Search

This is a review of sea level data performed at three selected stations (Quebec-Lauzon, Harrington Harbour,and Halifax) in eastern Canada in order to investigate the seasonal trends and other long-term and short-term changes which occurred since the beginning of the 20th century. Stations situated in riverine or estuarine regions (e.g., Quebec-Lauzon) are significantly affected by freshwater flow in their annual cycle of sea level changes and exhibit a definite maximum in spring and minimum in autumn-winter. Other stations situated in the eastern half of the Gulf of St. Lawrence (e.g., Harrington Harbour) or near the open Atlantic coast (Halifax) mainly follow the general cycle of subarctic regions, with lows in spring-summer and highs in autumn-winter. Such seasonal variations appear to be related to the atmospheric pressure and baroclinic current variations. Secular trends in mean sea level in eastern Canadian waterbodies show a mean rise of about 2.56 mm/yr -1 due to tectonic motions, that is, land sub...

There has been extensive loss of life and damage associated with severe tropical cyclones in the South Asian region. The Bay of Bengal and the Arabian Sea are the two marginal seas of the North Indian Ocean and have given rise to storm surges during the pre-monsoon (April-May) and the post- monsoon period (October-December). Table 1 lists the events where the loss of lives were 1000 or more.

Interannual variations of sea level along the Bangladesh coast are quite pronounced and often dominate the long-term sea level trends that are taking place. The El Nino/Southern Oscillation (ENSO) induced variation is an important component of interannual mode of variations. The present article deals with the relationship between the sea level variations along the Bangladesh coast and the Southern Oscillation phenomenon. The mean tide level data of monsoon season (June to September) pertaining to Hiron Point (in Sundarbans) and Char Changa (on the mouth of Meghna River) have been analyzed and correlated to the Southern Oscillation Index (SOI). The annual variation of mean tide level in the coastal areas of Bangladesh reveals that the tide level reaches its peak during the monsoon season. The maximum tide level during the calendar year is recorded in August. Thus, it is not surprising that the inundation of the coastal belt of Bangladesh due to the floods is most common during the summer monsoon season, es...

Hourly sea level records from three stations in eastern Canada (Quebec-Lauzon, Harrington-Harbour and Halifax) are analyzed both in frequency domain from 1970 to 1979 and in time domain during 1973. At the three stations, the deterministic model explains 90 to 96% of the total variability of sea level. The semidiurnal and diurnal tides contribute largely to its variations. The residual series, less than 10% of the initial variations of sea level, contain irregular values including extreme values of seiches and storm surges. Such random variations are analyzed and modeled following the method described by Box and Jenkins (1976). The long period variations (2 to 30 days) can be attributed to meteorological forcing (atmospheric pressure and winds). The short period variations (some hours to one day) can be attributed to longitudinal seiches, semidiurnal and diurnal atmospheric tides, and inertial oscillations. The water discharge from the St. Lawrence River contributes 29% of the monthly residual sea level a...

Tropical Cyclones which develop in the Andaman Sea and the Bay of Bengal during the inter monsoon months (April-May, October-December), move either westwards affecting the east coast of India or recurve to the north or northeast and eventually cross the coast of Bangladesh or Myanmar. Extensive damage is caused to the life and property by the storm surge as much of the coastal land around the Bay of Bengal is densely populated. The damage caused by a cyclone induced surge depends to a considerable extent on whether the surge peaks at or close to high tide. The main purpose of the present study was to develop a combined time-surge model for the off-shore waters in the Meghna estuary. It seems clear that the strong, predominantly southward current measured at Site A, south of Sandwip Island, has substantial magnifying and delaying effect on tidal elevation and current. But the areal extent of this modification of the tide is unknown at present. Further, it is impossible to say whether the fast southward current forms a narrow jet or a broad current many kilometers wide, but it is important to know which is the case before the effect can be modelled satisfactorily.

Trends in monthly precipitation, temperature, water equivalent of the snowpack, and streamflow are investigated for stations in three local areas in interior British Columbia of interest for hatching of salmon on the Adams River, Fraser River near Prince George, and Skeena River near Babine Lake. In general, since the 1960s, cold season (October and January) precipitation has been decreasing, temperatures have been rising, snow water equivalents have been diminishing, and streamflows have been going down. However, the results may not necessarily be part of a global warming but may rather be an amplification of the common Pacific/North American pattern.

The International Tsunami‐93 Symposium was sponsored jointly by the Tsunami Commission of the International Union of Geodesy and Geophysics and the International Coordination Group of the Intergovernmental Oceanographic Commission of UNESCO. It was organized by the Japan Society of Civil Engineers and was held in Wakayama, Japan, August 23–25, 1993. This review is based on the proceedings of the Conference, which were jointly edited by Professors Yoshito Tsuchiya and Nobuo Shuto. In total about 142 participants from 19 countries attended the symposium. They presented some 65 papers in 12 sessions dealing with various topics. Selected papers from the conference are briefly reviewed and highlights of the conference are noted.

A vertically integrated two‐dimensional (2‐D) and a five‐layer three‐dimensional (3‐D) numerical models were developed to compute the tides in the Gulf of Suez, Egypt. The computational grid used to schematize the Gulf has a horizontal resolution of 3 × 3 km and consists of a lattice of 23 × 100 points in the 2‐D model and five such lattices in the 3‐D model. Both 2‐D and 3‐D simulations clearly revealed the Kelvin wave nature of the tide with partial reflection. The M2 simulations also showed a strong tidal signature in the southern sector as compared to the northern part. For the 3‐D simulations, the horizontal and vertical eddy viscosity coefficients and the coefficient of bottom friction were respectively set to 3 × 106 cm2/s, 200 cm2/s, and 0.001. The tidal range decreases from the entrance of the Gulf of Suez toward the Bank of Tor where it reaches a small value and then increases again to about 1.5 m at Suez. A difference of six hours exists between the times of high water at the southern and north...

A two‐dimensional time‐dependent nonlinear tidal model was developed for the St. Lawrence estuary, with two open boundaries. This model is a substantial improvement over two earlier models which did not include the nonlinear terms. The model is run simultaneously, to hindcast the tide on any particular day, using as boundary conditions real data for the five important constituents, namely M2, S2, N2, K1, and O1. Coamplitude and cophase diagrams were constructed and compared with known shore‐based gauge observation. This comparison revealed that the model performs quite satisfactory. Further calibration was then carried out by adjusting the bottom friction coefficient. Such a calibration is necessary before the model can be used to simulate tsunami generation and propagation in the estuary. The total water level during the occurrence of a tsunami should include not only the tsunami amplitude, but also the interaction of the tsunami with the tide.

An earthquake occurred off the west coast of Nicaragua at 18 h 16 m on September 1, 1992. Despite a relatively modest magnitude (Ms = 7.0 Mw = 7.6), the earthquake generated a significant tsunami, which strongly impacted the coast of Nicaragua and was felt at several remote stations. In an attempt to better understand the tsunami propagation and coastal impact, we conducted a field survey on November 2–7, 1992, i.e., roughly 2 months after the event. We emphasized the observation of run‐up and inundation, but also collected a limited number of sand samples and mud cores. A total of 12 sites were surveyed, with multiple stations per site. This article presents the results of the survey, including a detailed discussion of the observation methods and data processing strategies. Observed run‐ups (relative to MSL) were typically in the 4–7 m range, with a maximum run‐up of 9.1 m at San Martin, an open coast site with a steep beach slope. Inundation extents were relatively modest, typically of the orde...

Numerical models of the hydrodynamics of Howe Sound and of the transport and deposition of suspended sediment from the Squamish River entering Howe Sound are described. Two different versions of a three‐dimensional hydrodynamic model of Howe Sound have been developed. One is a fine‐grid model employing a 390‐m grid size, the other a coarse‐grid model with a 1.95‐km grid size. In the present implementation, both models have 9 layers in the vertical, and calculate the three components of the velocity field as well as the time‐varying density field. The models are forced at their open boundaries by a larger three‐dimensional model of the Strait of Georgia, by surface wind stress and by the Squamish River flow. The coarse‐grid model is computationally less demanding and is used for long‐term simulations, whereas the fine‐grid model is used for detailed simulations for shorter time periods. The sediment model derives its velocity field from the coarse grid three‐dimensional hydrodynamic model. Sedimen...

Sea level records from nine stations in eastern Canada have been analyzed in order to investigate the seasonal trends and other long‐term changes that have occurred since the beginning of this century. The four stations situated in riverine or estuarine regions are significantly affected by freshwater flow in their annual cycle of sea level changes and exhibit a definite maximum in spring and minimum in autumnwinter. Other stations situated in the eastern half of the Gulf of St. Lawrence or near the open Atlantic coast follow the general cycle of subarctic regions, with a low in spring‐summer and a high in autumn‐winter. Such seasonal variations appear to be related to the atmospheric pressure and baroclinic current variations. Secular trends in mean sea level show a mean rise of about 25.6 cm/century along the Atlantic coast of Canada. The monthly mean sea level series are subjected to Fourier spectral analysis in an attempt to identify the major long‐period cycles present in the records. The sp...

A vertically integrated tidal model of Barkley Sound has been developed as the first phase of a study of circulation and its interaction with juvenile salmon moving through Alberni Inlet and Barkley Sound to the Pacific Ocean. This model is an implementation of the three‐dimensional baroclinic model GF8 using a single layer, with a horizontal grid size of 400 m. Elevations and velocity boundary condition data for this model were obtained from an irregular grid finite element model of the west coast of Vancouver Island. The resulting barotropic tidal model was found to be stable, and produced tidal elevations in close agreement with observations. Modeled tidal currents at the location of a current meter morred in Imperial Eagle Channel agreed well with observations, to the extent that barotropic currents could be deduced in a simple manner from observations containing both barotropic and baroclinic modes. The next phase in the modeling program was the development of an upper layer baroclinic model...

Canada has increased the number of tsunami warning stations on the Pacific Coast from two to three. The last gauge was installed at the north end of Vancouver Island, thereby filling a large gap previously existing and providing full coverage along the coast. The record of gauges at two of the three locations is accessible either by telephone or by means of meteor burst communication, alleviating the difficulties experienced during the tsunami threat of May 6, 1986, when telephone communications were disrupted by heavy use. The gauge at Langara Island will be relocated in a more accessible and also a more tsunami‐responsive location at Rennell Sound in the Queen Charlotte Islands. All tsunami gauges also serve as tide gauges, recording the water level every 15 min. In the event of a tsunami, the recording interval can be altered to every 60 s. Suggestions have been made that Canada attempt deep‐sea recording of tsunamis off its Pacific Coast. Although this would be of great scientific value, no s...

Some past tsunamis in various regions of Canada have been reviewed. These tsunamis have occurred from a variety of sources: under-ocean earthquakes, sub-marine landslides and human-made inadvertent large chemical explosions. Resonance amplification of the tsunami has been identified as important in the Alberni Inlet on the Vancouver Island, British Columbia, and also in the Burin Inlet in Newfoundland. The results of the various numerical models for tsunami generation, propagation onto the shelf, and into the coastal inlets that formed the scientific basis for the British Columbia tsunami warning system have been briefly discussed.

The teleconnections between the occurrence of El-Nino events and the generation of tropical cyclones in various ocean basins have been considered by several authors. Here we present data linking storm surges in the Bay of Bengal with El-Nino. In fact this preliminary study shows that there is no correlation between El-Nino occurrences and Storm Surge Events in the Bay of Bengal.

The Gulf war of 1991 resulted in some of the worst oil spills in the Arabian Gulf. While the amount of oil spilled was one of the largest in the world and the damage extensive, the desalination plants near Al-Jubail on the Saudi Arabian coast were not affected by oil contamination. Contamination of these plants would have threatened the water supply not only to the Saudi population but also to the military personnel that were involved in fighting the war. The amount of oil reaching this area of the Arabian Gulf was rather minimal compared to the total quantity of oil that was actaully spilled into the Gulf waters.

Global warming from increasing atmospheric concentrations of greenhouse gases could well cause global sea level to rise. As discussed in Chapter 2 of this book, it is the firm consensus of the scientific community that, in the decades to come, the increasing atmospheric concentrations of greenhouse gases should cause the world to warm. Global warming, in turn, should cause sea level to rise by expanding the oceans and increasing the melt rates of ice sheets and glaciers. Global climate change could also affect the frequency and intensity of cyclones and storm surges, and thus coastal flooding, in the Bay of Bengal.

One of the best examples of tsunami amplification due to resonance is at Port Alberni which is located at the head of Alberni Inlet on Vancouver Island on the Pacific Coast of Canada. During the 1964 Alaska earthquake tsunami, the tsunami amplitude at Port Alberni was two to three times greater than the tsunami amplitude at Tofino, which is located close to the open coast. This remarkable amplification of the tsunami as it travelled a distance of about 65 km through the Alberni Inlet can be explained by the fact that periods of the normal modes of the Alberni Inlet system agree closely with the dominant periods in the tsunami wave form from the Alaska earthquake tsunami as it reached the continental shelf off Vancouver Island.

While there are several excellent books dealing with numerical analysis and analytical theory, students and faculty in numerical applications to ocean dynamics have to sift through hundreds of references. This monograph is an attempt to partly rectify this situation. Major chapters (II, III and IV) deal first with the basics and then go on to various applications. Instead of covering the vast field of ocean dynamics, this book focuses on transport equations (diffusion and advection), shallow water phenomena - tides, storm surges and tsunamis; three-dimensional time dependent oceanic motion; natural oscillations; and steady state phenomena. The aim of this book is two-fold; it gives an introduction to the application of finite-difference methods to ocean dynamics, and it also reviews more complex methods.

This paper is an attempt to synthesize current knowledge about tides and tidal modelling efforts in the St. Lawrence Estuary. A sequence of numerical models, increasing in complexity from simple one-dimensional to baroclinic, is discussed. The performance of these models and their limitations are pointed out. In the last section, some recommendations are given for future work to improve our present knowledge of the tidal properties over the estuary. These include complementary in situ observations of offshore tidal elevations and principal tidal currents, in addition to new numerical modelling experiments combining all the positive attributes of the previous models.

One of the aspects of the seaward movement of fresh water from the Fraser River is its passage tangential to Roberts Bank and subsequent recirculation onto the bank. A synoptic study of the oceanography of the bank is described, part, of a larger study of surface currents in the Strait of Georgia. Seasonal changes in salinity and temperature over Roberts Bank are presented. The freshwater flows and motions of the salt wedges in the river and on the bank are described. Under the restraints of a retaining wall immediately north of the river, and on the incoming tide, a shallow layer of fresh water briefly forms an eddy, varying in size with river discharge and extending over the bank to the south. The influence of the two-layer flow on sedimentation is also discussed.

Maximum tsunami amplitudes that will result from major earthquakes on the outer coast of the Pacific Northwest region of North America were considered in Part 1 (Marine Geodesy 13(2):101–146). Here the tsunami threat for the inner coast is considered by developing a numerical model for the system consisting of the Strait of Georgia, Juan de Fuca Strait, and Puget Sound (GFP model). Three different source areas were considered for the outer coast models, and the resulting tsunami was propagated to the entrance of Juan de Fuca Strait. Using the output from the outer models, the GFP model was run. The results showed that whereas large tsunami amplitudes can occur on the outer coast, inside the GFP system, unless the earthquake occurs in the system itself, no major tsunami will result.

The semidiurnal L2 tide, in a model simulation of mixed tides in the straits of Georgia and Juan de Fuca, is the only tidal constituent with a nondegenerate amphidromic point in the system.

Tsunami research has suffered considerably due to the lack of an objective magnitude scale. The Imamura‐Iida Grade Scale, although it has been used until now as a magnitude scale, is not o...

The tsunami travel‐time charts that are presently in use by the Tsunami Warning Center were constructed originally in 1948 based on the hydrographic data available in the mid 1940s. Even the revised charts of 1971 made use of essentially the same data. It is shown here that the travel times deduced by these charts could be in error by as much as two hours in some cases. Even worse, the compiled travel times as deduced from these charts are generally greater than the observed travel times, which is a dangerous situation from a tsunami warning point of view.

A time-dependent two-dimensional numerical model was developed to study the generation and propagation of a tsunami resulting from a major earthquake predicted to occur in the Shumagin Seismic Gap area of the eastern part of the Aleutian Island chain. The grid size is approximately 28 km in the latitudinal direction and 36 km in the longitudinal direction. The leading wave of the tsunami took only about three hours to arrive at the southern British Columbia coast and the northern part of the coast of the state of Washington. On the other hand, it took about eight hours to propagate into the Bering Sea. Two important results emerged from this study. The tsunami waves show very strong directionality, and the presence of the so-called lateral waves can be inferred. In the deep ocean, tsunami amplitudes up to one meter appear to occur. To deduce the detailed distribution of the tsunami amplitudes at the coast, the present large area model has to be coupled to a coastal model which uses either finer rectangular grids or irregular triangular grids.

Maximum tsunami amplitudes that will result from major earthquakes in the Pacific Northwest region of North America are considered. The modeled region encompasses the coastlines of British Columbia in Canada, and Washington and Oregon in the United States. Three separate models were developed for the outer coast and one model for the system consisting of the Strait of Georgia, Juan de Fuca Strait, and Puget Sound (GFP model) (Part 2). Three different source areas were considered for the outer coast models and the resulting tsunami was propagated to the entrance of Juan de Fuca Strait. Using the output from the other models, the GFP model was run. The results showed that large tsunami amplitudes can occur on the outer coast, whereas inside the GFP system, unless the earthquake occurs in the system itself, no major tsunami will result (Part 2).

Tidal flow patterns computed at various stages of the tide (high tide, falling tide, ebb tide, low tide, and flooding tide) and at different chronological stages (9000, 6000, 4500, 3000, and 2300 years B.P. and the present) using tidal numerical models were presented and interpreted in relation to sedimentary processes and the growth of the Fraser River delta during the past 9000 years. The overall tidal flows are least affected by the deltaic growth at high tide than at any other state of the tide. The net tidal flow with the overall northerly component has a sedimentary effect, in that there is an asymmetry in the transportation of silt and clay with more of these fine materials moved to the north of the delta than to the south. There is also independent seismic evidence for the existence of a tidal channel some 2000 B.P. between Point Roberts and the Fraser delta. This study attempted to link closely the sedimentary structure of the deltaic growth of the Fraser River with the physical oceanogr...

The propagation of tides in the Fraser River is discussed, with particular attention to the nonlinear interactions that characterize tides in such a system. The tidal regime is reviewed, and then the distortion of the tidal wave as it propagates upstream is discussed. The seasonal variation in tidal constituents can be related to the seasonally varying magnitude of the Fraser River discharge. The distortion in the tidal wave is related to the production of overtides through the nonlinear advective and friction terms. By considering the astronomical constituents of the tide, expressions are derived for the production of such overtides as the M4 and are quantitatively related to the relative magnitudes of tidal and mean velocity contributions.

We comment on the statements that amphidromes in marginal seas with cooscillating tides in the Northern Hemisphere have anticlockwise rotation, of cotidal lines, and that rotation in the Southern Hemisphere will be clockwise.

A computer program has been developed to draw paleo‐continental maps for the period 190 million years ago to the present. The user has the option of specifying the projection and the scale. Maps can be drawn for the globe as a whole, for the northern hemisphere alone, and for the southern hemisphere alone, or indeed for any specified part of the globe. These paleo‐geographic maps are useful for the computation of paleo‐tides, paleo‐ocean circulation studies and climate.

Updated maps showing the age of the diurnal tide in the Atlantic, Pacific, Indian, and Arctic oceans are given. In addition diagrams depicting the O1/K1 amplitude ratios for the world oceans are included.

The delta of the Fraser River has been growing for the last 9000 years into the strongly tidal strait of Georgia (5 m tidal range), an estuary between Vancouver Island and the mainland of the province of British Columbia. As the delta grew, it had an impact on the tidal flows adjacent to its shores. The objective of this research was to model tidal flows at different stages of the delta's growth during the late Quaternary. This chronological analysis of tidal flows and delta growth was possible because of the number of published radiocarbon ages as well as the detailed knowledge of the surface geomorphology and subsurface geology of this delta that is now available. Four major units form the lithostratigraphic sequence of the delta: Unit 1 is a silt with thin sand beds formed in a delta‐front environment; Unit 2 is a sand with increasing amounts of silt in the upper part of the unit, formed in the lower tidal flats; Unit 3 is an organic rich silt containing peat interbeds and the Mazama ash, form...

Possible tsunamis in the Pacific Ocean, especially in its northeastern part, are discussed in relation to a predicted major earthquake in the Shumagin Seismic Gap (located in the eastern part of the Aleutian Island Chain) and to a major eruption of the St. Augustine volcano in Cook Inlet, Alaska. The deep-water propagation of the tsunami generated in the Shumagin Gap is simulated through the use of a spherical polar coordinate grid of the approximate size of 14km. The tsunami generated by the St. Augustine volcano is studied through the fine mesh grid confined to the Cook Inlet only. The numerical models were calibrated against historical tsunami data. The properties of the tsunami signal are described by the maximum amplitude which occurs in the tsunami record. This allows us to single out the direction along which a maximum tsunami is to be expected.

Continuous and long sea level records from four stations along the Canadian Atlantic coast have been examined and analyzed to investigate the seasonal and long‐term changes. Seasonal sea level variations at stations near the open ocean (Halifax, Nova Scotia, and Port‐aux‐Basques and St. John's, Newfoundland) follow mainly the general cycle for subarctic regions, with a low in spring and summer and a high in autumn and winter. Atmospheric pressure, water transport, and the changing density of seawater are found to be the major factors affecting the mean sea level at these stations. The most pronounced effects of river runoff and tides are found at Saint John, New Brunswick, in the Bay of Fundy, with a spring maximum and an

This article compares observed and computed elevations for the principal semidiurnal and diurnal tides, M2 and K1 in the Northwest Passage, the main east‐west channel through the Canadian archipelago. The official geographical name for this waterway is Parry Channel, although certain stretches of the passage are given individual names as shown in Figure 1. In fact, the western end of the passage, M'Clure Strait, is normally unnavigable due to ice, and the Peel Sound‐Franklin Strait‐Larsen Sound system is more important for access to the Beaufort Sea. Parry Channel is generally well over 200 m deep except in Barrow Strait where depths range from 130 to 170 m. Most of the side channels are also fairly deep. In the area discussed here, the water depths are sufficiently great to eliminate the likelihood of variations in ice cover having any significant effect on the phases and amplitudes of the barotropic tide.

The paleo‐tidal flows in the waters in the vicinity of Point Roberts and the Fraser River delta were numerically simulated through the development of a coarse grid model (2 km mesh size) and a fine grid model (2/3 km mesh size). The basic hydrodynamic equations, the numerical grid layout, and the finite‐difference forms are given. Special attention was given to transferring the boundary grid data generated from coarse grid model to the boundaries of the fine grid model. Diagrams are presented to show that the fine grid model yields much superior results in the sense of providing a greater amount of detail, in the current patterns and the eddy structure, than is provided by the coarse grid model, in the application to estimating sedimentary patterns.

The data of Reed (1983) are analysed to produce the following empirical equations for the amplitude p0 (overall fluctuation) in Pascals of the air pressure wave associated with a volcanic eruption of volume V km3 or a nuclear explosion of strength M Mt: Here s is the distance from the source in km. $$\begin{gathered} \log _{10} p_0 = 4.44 + \log _{10} V - 0.84\log _{10} s \hfill \\ {\text{ }} = 3.44 + \log _{10} M - 0.84\log _{10} s. \hfill \\ \end{gathered} $$

Hydrodynamical techniques were used to investigate the movement of oil slicks in the Strait of Georgia by assuming that the oil moves with the current and not with the wind. The divergence of the isobaths near Alden Bank probably makes the current branch while that southeast of Pt. Roberts narrows and intensifies the current. The effect of mutual interaction among several slicks both with and without a steering current such as a tidal or wind-generated current is found to be important in determining their trajectories. For oils with large viscosities, an oscillatory tidal current could give rise to an unstable situation in which the oil spreads in the form of long lines radiating out from the spill similar to the spokes in a bicycle wheel. Stratification of the water causes the oil plume to bifurcate and due to this phenomena, even if one branch of the plume carrying the oil goes away from the coast, the second branch carries oil to the coast.

Negative storm surges in eastern Canadian water bodies are studied. These water bodies include the Atlantic Coast of Canada, the Labrador Sea Coast, Bay of Fundy, Gulf of St. Lawrence and St. Lawrence estuary. Using hourly observed and predicted sea level variations for a period of eleven years (1956—1975), statistical analyses were made to examine the extreme values, the return periods for negative surges of various amplitudes, the influence of an ice layer and the interaction between tides and surges. The analysis for the Labrador coast, Bay of Fundy and Gulf of St. Lawrence shows that negative surges with amplitudes up to 1.7 m have occurred during the 11-yr period of study at Pointe-du-Chene. The negative surges generally are greater where strong and abrupt water depth changes occur. Largest negative surges occurred in winter, and smallest negative surges occurred in summer. This is due in part, to the presence of ice cover and, in part, to the fact that the weather systems that generate these surges are much stronger in winter than in summer. In these regions, the return periods of the extreme surges vary between 1 and 11 yr, but usually are less than 5 yr for most of the stations. Extreme values for negative surges in the St. Lawrence estuary are, in general, greater than for the other regions considered, with maximum value of 2.2 m observed at Quebec. All other locations in the upper estuary have amplitudes with value of 1.4 m between 1.8 and 1.9 m. The smallest amplitude negative surges occurred at Pointe-au-Pere and Baie- Comeau. In terms of frequency, the negative surges are most prominent at St-FranCois and least prominent at Ste-Anne-des-Monts. Based on the amplitudes and frequencies for both positive and negative surges, we conclude that stations located at the lower part of the estuary (Pointe-au-Pere, Baie-Comeau, Riviere-du- Loup) are reasonably safe from storm surges, while Quebec City and St-FranCois are extremely prone to storm surges. However, one reassuring feature at Quebec City and St-FranCois is that positive surges occur mostly at the time of low tide while negative surges occur at the time of high tide. On the other hand, at St- Jean-Port-Joli and St-Joseph-de-la-Rive positive surges occur at the time of high tide, which is a factor that makes the overall water level variations at these locations quite significant. Similar behaviour of the tidal wave propagation in the estuary and that of storm surge amplitudes is demonstrated.

During an eruption of the Alaskan volcano Mount St. Augustine in the spring of 1986, there was concern about the possibility that a tsunami might be generated by the collapse of a portion of the volcano into the shallow water of Cook Inlet. A similar edifice collapse of the volcano and ensuing sea wave occurred during an eruption in 1883. Other sea waves resulting in great loss of life and property have been generated by the eruption of coastal volcanos around the world. Although Mount St. Augustine remained intact during this eruptive cycle, a possible recurrence of the 1883 events spurred a numerical simulation of the 1883 sea wave. This simulation, which yielded a forecast of potential wave heights and travel times, was based on a method that could be applied generally to other coastal volcanos.

Frequent transverse currents have been observed in the middle of the lower St. Lawrence estuary, which are directed from the south towards the north shore with a mean residual velocity of 0.6 knots in the upper layer. The explanation that is invoked for the existence of these currents follows along the lines of the classical Rossby Adjustment Problem. The mutual adjustment of pressure and velocity fields in a fluid to which momentum has been imparted (for example by winds) takes place in a few hours in the homogeneous case but occupies several days (7 to 10) in the stratified case. Since the weather systems usually travel over the St. Lawrence estuary with roughly the same frequency (as the duration of the adjustment in the baroclinic situation) and impart momentum, continuity and frequent occurrence are achieved by the transverse currents.