope) to arrest the flow of ice pans. Bridging may occur at several locations along the river independently from one another with separate ice cover formations occurring simultaneously. Juxtaposition of incoming frazil ice and ice pans extending from the ice bridgings can cause backwater levels to rise upstream and cause flooding. The separate ice covers will eventually merge to completely cover the river sooner than if no additional bridging had occurred. This makes it difficult to predict the time required for a river to be completely ice covered and the locations and timing of related potential flood events. A geospatial model is introduced in this paper that applies a principal component analysis to cluster geomorphological features – such as sinuosity, channel width and channel slope – into typologies. Certain combinations of these typologies reveal predisposition of certain geomorphological features to ice bridging or non-bridging. The Dauphin River in Manitoba is used as a test case for the development of the model.

Evaluating the impact of fluvial geomorphology on river ice cover formation based on a global sensitivity analysis of a river ice model

In 2011, Manitoba was stricken by wide-scale flooding causing high flows along the Dauphin River. The unprecedented high discharges at freeze-up created the potential for excessive ice cover thickening and backwater staging exacerbating the flood risk already threatening the communities along the river and upstream-lying Lake St. Martin. Hence, the river ice model RIVICE was implemented to determine flood protection elevations to which existing dikes needed to be raised and extended. Two reaches of the river were modelled separately representing distinct geomorphological characteristics: mildly sloping, more sinuous upper reach and steeper, more channelized lower reach. A global sensitivity analysis based on a Monte Carlo analysis was carried out to determine how differences in these morphological features influence different processes of the ice cover formation. It was found that the sinuous and braided morphology of the upper reach has a marked impact on the sensitivity of the hydraulic roughness and strength parameters. The structure of the ice (porosity) and the discharge were most sensitive to the backwater level outcomes of the steeper and straighter lower reach.

Modelling ice cover formation of a lake–river system with exceptionally high flows (Lake St. Martin and Dauphin River, Manitoba)

Abstract In 2011, Manitoba experienced extreme flooding on a wide extended scale that affected many large watersheds, in particular the Upper Assiniboine, Qu'Appelle and Souris River watersheds. In order to protect Winnipeg from excessive flooding, some of the flood waters were diverted via the Portage Diversion into Lake Manitoba and Lake St. Martin, causing extreme high water levels and flooding of these lakes. The outlet of this entire drainage basin is the Dauphin River flowing from Lake St. Martin to Lake Winnipeg. The river's conveyance capacity was exceeded by the flood and the lakes could not be drained fast enough to avoid damage from high flood levels and shore line erosion and property damage from accentuated wind setup (seiche) and wave uprush (runup) of the lake water. An Emergency Channel was constructed to compensate for this constriction, which now conveys additional water from Lake St. Martin towards Lake Winnipeg. Due to the extreme flooding, the Dauphin River had unprecedented record flows in the autumn of 2011. Ice cover formation during winter freeze-up of these waters could potentially exacerbate the flooding situation for communities alongside Lake St. Martin and the Dauphin River. Hence, the river ice model RIVICE was implemented to determine flood protection elevations to which existing dikes needed to be raised and extended. The modelling takes into consideration the potential frazil ice generation that can raise water levels by up to 3 m from open water levels. High discharge and water level data was available to calibrate the model both for an open-water and an ice-covered case. Three scenarios were then simulated to determine ice-covered water levels of the river for early, best-estimate and late freeze-up events to determine flood protection levels along the Dauphin River and Lake St. Martin. The model was successfully validated using data collected during a freeze-up event in December 2011.

Nocturnal Foraging in the American White Pelican

Nocturnal foraging was examined in American White Pelicans (Pelecanus erythrorhynchos) at the Dauphin River, about 50 km from a breeding colony on Lake Winnipeg, Manitoba, Canada. From two to three times as many pelicans foraged at night as in the daytime, with foraging flocks being larger at night. In contrast, more pelicans were present at adjacent loafing sites during the day. Capture rates were highest for flocks of up to 100 foragers in the daytime, but showed no relationship to flock size at night. Rates of bill dipping and mean duration of dips were significantly greater at night, but capture rates were significantly lower. Lower capture rates at night were to some extent offset by the capture of more large fish, probably because they were more accessible at that time. Most of the day-night differences in capture efficiency appeared to be due to lower visual sensitivity of pelicans at night.

Subpopulation Status of Lake Whitefish (Coregonus clupeaformis) in Lake Winnipeg

Spawning lake whitefish in Lake Winnipeg from Dauphin River–Lake St. Martin, Traverse Bay, Berens River–Poplar River–Big Black River and Grand Rapids and in Little Playgreen Lake are accorded subpopulation status based on morphometric measurements, meristic counts, and differences in frequencies of alleles at the muscle glycerol-3-phosphate dehydrogenase (g-3-pdh-b) locus. The frequency of the g-3-pdh-b-3 allele in samples from Dauphin River–Lake St. Martin was 0.77 compared with 0.46 for whitefish from the remainder of Lake Winnipeg and Little Playgreen Lake. Two loci for isocitrate dehydrogenase and one lactate dehydrogenase locus are also polymorphic in these fish but no significant allele frequency differences were found at these loci among the samples of spawners compared above. Hatchery-reared lake whitefish, obtained from stock in Clearwater Lake and William Lake and planted as fry or eyed eggs in Lake Winnipeg at Dauphin River and Grand Rapids, do not appear to have made a detectable genetic contribution to the local whitefish stocks, based on differences in g-3-pdh-b and idh Bβ allele frequencies among parent stock and spawners captured near the release sites.

Artificial Feeding of Whitefish Fry

Through experiments conducted at the Winnipegosis and Dauphin River hatcheries in Manitoba, it has been determined that the artificial feeding of whitefish fry (Coregonus clupeaformis) is of value in those localities where the eggs hatch early in the spring and fry must otherwise be planted while the lakes are covered with ice. At that time the existence of an adequate supply of food organisms is uncertain. Various artificial foods were tried in the experiments but the most satisfactory results were obtained with (1) natural plankton (principally Daphnia pulex), and ground beef heart, and (2) natural plankton alone. During the hatching period three successive batches of fry may be held in tanks and fed for a 10-day period before they are released. Fry gain in length, weight and viability during the period of artificial feeding and are better fitted for survival in the natural environment. Through the delay in planting, the supply of available natural food and the temperature of the water in the lake environment are enabled to become more favorable by the time the fry are ready to be transferred from the tanks to the lake.