Oceans cover nearly three quarters of planet Earth. Our lives are intimately linked to them and to their tides in diverse ways, most of which we poorly appreciate. Subject to fanciful theories and speculations for thousands of years, tides have long piqued our curiosity. Sir Isaac Newton (1642– 1727) first identified gravitational forces as the prime movers in tide generation. He laid the basis of tidal theory, conceiving an "equilibrium tide" that would apply to ideal conditions of a global ocean. Pierre Simon de Laplace (1749– 1827) shares honours with Newton because he formulated equations of motion for tides on a rotating Earth, and was first to distinguish tidal phenomena according to different types (species) of tides. Laplace perceived the harmonic method of analyzing tides, later elaborated by Lord Kelvin (1824– 1907), inventor of the earliest tide-predicting machine in 1872.

Yet despite the ease with which tide tables are now constructed, no one to this day really comprehends how gravitational forces work. Neither has anyone been able to conceive a purely dynamic theory to directly relate tide-generating forces to actual tides (Clancy 1969; LeBlond and Mysak 1978). Laplace recognized that there could be no perfect theory, the chief difficulty being Earth's rotation. He preferred to write in terms of tidal waves, with the same periods as those induced by rhythmical components of gravitational forces. The problem of tides is that of a fluid motion modified by the geometry (including depth) of ocean basins, by friction, and by such forces as the Coriolis effect due to Earth's rotation.

The ebb and flow of tides provide fascinating measures of an ocean's pulse (Defant 1958). Nowhere does this beat more impressively than in the Bay of Fundy on the western shore of the North Atlantic Ocean. Here the tide attains world record levels, with tidal range approaching 16 m at times of particular astronomical conditions. Storm surges, tidal bores that gain height in an estuary, and great waves that batter the shore: all stir the blood and instil respect for the raw power of nature.

Other processes related to tides, among them currents, shifting sediments, erosion, development of tidal estuaries, and salt marsh growth and decay, work more subtly, creating changes over decades, centuries and millennia, rather than in a matter of hours or days. These processes, in turn, evolve in response to postglacial sea-level recovery. The unmistakable hand of Homo sapiens too, is now everywhere evident along the shore, for it is in the world's coastal zones that humankind has by choice become most concentrated.

The main purpose of this review is to provide a general survey of tides and an overview of their relationship to, and effect on, the Bay of Fundy (henceforth sometimes referred to as "the Bay"), home to the world's highest tides. Worldwide there is now an increased focus on understanding tidal processes and documenting changes on short-, medium-, and long-term time scales by various historical and scientific means. The Bay of Fundy and its tidal processes form a dynamic entity, and serve as a model for comparison with similar tidal regimes elsewhere.

Defant (1958) described tides, colourfully, as "… the heartbeat of the ocean, a pulse that can be felt all over the world." More technically, "tide" is the periodic rise and fall of the ocean surface in response to gravitational forces of the Moon and Sun. The periodicity of the tides is imposed by astronomical cycles determined by the relative motions of the Earth, Moon, and Sun. The regularity of tidal movements makes accurate prediction possible, and sets tides apart from other changes in sea level and from irregular phenomena like earthquakes, storms, and volcanic eruptions. In practice, tides are a mixture of diurnal (daily) variations, with one low and one high tide each day, andsemidiurnal (twice daily) variation with two low and two high tides each day. In relatively shallow coastal waters these motions are of course greatly magnified. The "range" of the tides generally refers to the vertical movement of the water surface between the Low Water and the High Water levels of the tide. This factor, and the regime or mixture of the two types of tides varies from one area to another and also over time at any given location.

In Canada, the tidal levels in use (Canadian Tide and Current Tables 2004) are:

• MWL – Mean Water Level – average of all hourly water levels observed over the available period of record; in comparison, Mean Sea Level (MSL) is a statistically established entity.

• HHWLT – Higher High Water, large tide – average of the highest high waters, one from each of 19 years of predictions.

• HHWMT – Higher High Water, mean tide – average of all the higher high waters from 19 years of predictions.

• LLWMT – Low Low Water, mean tide – average of all the lower low waters from 19 years of predictions.

• LLWLT – Lower Low Water, large tide – average of lowest low waters, one from each of 19 years of predictions.

• LNT – Lowest Normal Tide – in present usage it is synonymous with LLWLT, but on older charts it may refer to a variety of low water chart datums – it is also called Chart Datum (CD), a most important term of reference.

All tidal measurements are made from the local Chart Datum (CD), an elevation so low that the tide at that place will seldom if ever fall below it. Thus, soundings on hydrographic charts show mariners the minimum depth of water. The tidal range gives them an extra margin of safety. Generally the tidal range is small and so is the margin of safety. However, for the Bay of Fundy, on charts showing a number of tidal stations, the difference between Chart Datum and Mean Water Level in one section of the charted area may be quite different than it is in other sections. The soundings on such charts do not allow one to construct a proper three-dimensional picture of the shape of the Bay. On land, the datum used by geodesists, surveyors, and engineers is the Geodetic Survey of Canada Datum (GSCD, or GD). This datum is based on the value of mean sea level prior to 1910 as determined from a period of observations at tide stations at Halifax and Yarmouth, N.S., and Pointe au-Père, Quebec, on the east coast, and Prince Rupert, Vancouver, and Victoria, British Columbia, on the west coast. In 1922 the datum was adjusted in the Canadian levelling network. Because in most areas of the Maritime Provinces the landmass is submerging relative to mean sea level, geodetic datum drops gradually below mean sea level. However, there is a dearth of data, and no one is certain exactly what the difference is between GSCD and MWL at different stations. This situation is troublesome for engineers and biologists who need to know the proper relation between the two datums at particular places.

Between the Atlantic Ocean and the eastern seaboard of the North American continent lies the continental shelf. In large measure a result of planation during the ongoing Pleistocene– Holocene Ice Age, the continental shelf has depths of less than 250 m. Near Cape Breton Island, it extends more than 200 km from the coast, although near the southern tip of Nova Scotia it narrows to 130 km (Fader et al. 1977). Farther south, along the coast of the United States, the shelf varies between 25 and 200 km. Off the continental shelf the seabed drops to depths of 5000 m over a distance of about 200 km. This 200 km-wide margin is known as the continental slope. Along the outer edge of the shelf the average tide ranges between 80 and 100 cm. After crossing the continental shelf, the tidal range along the shore of Nova Scotia is increased to between 120 and 140 cm.

When the relative positions of the Earth, Moon, and Sun generate exceptionally large tides, the tidal range along the shoreline of Nova Scotia may reach as high as 200 cm. This increase is 40– 50% above average. Conversely, when the relative positions of the Earth, Moon, and Sun are such that tides are weakened, ranges drop by about 40% below the average. Variations in tidal strength of 60% to 140% are observed in the Gulf of Maine and the Bay of Fundy (Fairbridge 1966).

Between the Gulf of Maine and the Bay of Fundy, along the edge of the continental shelf off southwestern Nova Scotia and Cape Cod, Massachusetts, a series of shoals and banks acts as a sill obstructing tidal flow (see Fig. 3); in places the water is barely 4 m deep. Three channels cross this sill, of which the Northeast Channel is by far the largest between ocean and Gulf. Located between Georges Bank and Browns Bank, it is 230 m deep, 40 km wide, and 70 km long.

Tides on the ocean side of the Northeast Channel have an average range of 90 cm, but 320 km eastward at Bar Harbor, Maine, they have a mean range of 310 cm. High Water on this part of the coast is 3 hours later than along the edge of the continental shelf. The strength of tides in the southern bight of the Gulf of Maine, up to Race Point off Cape Cod, is rather uniform, varying between 210 and 310 cm for average tides; however between Bar Harbor and Jonesport 50 km to the east, their strength steadily increases toward the upper reaches of the Bay of Fundy. The average tidal range at the entrances to Minas Basin, Cumberland Basin, and Shepody Bay, each about 320 km from Bar Harbor, is 960 cm. However, in the Minas Basin the tides are 100 minutes later than at Bar Harbor. Tidal ranges are greatest in the estuaries of the Salmon and Shubenacadie rivers in the Minas Basin, 400 km from Bar Harbor. Depending upon astronomical conditions, this average range increases to 1360 cm. To the dismay of many inhabitants along this coast, even this enormous tidal range may be greatly extended by storm conditions.

Figure 10 (see appendix) shows the tidal ranges along the eastern Canadian seaboard during mean and large tides. At most places the tide reaches a maximum level on an average interval of 12.4 hours. Usually one of the two daily High Waters is higher than the other and is called the Higher High Water. Similarly, a Lower High Water and a Lower Low Water occur during a day-long tidal cycle. The distance that the tide moves up or down from the Mean Water Level is called the amplitude of the tide, and the total vertical distance between High Water and Low Water is the tidal range. The range given in the Canadian tide tables for a given locality is specifically for the distance between Higher High Water and Lower Low Water. This gives a somewhat larger value than ranges listed in American tide tables, which are calculated from the mean of the semidiurnal tides.

The center line of the band representing the tidal ranges (Fig. 10) depicts the local Mean Water Level; the pair of lines closest to this center line delineates the range during average tidal fluctuations in the area, and the outer pair delineates the fluctuation during large tides. Large tides occur when the main tide-producing astronomical forces are at maximum strength and working more or less in unison. The rise in feet of the Higher High Water above Mean Water Level during large and mean tides, and the fall of the Lower Low Waters below that datum, are shown on the chart beside the inserts which indicate the tidal characteristics near the principal tidal stations in the region.

Along the eastern Canadian seaboard the tidal ranges are rather uniform at 4 to 5 feet (1.22– 1.52 m). The smallest tidal ranges in the charted area are recorded on the west coast of the Magdalen Islands and in the western Northumberland Strait, with mean ranges of 1.5 and 2.0 feet (0.45 m and 0.61 m). However in embayments such as the Bay of Fundy, the Bay of Chaleur, the channel leading to the St Lawrence River estuary, and the eastern and central parts of the Northumberland Strait, tidal ranges gradually increase at points more distant from ocean tides (White and Johns 1977).

Tidal characteristics near the principal, or reference, ports (e.g., Halifax, North Sydney) are included as marigrams in Fig. 10. The examples show the predicted tidal movements for March, 1966, a date chosen arbitrarily, the particular astronomical conditions for which are given with the sample set in the legend (inset at lower left of Fig. 10). The marigrams for this month are typical of any monthly period except that the sinusoidal diurnal and semidiurnal variations are offset according to astronomical conditions. In order to accentuate these characteristics, different vertical scales are employed. For this reason the local tidal amplitudes of Higher High and Lower Low Water during large and average tides are printed to the left of the inserts (see "Typical Tidal Variations over one Month"). Note that ranges are largest shortly after full moon and new moon during spring tides. When the Moon shows its quarter phases, the amplitudes are small and the tides are neap. The influence of the distance of the Moon from the Earth is such that when the Moon is in perigee (closest to Earth), spring tides are higher than when the Moon is in apogee (farthest from Earth). The diurnal inequality between high waters and low waters is nil when the Moon moves through the plane of Earth's equator, and as seen from Earth, reaches its peak when the Moon is in its most northerly or southerly position in the sky.

Local tidal characteristics along the eastern Canadian seaboard result from a combination of diurnal tides and semi-diurnal tides. The diurnal tide repeats itself every 24.8 hours, and the semidiurnal tide every 12.4 hours (see insert, lower left hand side of Fig. 10). At most locations the semidiurnal tide is dominant. An exception is in the southern Gulf of St. Lawrence where, under certain astronomical conditions, only one High Water and one Low Water occur daily; here semidiurnal tides may have less than half the range of the diurnal component. Diurnal tides dominate along the north coast of Prince Edward Island, the west coast of the Magdalen Islands, and in the western portion of the Northumberland Strait.

Tidal characteristics also depend on the phase relationship between the diurnal and semidiurnal components. This manifests itself prominently in the Gulf of St. Lawrence where the strengths of both components are similar. Here there is less variation than in the Fundy tides, and there is no annual progression of 1.5 months in the highest tides. Tides in the Gulf of St. Lawrence tended to be higher in 1969 and 1973 (Fig.11), years of relatively low tides in the Bay of Fundy (Canadian Hydrographic Service 1964, 1969a, 1969b, 1973, 1979, 1981).

Several other interesting contrasts exist among regional tides. In the Gulf of St Lawrence, when the High Water of a diurnal tide coincides with one of the two High Waters of the semidiurnal tide, the Higher High Water of the day is higher than the Lower High Water, assuming that the semidiurnal tide dominates at the location. This situation is because the other semidiurnal peak will coincide with the diurnal Low Water. Under these conditions the semidiurnal Low Waters will occur when the diurnal tide is near Mean Water Level and both will drop to almost the same level. This combination occurs in the western section of the Gulf of St. Lawrence (Fig. 10) where the High Waters at Rustico, P.E.I., Portage Island, N.B., and Pointe St. Pierre, Quebec, are usually unequal, in contrast to the almost equal Low Waters. (In Fig. 10 this is shown on the inserts indicating tidal characteristics near certain ports.) Note too, in Fig. 11, that the highest tides in the Gulf of St. Lawrence occur at mid-year and at year end, contradicting the age-old notion that equinoctial tides (March and September) will be the larger tides. Conditions similar to those of the Gulf of St. Lawrence prevail along the east coast of Newfoundland near St. John's and Argentia. The opposite conditions prevail in the Northumberland Strait where one of the semidiurnal Low Waters nearly coincides with the diurnal Low Water resulting in generally unequal Low Waters and almost equal High Waters (e.g., at Charlottetown, P.E.I., Pictou, N.S., and Shediac, N.B.). As shown in Fig. 12, for Halifax, North Sydney, and Charlottetown, the winter tides are predicted to be higher than summer tides. Note, too, that the Low Waters of Charlottetown extend much further below MSL than the High Waters reach above this level. This is because of the diurnal inequality of the tides, which causes High Water to be nearly the same each day and Low Water level to vary.

As a result of the Coriolis effect, in large bodies of water such as the North Atlantic Ocean and the Gulf of St. Lawrence, the tides and especially the semidiurnal tides circulate along the surrounding coastlines in an anti-clockwise progression forming an amphidromic system. At and near the system's central point the tidal range is nil. The location of the amphidromic system in the Gulf of St. Lawrence is shown in Fig. 10. Farther removed from the central point the tidal range gradually increases. Illustrating this progression the chart shows (in lunar hours, 1 to 12) the approximate positions of High Water of the semidiurnal component of the tide in this system. This semidiurnal system splits at the western side of the Northumberland Strait, one wave going directly into the Strait and the other traveling around Prince Edward Island and moving partly into the eastern entrance of the Strait; as it moves westward to Shediac Bay (see Fig. 10), it meets the wave trough of the one entering the west entrance. The result is that the semidiurnal tide is scarcely noticeable in this area, and so the diurnal tide becomes dominant. Only when the Moon is near the Equator will semidiurnal tides be observed.

Tides in the Gulf of St. Lawrence are generated and maintained by tidal movements in the Atlantic Ocean. The two systems appear to be synchronized, with simultaneous High Waters and simultaneous Low Waters at their meeting place in the Cabot Strait. Thus, the diurnal tide in the Gulf maintains a see-saw relationship with that of the Atlantic Ocean, being high in the Gulf when it is low in the ocean. The Coriolis effect causes the diurnal tides to peak in the northeastern part of the Gulf before they reach their maximum level along the southwestern coastline.

During spring and summer in the Bay of Fundy and in the Northumberland Strait, the Lower Low Water occurs during the daytime (and HLW at night!) so that the tidal flats absorb considerable heat from the Sun. As the tide rises over the warm tidal flats, the water temperature rises, especially in the Northumberland Strait where the water cover is comparatively thin.

During the same seasons, in the western section of the Gulf of St. Lawrence the Lower High Water occurs during the daytime, covering to a limited depth the Sun-heated beaches and shoals with almost the same temperature-raising effects as occur along the Northumberland Strait. This has ecological implications. For example, both sections of the Gulf have a mollusc fauna similar to that found near Virginia along the eastern U.S. seaboard. The species involved need a water temperature above 68°F (20°C) to spawn. Yet, even during the warmest months of the year the mean air temperature in the Gulf is less than this value. Tidal characteristics of the area are therefore probably important factors in the propagation of these species in this part of the Gulf.

Conditions are reversed during fall and winter, exposing tidal flats to heat loss at Lower Low Water during the night. This results in the following Lower High Water being cooled well below normal sea water temperature as it flows over these frigid flats, circumstances that probably aggravate ice conditions.

A tidal wave with a period of about 12.4 hours appears in the Atlantic Ocean in front of the Gulf of Maine, causing a wave to enter through the Northeast Channel. The wave propagates over the Gulf, reaching the shore about three hours later. As this progressive wave moves along the mouth of the Bay of Fundy, a standing wave is formed, and energy becomes concentrated toward the head of the Bay as its cross-sectional area is gradually reduced.

A standing wave can form in a basin in which the water moves periodically from one side to the other, high water at one side corresponding to low water on the other. The strength of the tide is strongest at the ends of the basin. Although in the center of the basin the water surface does not move up and down, strong currents are present because a large volume of water must move to and fro in order to create the High Water alternately at both ends.

The funnel-shaped Bay of Fundy branches at its northeast end into the Minas Basin and Chignecto Bay. The dimensions of the Bay of Fundy are such that its natural period of oscillation closely approximates that of the semidiurnal tidal component, thus greatly increasing resonance. There are fascinating variations on this theme. For example, as detailed in section 1.4.3 (on applications to tides and water currents), due to the Coriolis effect the entering tidal stream is deflected toward the southern shore of the Bay and the retreating stream to the north.

On average, the up and down sweep of the tide solely due to the Coriolis effect is larger on the Nova Scotia side for the incoming tide than for the outgoing tide on the corresponding opposite New Brunswick shore (see Fig. 13). Thus, when the water reaches its highest level (A), the tide is still coming in, filling the upper reaches of the Bay, and the water maintains a lateral slope AB (Fig. 13a). At the turn of the tide, it is level (BC), although the level on the Nova Scotia side will have already dropped (A to C). Note that, like the sweep of the tidal level due to the Coriolis effect, the range of the incoming tide in any Fundy cross-section is higher perpendicular to the axis of the Bay on the Nova Scotia side. Likewise, tidal flow entering the Gulf of Maine, results in a higher tidal range on the Nova Scotia side. The same argument applies to any cross-section as the tide ebbs (Fig. 13b). Approaching the turn of the tide, the water continues flowing seaward for some time, just as it still has kinetic energy which allows it to flow upward to higher level (D) on the New Brunswick side than on the Nova Scotia side (F). At the turn of the tide, it is level (DE).

Thus, the south shore experiences a greater tidal range than does the north shore, as well as a counter-clockwise system of residual tidal currents (see Fig. 13). Turbulent current in the narrow strait between Minas Channel and the Minas Basin can exceed 11 knots.

In the upper reaches of the Bay of Fundy, the standing wave fills the estuaries. However, the rate of rise of tide is so rapid that large flows of water into the mouths of the estuaries would be required if the water surface in the estuaries were to rise at the same rate as in the Bay. Because the estuarine cross-sections are relatively small, particularly at low tide, such large flows of water are not able to enter. Therefore, filling of the estuaries is delayed. Physical restrictions on flow velocities are much more severe in shallow water than in deep water. The result is that water flowing in the more shallow parts is overtaken in time by water which, although entering the estuary later, moves faster in the deeper water. Thus, on the gently sloping bottom of the estuary a tidal bore commonly develops, most impressively so when the tide has its largest amplitude.

Theoretically, in basins where a standing wave occurs, hardly any currents are at the end where the tides are strongest. However, if there are shallow estuaries to be filled, strong tidal currents will develop in a relatively short time. These currents are the main agents of transportation and deposition of sedimentary materials in the Bay.

High Water occurs progressively later further up the estuaries. This High Water can happen quite a distance up the estuary, and considerably later than the time of High Water at the mouth. The time differential depends in turn on the gradient of the estuary bottom within the tidal range. As we shall see (section 9.5), these two factors are intimately linked to the evolution of the tidal marshes within estuaries in the Bay of Fundy.

A profound difference exists between the tidal regime of the Gulf of St. Lawrence and that of the Bay of Fundy. This much is evident in the frequencies and occurrence of monthly maximum and minimum tidal water levels. Predicted tide water levels over typical 8- and 10-year intervals are shown for three ports in, respectively, the Gulf of St. Lawrence (Fig. 11) and the Atlantic coast (Fig. 12). Note that in all years the highest High Waters for Charlottetown, Halifax, and North Sydney (i.e., the Atlantic coast) were predicted around the winter solstice (Fig.12), whereas the lowest Low Waters occurred near the summer solstice. However, for the three ports in the Gulf of St. Lawrence (Fig. 11), the highest High Waters and the lowest Low Waters occurred near the solstices, but with no noticeable difference between those near the winter solstice and those near the summer one. The smallest ranges of the tides were predicted to occur near the equinoxes.

The above relationships clearly demonstrate that the tidal regime along the Atlantic coast is different from that in the Gulf of St. Lawrence. Both are unlike the one in the Bay of Fundy. While in most waters, the largest tides occur in the same part of the year, in the Bay of Fundy they are more influenced by the shifting coincidence of spring and perigean tides with the result that each year they occur about 47 days later than in the previous year. Overall, regional tides along the eastern Canadian seaboard are best described as a succession of standing and progressive waves. The standing waves occur in the Atlantic Ocean, the Gulf of St. Lawrence, and the main body of the Bay of Fundy, whereas the progressive waves occur in the Gulf of Maine, the upper reaches of the Bay of Fundy, and in the estuaries leading into the Gulf and Bay. Progressive waves, unless supplied by external sources of energy such as air currents, do not grow in strength, for they lose energy when moving through confining channels. Likewise, progressive waves entering a restricted harbour mouth lose strength when they expand over the wider harbour surface. However, as we shall see, the wave entering the Gulf of Maine through the Northeast Channel seems to gain in strength while spreading over the Gulf. In fact it becomes 3.5 times stronger on reaching the shoreline.

According to Pliny the Elder, writing in his Historia naturalis: "… tides swell more during the equinoxes, more during the autumn than in the spring, but… were empty at midwinter and even more so at midsummer". Unknown to Pliny, the tides are much complicated by the shape of the oceanic basins. The example of conditions prevailing in the North Atlantic on 5 September, 1975, with the new moon in perigee (and rather small declinations of the Sun, 6.5°, and Moon, 5.8°) serves very well to illustrate (Clancy 1969). With the Moon closest to Earth on its monthly orbit, higher than normal tides can be expected. The theoretical movements of individual particles in the ocean on this date (for details see Desplanque and Mossman 1998a) are rather small, clockwise, and predominantly diurnal. Nevertheless, the tides on that day were large. Table 4 provides the times and heights of the tides as they were predicted to be at sixteen ports along the eastern seaboard of the North American continent.

Do movements of the North Atlantic ocean follow water particle movements generated locally, or are they a result of an oscillating motion of the ocean with a period of 12.4 hours? First let us compare the times of High Water with the location of water particles in their cycle near 45° N latitude.

Table 4 shows that the highest tides on the western side of the Atlantic reach the highest levels of the day between 18:45 and 20:49 AST (moving from north to south, as indicated by the theoretical particle movements calculated for that day). The smaller high tides for that day occurred between 6:30 and 8:30 AST.

Semidiurnal tides are predominant in the Atlantic Ocean, the diurnal component being small (Fairbridge 1966). On the eastern side of the Atlantic (Table 5), the diurnal inequality of the tide is small. The High Waters from Morocco to France (about 15°N and 45°N latitude respectively) occur between 9:40 and 11:35 AST (13:40– 15:35 GMT) and 22:00 and 24:50 AST (2:00– 4:00 GMT), the tidal wave moving from south to north (Schwiderski 1980). These times coincide closely with those for the easternmost locations of water particles (Desplanque and Mossman 1998a).

Given the Atlantic Ocean's 5000 km width and 5000 m depth, a scale model of this ocean as an oscillating basin would be approximately 1 m square and 1 mm deep. As it turns out, such a model is unrealistic (see Kwong et al. 1997) although it clearly illustrates the shallowness of the ocean. For example, if the ocean is oscillating, then when the water is high at one side of the ocean it should be low at the other side. So, if it is high between 9:40 and 11:40 AST (Table 5) at the eastern side one should expect it to be low at the same time on the western side. However, according to Table 4, low tide occurs in the west between 12:25 and 14:29. It appears, therefore, that ocean tides are more in tune with water particle movements than with an oscillating system.

Except for the Gulf of Maine and the Bay of Fundy, the ranges of tides on the eastern side of the Atlantic Ocean are substantially greater than those along the North American coast (Apel 1987). Harmonic analysis, the principle first applied to tidal problems by Lord Kelvin (1824– 1907), is the most satisfactory method of comparing characteristics of tides observed at various locations (Schureman 1941, 1949; Fairbridge 1966). This method relies on the fact that the observed tide is the result of several astronomically driven partial tides. Just as complicated sea waves can be reconstructed by combining several wave trains, so also can tides be calculated by combining partial tides, or tidal constituents. For most ports bordering the North Atlantic Ocean (Desplanque and Mossman 1998a), the M2 constituent is dominant (see Table 3; Fig. 14).

Since 1977, tidal records have been collected along the edge of the continental shelf off North America, and more recently at several stations in the mid-Atlantic between Newfoundland and Portugal. Tables 6 and 7 show the results of analyses of these data together with data from shore stations. The ranges of the major constituents at these stations are given and the strength of other constituents is compared with M2. This can be used to determine if the local S2 tide is close to its theoretical strength of 46% of the M2 strength.

The value of the initial phase p (known as the phase lag and representing the phase of the constituent at the time of origin) is given as the epoch (Table 6), based on the local meridian of the observation station, and the 65°30' W meridian. The use of the latter as a time meridian is prompted by the fact that this meridian crosses the main entrance of the Gulf of Maine and also the entrance of the Bay of Fundy. Note in Table 6 that the percentage of S2/M2 is close to the theoretical value of 46% near Newfoundland, but slightly less in the eastern North Atlantic. The relative strength of the N2 constituents remains close to 22%, while the already weak diurnal K1 and O1 constituents are even weaker in mid-ocean. One might expect that the tidal wave would follow the apparent westerly movements of agents that account for the existence of the various constituents. However the phase lags indicate that the co-tidal lines move from west to east, in the opposite direction of the imaginary heavenly bodies.

Table 7 shows the same type of data for ports along the eastern seaboard of North America, with additional data from some stations offshore. Indications are that the time of High Water is virtually the same for all stations over a distance of 2200 km between Banquereau Bank, south of the Cabot Strait, and Charleston, South Carolina. The strength of the Sun's influence is much less than the theoretical 46% of the Moon's influence, dropping from 24% at the Banquereau Bank to 17% at Charleston. Again there is little change in the relative strength of the N2 constituent. Thus, variations in the tidal strength caused by the variable distance between Earth and Moon gain in significance over variations in the influence of the Sun's gravitation. Stated otherwise, the perigean tides become more important than the spring tides.

The influence of the Sun is far more noticeable along the coast of Newfoundland and Labrador. The percentage of S2 with respect to M2 ranges from 36% on the south coast of the Avalon Peninsula, north of the Cabot Strait, to 68% near Fogo Island at the northeast corner of Newfoundland. It drops again to 50% near the northern tip of the island, and to 36% along the Labrador coast. Evidently tides near Newfoundland have characteristics in common with those in the mid and eastern North Atlantic whereas tides south of the Cabot Strait have a different pattern.

Other characteristics also indicate that tides between Newfoundland and Europe are of a different regime than those along the North American mainland. One of these characteristics concerns the stronger equinoctial tides. Plinys oft quoted statement probably accounts for the assumption by most oceanographers that the so-called equinoctial tides (March and September) reach higher levels than tides close to the solstices (June and December). Pliny may have been correct, for he was only aware of tides that can be observed on the east side of the Atlantic Ocean. Table 8 shows the highest predicted tides for each month in 1953 and 1975 for stations on the east side of the Atlantic. The semi-annual peaks are underlined. These data support Pliny's claim that the tides near the equinoxes are generally higher than those occurring during the remainder of the year. Even the fall tides are stronger than those in the spring, and the midwinter ones are stronger than the midsummer ones. However, as shown below, Pliny's claim is not valid along the east coast of North America, or along the continent's west coast south of the Alaskan panhandle.

Data for 1953 and 1975 from stations on the west side of the Atlantic (Table 9) are rather less supportive of Pliny's case. However, in order to test for possible bias in the two sample years, Table 10 supplies a more complete set of data for Halifax. Note that the peak tides do not fall in the same months as they did on the other side of the ocean, and that they are spread over all seasons. The only month in which no peak occurs is September, one of the equinox months. Each successive year the peak appears to arrive somewhat later. This can be explained by the dominance of perigean tides on the west side of the ocean. The Spring Tide cycle has a period of 14.77 days, and the perigean cycle one of 27.56 days. After 205.89 days, these cycles overlap each other causing interference and extraordinary strong tides. Every second overlap occurs the following year, but 47 days later in the season. It is, therefore, incorrect to claim that on the western side of the ocean the equinoctial tides are the highest tides of the year. The highest tides of the year occur in the late fall and early winter season, because the mean water levels are higher in winter than in summer due to world-wide shifts in atmospheric pressures. Another phenomenon discernible in Table 10 (last column) is that every 4 to 5 years the tides reach another peak. This peak occurs because after 112 cycles of Moon declinations, their peaks occur again at almost the same time. The result is a 4.53 year cycle of extraordinary high tides.

As discussed earlier, the theoretical movements of water particles in the oceans, due to the direct gravitational effects of the Moon and Sun, generally result in diurnal loops. Only during either new or full moon close to the equinoxes do strictly semidiurnal movements occur near the Equator. During fall and winter the most westerly displacements are reached during the morning hours, and during spring and summer in the evening hours. The diurnal inequality of tides along the North American seaboard shows the same tendency. During the summer months the most easterly displacements are reached around 04:00 and 05:00 hours local time, and during the winter, around 16:00 and 17:00 hours. As shown in Table 11 the tidal observations along the eastern side of the ocean are not in step with these particle movements. Clearly, tidal characteristics on the eastern side of the Atlantic differ significantly from those on the western side. It seems likely that the North American ocean basin, with its Sohm, Hatteras, and Nares abyssal plains, enjoys its own tidal regime, distant from that of the basins between Newfoundland and Europe.

We turn now to the behaviour of tidal waters along the coast of northern New England where North Atlantic tides impinge on the Gulf of Maine and the Bay of Fundy.

Like the rest of the continental shelf, the Gulf of Maine results from drowning of the coastline due to the postglacial rise in sea level. The tides which beset its shores are perhaps best described as a degenerate amphidromic system. The amphidromic point, if such existed, would lie southwest of Cape Cod. However, the waters are too deep in the Gulf, considering its area, for a full amphidromic system to develop. Thus the tides of the Gulf of Maine are generated by the tides of the North Atlantic Ocean rather than directly by the Sun and the Moon. According to Redfield (1980) "The coast of New England north of Cape Cod opens on the Gulf of Maine where the behavior of the tide determines the tide in the numerous passages along the coast." In effect, the Gulf of Maine and the Bay of Fundy should be viewed as a single tidal system.

As mentioned previously, fronting the Gulf of Maine, at the edge of the continental shelf, is a natural sill which separates ocean from Gulf. This sill consists of a sequence of shoals and banks running from Cape Sable at the southern tip of Nova Scotia to Chatham on Cape Cod. Water that fills the Gulf of Maine between Low Water and High Water must move over the sill or pass through three channels between the banks (Apollonio 1979). Figure 15 shows the location and length of the sill along a line that approximately follows its shallowest parts.

The distance along the shallowest part of the sill fronting the Gulf of Maine is 589 km, although the great circle distance between Cape Sable and Cape Cod is only 414 km. The crest of the sill runs from Nova Scotia, over German and Browns Banks, continuing on the other side of the Northeast Channel over Georges Bank, the Georges, Cultivator and Little Georges Shoals, and the Nantucket Shoals to Nantucket Island and Monomoy Island near Cape Cod. In places such as Georges Shoals there is barely 4 m of water. Over much of its extent, Georges Bank, which forms the main part of the sill, is less than 60 m deep (see Fig. 15 and Table 12 for location of various features and place names).

Three channels through the sill provide the main connections between the Atlantic Ocean and the Gulf. They are:

1. The Northeast Channel between Georges Bank and Browns Bank, which is over 230 m deep, 40 km wide and 70 km long.

2. A channel 120 m deep and 15 km wide between Browns Bank and German Bank off Cape Sable.

3. The Great South Channel between Georges Bank and Nantucket Shoals.

The shallowest parts of Browns Bank are small in area. The largest opening in the sill is between Georges Bank and German Bank; thus most of the water moving in and out of the Gulf of Maine flows through the section formed by the first two channels. Using bathymetric and hydrographic charts of the area, the width of the passage and its cross-sectional area at and below certain levels can be estimated. Because the Northeast Channel is the main passageway, these data are provided for the entire sill and for the Northeast Channel only (Table 13).

Approximately 80% of the available cross-sectional area is in the Northeast Channel section. Flow through this section has less friction to overcome than over the shoals, marked as they are in heavy weather by strong tidal currents and dangerously high seas. Thus an estimated flow of 80% through the Northeast Channel section is almost certainly a minimum.

The Gulf of Maine has a surface area of approximately 95 000 km², not including the 13 000 km² of the Bay of Fundy (Conkling 1995). Average depth of the Gulf is about 150 m. Within the Gulf are 21 basins, of which Georges Basin (maximum depth 377 m), Jordan Basin (304 m), Crowell Basin (304 m), and Wilkinson Basin (280 m) are the most prominent. These basins are separated by ridges and swells where the depth is between 150 m and 200 m. Along the 1100 km shoreline is a shallow fringe with depths less than 80 m. In detail, the shoreline is replete with numerous bays, islands, and tidal estuaries, its total length approaching 8000 km (Pilkey et al. 1989).

In the millions of years before present that the Gulf of Maine was above sea level, subaerial erosion reshaped the exposed lowlands. Much of the sediment deposited during the early Cenozoic was removed, and valleys and deep stream beds were carved. The steep northern slope of Georges Bank may have been formed at this time. The main drainage outlet of the area was probably directed through what is now the Northeast Channel region.

Throughout the Pleistocene epoch, glaciers advanced and retreated several times on a global scale, deepening and widening valleys as well as causing sea-level fluctuations. Sea level dropped about 120 m as the last glacial advance removed water from the oceans, transferring it to continental glaciers. Advancing ice sheets were deflected eastward at the scarp-like northern edge of Georges Bank, and flowed as huge valley glaciers through what is now the Northeast Channel (Grant 1989; Keen and Piper 1990). On retreat, the ice left extensive terminal moraines and large areas covered with glacial out-wash. These and numerous other trademark glacial features have survived the drowning of this coastline by the postglacial rise in sea level.

Clearly, when the Gulf of Maine area was dry land, or covered with ice, the local tide action was of totally different character than it is now. At present, the amount of water exchanged during a tide of average strength within a 12.4 hour cycle, between the Atlantic and the Gulf, is about 300 km³. Of this total, about 100 km³ moves in and out of the Bay of Fundy. During the largest possible tides these amounts can be as much as 1.4 times average.

The watersheds of rivers flowing into the Gulf of Maine and the Bay of Fundy discharge about 95 km³ per year, or 0.135 km³ per tidal cycle. This is approximately 0.045% of the tidal prism of an average tide. Even during the spring when the river discharges can be ten times average, it is doubtful whether such a relatively small river flow can influence the circulation pattern in the Gulf.

During an average tidal cycle lasting 6.2 hours (22 357 seconds), flow through the available cross-sectional area of 36 km² needs on average to be only 0.37 ms^-1 (0.73 knots) to fill the Gulf with ocean water. Assuming a harmonic motion, represented by a sine curve, the maximum current speed is 0.59 ms^-1 (1.14 knots). This speed can be generated if a mass is dropped 1.8 cm at the surface of the Earth. It can also be caused by a column of water 1.8 cm high. Thus, if the water in the ocean is 1.8 cm higher than in the Gulf, the maximum velocity into the Gulf can be generated. At the seashore the average amplitude of the tide is 0.45 m. At mean sea level where the current is fastest, such a tide can rise 1.8 cm in 280 seconds (i.e., less than 5 minutes) (DeWolfe 1981). Therefore in order to receive all the water required to follow the ocean tide, the water surface in the Gulf has to follow the surface of the ocean by only 1.8 cm difference and less than 5 minutes in phase.

It has been argued that as the last glacial waned there was a time during which sea level was too low to allow tides of the present dimensions. The argument is that Georges Bank and other shallows were dry at the time, shortening the length of the opening. However, the remaining cross-sectional area in the Northeast Channel was still large enough to convey the water required to generate tides comparable to those today. The time lag would have been about 4 to 5 times larger than it is now but still small enough to prevent an appreciable change in the tidal regime.

Water entering the Gulf of Maine through the Northeast Channel moves as a wave through the channel, spreading quickly thereafter as a refracted wave throughout the Gulf. This wave travels 335 km to the coastline of Maine between Bar Harbor and Jonesport about 3 hours after entering the channel. Within half an hour, High Water will occur from Cape Cod Bay to the upper reaches of the Bay of Fundy except for Minas Basin, where there is a further delay of one hour.

According to Greenberg (1979) the average tidal range increases from 0.90 m at the entrance to the Northeast Channel to 3.10 m at Bar Harbor (Fig. 16). The ranges in the southern bight of the Gulf vary between 2.06 and 3.10 m, with High Water occurring a bit later. Stations in the center of the Gulf show intermediate ranges and times of occurrence. It seems remarkable that tides in the Great South Channel off Nantucket range from 0.63 m to 0.50 m, smaller than either ocean or Gulf.

Somewhere between Bar Harbor and Jonesport the tides assume a different character. Instead of having almost identical ranges as in the southern bight, the tidal range increases steadily in an east-northeasterly direction, reaching a maximum in the Minas Basin, 320 km from Bar Harbor. Here the highest High Water levels are reached in the estuaries of the Salmon and Shubenacadie rivers near Truro, about 400 km from Bar Harbor.

Between Cape Sable and Tiverton at the end of Digby Neck, the average tidal range increases from 2.13 to 4.80 m over a distance of only 115 km. Tides in this stretch of the Gulf shoreline of Nova Scotia adjust to the fact that the main thrust of the tidal wave is directed toward the Maine coast near Bar Harbor. The Gulf shore of Nova Scotia provides a short cut for the Fundy tides, resulting in a greater than usual increase in range per unit of distance along the shoreline.

A computer simulation developed by Greenberg (1979, 1987) of the behaviour of the M2 constituent indicates that the amplitude of M2 over Georges Bank and the Nantucket Shoals is the least of all (Fig.16), decreasing to as little as 0.25 (0.51/– 2). The close co-tidal lines between Georges Bank and Nova Scotia indicate the progressive wave character of the tide in that area, while the long distance in the southern bight and the Bay of Fundy indicates more of a prevailing standing wave (Greenberg 1987). High Water occurs last over the Nantucket Shoals and in the upper reaches of the Minas Basin (see Fig. 17). Standing waves in bays like Fundy, open-ended to the Gulf of Maine, behave much differently than standing waves in closed basins. From the edge of the continental shelf to the Minas Basin the tide takes from 3.3 to about 4 hours, an interval approaching a quarter of the oscillation period of the combined Gulf and Bay, accounting for the large tidal amplitude of the region.

The above considerations are of great importance in studies of tidal power generation because, among other things, construction of a tidal power dam in the Bay would lead to elevated tide levels. Controversy continues about the true resonant period of the Bay of Fundy, one problem in modelling being the determination of appropriate boundaries between Bay, Gulf, and ocean. Any major obstruction to tides in the Bay of Fundy would result in a change in the resonant period, which in turn could cause significant increase in the already extreme tidal amplification.

So much for the tides of a so-called "degenerate amphidromic system". Next we investigate the extent to which the Gulf of Maine and the Bay of Fundy act together as one tidal system resonating with the semidiurnal tides.

Bay of Fundy Tidal predictions are published annually for the principal hydrographic station at the reference port of Saint John. A mathematical approach using harmonic analysis is employed to compute the actual tide as the combined effect of all the tide-generating variables. The result is a large number of cosine curves, so-called tidal constituents or "partial tides", each representing the influence or characteristic of the local tide. Local tidal characteristics along the eastern Canadian seaboard result from a combination of diurnal and semidiurnal tides. However, semidiurnal tides are prevalent in the North Atlantic, and Fundy tides amplified by resonance across Georges Bank and through the Gulf of Maine are an integral part of the system (Gehrels et al. 1995; Davis and Browne 1996).

Recent observations of environmental characteristics of the Bay of Fundy suggest modern change in the dynamics of the system. They include the changing grain size distributions on the tidal mud flats (Amos et al. 1991), anecdotal observations from the fishing community of increasing water depths in some areas, and changing benthic communities (Percy et al. 1996). There is a need to better understand the dynamics of the Bay of Fundy and to promote efforts by concerned groups for a more detailed knowledge of seabed, oceanographic and biological conditions. Could the changes observed apply to the entire tidal regime? Are the tides stable or are they increasing due to changes in water depth or resonance length of the Bay? What is the future of the tides and associated currents in the Bay? These questions are central to the evolution of the Bay of Fundy and the sustainable management of its resources. In the following overview of the geology and evolution of the tidal regime in the Bay of Fundy, we elaborate on the tides at Herring Cove in Fundy National Park, N.B., and then examine the geological significance of the tides.

In strict terms, the Bay of Fundy is underlain by a half-graben with key faults on the northwestern margin. These faults, of Paleozoic or older origin, were reactivated at the onset of the opening of the present day Atlantic Ocean due to plate tectonic movements in the early Mesozoic (Keppie 1982; Greenough 1995; Withjack et al. 1995; Wade et al. 1996). Sedimentary infilling of the half-graben, termed the Fundy Basin, commenced with continental sedimentation about 230 Ma, during the Mid Triassic (King and MacLean 1976; Stevens 1977; Olsen et al. 1989; Withjack et al. 1998). During a late rifting stage immediately following the Triassic-Jurassic transition, voluminous basaltic lava erupted upon Triassic strata and was followed by mainly clastic sediment deposition possibly into the earliest Mid Jurassic (Olsen et al. 1989; Wade et al. 1996; Mossman and Grantham 1996). The entire sequence was then folded, uplifted and tilted southwestward in a saucer-shaped structure (Withjack et al. 1995). Thus rock exposures along the Fundy graben mark the boundaries of a major synclinal structure, the axis of which is located toward the center of the Bay. Mesozoic strata underlie much of the Gulf of Maine and the Bay of Fundy. Cretaceous sedimentary deposits preserved in isolated pockets in lowlands adjacent to the Bay of Fundy rest unconformably, usually on Carboniferous rocks (Stea and Pullan 2001; Falcon-Lang et al. 2003), suggesting that the Bay of Fundy Basin may have been covered by Cretaceous (and possibly Tertiary) deposits and subsequently exhumed.

In addition to basalt, much of the 1400 km coastline of the Bay of Fundy consists of erosion-prone sandstone and conglomerate. Erosion rates can exceed 1m/yr (Amos 1978), giving rise to sandy estuaries in which fine-grained clastic material accumulates in sheltered embayments. Some coastal sections of mainly Paleozoic siltstone and shale, for example in Chignecto Bay (Amos et al. 1991; Amos 1995a), contribute materials that persist in suspension through wave action and tidal cycling. Elsewhere, sections such as the basalt along much of the Nova Scotia coast, and the more massive igneous and metamorphic rocks along the New Brunswick coast, are relatively erosion resistant.

Glaciation has exercised important controls on the geo-morphology of the Bay of Fundy, as on its tidal regime. About 18 000 to 20 000 years BP, the Laurentide ice sheet blanketed most of Canada and extended far south of the Great Lakes. Crossing the Bay of Fundy and the Gulf of Maine, its approach to the edge of the continental shelf left blankets of glacial out-wash and huge terminal moraines peripheral to the ice sheet, as well as extensive drumlin fields. These deposits now form many of the banks and shoals along the Maritime and New England coastlines. Associated valley glaciers, such as one believed to have occupied the Northeast Channel in the Gulf of Maine, were also significant (Grant 1985, 1989; Keen and Piper 1990). During the last glacial maximum they contributed to a global sea level lower by 100– 130 m than at present. Overall then, there are two major phases to sea-level history in the Bay of Fundy: an early glacial emergence, and a present continuing submergence.

During the last 14 000 years the depth of the Bay of Fundy has changed appreciably as the land surface rebounded and sea level rose as the last ice sheets receded. One of the main features of the postglacial evolution of the Bay was the depth of water over Georges Bank; inflow of tidal waters was evidently restricted at the lowest point of relative sea level (Scott and Greenberg 1983). With progressive submergence of Georges Bank, the Bay of Fundy became more directly subjected to tidal forces. Mathematical modelling of the Gulf of Maine tidal system indicates that 7000 years BP, tidal ranges will have been 20– 50% of the present range (Greenberg 1979; 1987); by 4000 years BP they would have grown to 80%, reaching present strength about 2500 years BP, when mean sea level was approximately 7 m lower than present. There is general agreement too, that increasing erosion of the sea bed, with localized bottom scour and deepening, signals an increase in the dynamic energy of the tides (Bleakney 1986; Godin 1992; Fader 1996). However, the timing of increased resonance is controversial, because evidence from detailed salt marsh records indicates that tidal range was relatively subdued about 4000 yr BP (Shaw and Ceman 1999), a point earlier inferred by Grant (1970).

Tidal waves, as recorded on marigrams, are typically symmetrical, the shape of the falling limb being the mirror image of the rising limb. However a process of translation takes place as a tidal wave enters a shallow estuary. The wave becomes distorted (see equation 7), the rising limb being compacted within a shorter period, whereas the period of the falling tidal wave increases (Carter 1998). Friction is often given as the cause, although this is not necessarily so. This is because the rising tide is a function of the predetermined tidal wave, whereas the return is a matter of hydraulics peculiar to that particular channel. We define a tidal river as that stretch of a freshwater river where the dominating factor in shaping the river channel is the freshwater regime. A tidal estuary is the mouth of a tidal river where the tidal flow regime shapes the channel bed. Because tidal range and freshwater flow are variables, the tidal river and the tidal estuary are separated by a transitional section where the dominance of tidal regime and freshwater regime alternate.

The volume of an estuary upstream of an arbitrarily chosen cross-section can be conveniently divided into two parts. The tidal prism is the top part of the estuary between the Low Water and the High Water levels. More simply it is the volume of water flowing in and out of the estuary with the rise and fall of the tide. The tidal volume is that portion below Low Water Level. Because tides are variable in strength, the tidal prism and the tidal volume are variable, as is the wet cross-sectional area through which the water that fills the tidal prism must flow. In general, where a moderate tidal range prevails in combination with deep waters and a limited surface area, only small currents are required in the estuary to have its tide follow the outside tide in terms of vertical movement. An increase in the current speed can only happen when the water surface drops in the direction of flow. In outside tidal waters, the current speed is generally so small that it can be neglected. During Low Water, the upper parts of an estuary may become virtually empty. Water levels in the estuary are only able to match the rising tide if enough water flows through the available cross-sectional area to fill the tidal prism to the level reached by the tide. Except for the extremely critical conditions of tidal bore development, the maximum current speed possible in a cross-sectional area is limited, being proportional to the square root of the water depth. Rectangular cross-sectional areas are able to discharge much more water than either parabolic or triangular cross-sectional areas. Sediment transport is of course associated with the tidal cycle, and at any given location velocity changes in tidal currents bring about periodic changes during flood and ebb portions of the cycle.

The vertical movement Y of the tide in seas and deep bays can be roughly expressed as:

-> See the list of figures

where A = the amplitude of the tide, t = the time expired since the occurrence of the High Water selected at the time of origin, n = the speed number, or the change in angle during a unit of time.

An appropriate speed number (n) for describing average semidiurnal tidal conditions in the Bay is 28.984° per hour or 1.405 · 10^{– 4} radians per second. The reciprocal v, of the last value is 7116.5. When seconds are used as time units, (22) can be replaced by:

-> See the list of figures

The rate of rise or fall Y' per second, of the tide in front of the mouth of the estuary at any value of t or Y can be expressed as:

-> See the list of figures

-> See the list of figures

These equations indicate that the rate of rise and fall is strongest when Y = 0, i.e., when the tide is at Mean Water Level. In the upper reaches of the Bay of Fundy, the mean tidal amplitudes are 6 m. Here, the tide will rise at a rate Y' of 6/7116.5 = 8.4 · 10^-4 ms^-1 or over 3.0 m per hour. To fill a relatively large estuary at this rate requires large amounts of water. The momentary rate of rise p't in the tidal prism can be expressed as:

-> See the list of figures

where Vm = the maximum possible flow velocity in cross-section A, and S = the surface area of the tidal volume at a certain time t.

However, there are conditions in which the rate p' cannot equal Y', either because the values of Vm and A are relatively small compared to S, or because the value Vm depends on the depth of the flowing water. In order for the incoming tide to gain speed, the water level of the tidal volume must be below the level of the outside water. If the wet cross-sectional area is A, the required velocity V through the gap must be at least Y' · (S/A) ms^-1, because eddies and local currents prevent the total area being available for the discharge. These relationships are explored in greater detail in section 6.3.

Basically, a tidal bore is a wave of translation, a hydraulic shock wave formed by a rolling wall of water. Analogous to a sonic boom (Lynch 1982), it develops when the rising tide forces water to move up a river channel faster than it otherwise could propagate in a given depth of water. Ideally the tidal range should be high, the river shallow, and the channel gradient slight. In reality a tidal bore is a single wave or a group of symmetrical sine waves that travel upstream signalling the turn of the tide at Low Water. In deep water these waves move at different speeds and separate from each other. However, in shallow water the component waves of the bore are translated into a single wall of water.

An estuary generates a bore when the level of the water surface is unable to keep up with the rising tide. Too small a cross-sectional area is available in the river channel. Thus at the mouth of the estuary more water enters than can be moved along at the front of the bore. The developing wave travels more quickly in deep water than in the shallow water of the estuary. Upstream, cross-sectional area and depth decrease. Consequently, possible current speeds are smaller in front of the advancing wave and the volume of the flow is decreased. As excess water overtakes the bore its height increases. Bore height continues to rise until the rate of rise in the tide becomes insignificant. The bore terminates when the tide reaches its most upstream point in the estuary. Under favourable geographical and astronomical conditions, a bore may be a metres-high wall of water moving up an estuary with a roar heard tens of kilometres distant; under other conditions, it is a leisurely moving ripple.

Waves of translation are surprisingly common. For example, during a heavy rainfall a sheet of water may be seen flowing down a steeply-sloping asphalt street. Here at regular intervals, waves form which move faster the higher they become. Another example occurs with the last uprush of water upon a sandy beach, pushing bits of shell, seaweed and other light materials to the line of its farthest reach. Here it is momentum that carries the water up the beach and over a slope upon which water from the previous wave has mostly been drained. The steep front of the breaking wave is, in effect, the crest of a miniature bore. These waves also occur in canals used for inland navigation when water is suddenly discharged from a reservoir or lock into the canal. Wave height here depends on the amount discharged and the dimensions of the canal. However, this is a negative wave of translation that results in a distinct drop in the level of water moving down a canal. Likewise, a negative bore also may occur when a lock is filled at an appreciable rate. The following theory outlines the conditions governing maximum flow.

Elementary physics provides the following equations for free fall near the surface of the Earth.

g = the acceleration (9.806 ms^-2) due to the gravitation of the Earth near its surface.
h = 0.5 g · t²
h = vertical drop in free fall
v = g · t, the speed attained after at the end of a vertical drop
t = time

Thus:

-> See the list of figures

Note that this relationship between drop and velocity holds whether the resulting movement is vertical or not. Pressure due to differences in surface elevations will cause the same speeds. Also,

Q = c · v · A , where Q = volume of water passing through the channel in 1 s and
c = coefficient of loss in average speed due to eddying and heat production,
v = theoretical speed = (2g · h)^0.5
A = wetted cross-sectional area of channel, and
D = water depth.

Assume that water is moving from a deep reservoir under the conditions illustrated in Fig. 30. In order to attain a speed v, the surface of the water near the entrance has to drop h metres, causing a pressure on the water in the cross-sectional area equal to the weight of the water in the column h metres high. In a rectangular channel B metres wide (Fig. 30) and the water D metres (H – h) deep.

-> See the list of figures

If h becomes larger, A = (H– h) · B becomes smaller.
Q will be zero if h = 0 (No speed)
Q will be zero if h = H = 0 (No area)
Q will be maximum when dQ = 0^0.5

Fig. 30 Sketch illustrates the condition of maximum flow in a rectangular channel caused by a decrease in head (h) of waterin a reservoir feeding into a channel H m below surface of the water reservoir. The limits for the ratio d/H for rectangular, triangular, and parabolic channels respectively, are 0.667, 0.80 and 0.75. For trapezoidal receiving channels, the limit is variable, but somewhere between these limits.

-> See the list of figures

-> See the list of figures

When dA is positive, dh will be negative, and since dA = B · dh, where B is the surface width of the flowing water:

-> See the list of figures

Note that the condition of maximum flow is caused by the drop (or head) h metres, which is equal to the wet cross-sectional area A prevailing in that condition, divided by twice its surface width. This is valid for all shapes of channels. For rectangular (r) or very wide channels A = B · D and:

-> See the list of figures

For triangular (t) channels A = B · D/2 and:

-> See the list of figures

For parabolic-shaped (p) channels A = 2B · D/3, where (B = K · D^0.5), and:

-> See the list of figures

Thus the speed of the water flowing into the channel will be highest when the channel has a flat bottom. The maximum flow Qr in a rectangular or flat-bottomed channel will be:

-> See the list of figures

The maximum flow Qt in a triangular channel, where B = kD, k = the relationship between surface width and maximum depth of water, and A = 0.32 k · H², will be:

-> See the list of figures

The maximum flow Qp in a parabolic channel, where B = k · D^0.5 and A = 0.433 k · H^1.5 will be:

-> See the list of figures

Thus, the maximum discharge in a receiving channel depends on the height (H) between the bottom of the channel and the surface elevation of the body supplying the water.

In a sinusoidal tidal movement the height of the tide can be expressed as:

-> See the list of figures

where At = tidal amplitude (half range), and t = is time from Low Water, in seconds. The rate of rise and fall of the tide can be expressed with the derivative of the above equation:

-> See the list of figures

If the bottom of the tidal channel is at elevation F, and a velocity head h is required to move water at a certain speed v, over the flat, then the depth D, in the channel is:

-> See the list of figures

In order to establish the dimensions of discharge sluices in major Bay of Fundy estuaries, it was necessary for the MMRA to determine the volume P of the tidal prism upstream of the construction sites. Following surveys it was found that there was a rather constant relationship between the tidal cross-section A and the volumes P of the tidal prisms upstream of them. A general relationship for estuaries or tidal basins has been proposed by Jarrett (1976) and by Van de Kreeke (1998). Regional characteristics determine the constant and the power of A. This relationship can be empirically expressed as:

-> See the list of figures

Tidal estuaries where bores occur are flat bottomed and the width is large compared with the possible water depth. Thus the cross-section is close to rectangular in shape. The following discussion assumes a rectangular cross-section. Thus A = B · D.

Taking d(h) as zero, so that the volume of a rectangular tidal prism to be filled in 1 second becomes:

-> See the list of figures

and

-> See the list of figures

whence

-> See the list of figures

To calculate the velocity v of the water, this discharge Q is divided by the area A = B · D

-> See the list of figures

From elementary physics the required velocity head h to generate this speed is v²/(2g), whence

-> See the list of figures

The maximum flow discharge is reached when this velocity head h = D/2. Whence

-> See the list of figures

Thus, maximum flow will be reached when the bottom F, of the channel, is at elevation T– 1.5D. If the channel bottom is higher, the inflow of water will be insufficient to fill the estuary at almost the same rate as the tide is rising. The level in the estuary has to lag behind because of the velocity head required to generate the discharge into the estuary. This velocity head is, of course, also changing with time. The examples given in Table 15 illustrate specific conditions that favour bore formation.

Note that the critical depth does not change much when the tide is near mean sea level. Thus the simplification of neglecting d(h) /dt has no serious effects. Note too that the critical bottom elevation is at lower elevations when the channel is wider. Therefore the bore can start as soon as these conditions are met.

As not enough water can enter to raise the water levels at the same rate as the tidal waters, the water that is able to enter the estuary will flow at the maximum speed, depending on the depth of the water. The first water to enter flows considerably slower than that which follows, as depth increases. In fact the later water will overtake the earlier water, with the result that the wave entering the estuary will do so as a bore of increased height. The water in the bore front is thus a mixture of volumes of water that had different initial speeds and momentum. For this reason the speed Vb of the bore is likely to be somewhat less than maximum flow speed indicated by the height H of the bore. Thus:

-> See the list of figures

Note that the maximum flow condition is actually unstable, although depths can deviate slightly from the critical depth without changing the discharge very much. This could explain the presence of "whelps" behind the bore front. Whelps are waves that are very conspicuous after the passage of the bore. But even if maximum flow conditions are not reached, the water surface elevation in the estuary will always be lower than the rising water in the tidal sea. This is because of the required velocity heads. Thus water near the head of the estuary will be flowing on a slope with decreasing cross-sectional area. Consequently the sinusoidal wave is distorted and becomes less and less symmetrical. This can be illustrated by plotting the centre line in marigrams, obtained at different locations along the estuary (see section 7.2).

A bore will be formed if the frontal part of the distorted tidal wave becomes vertical. It will grow in size when it flows over horizontal bottom plain(s). However, if the elevation of the estuary bottom is rising steeply, there is less chance of a large bore due to inadequate distance and time. When a tidal bore flows over a dry tidal flat, or shallow water, its progress is unrestricted. But as soon as the original water depth is 2/3 or more of the water depth of the bore, maximum flow conditions no longer prevail and progress becomes restricted. This figure approaches the 1.4 ratio, evidently the dividing condition between breaking and non-breaking bores. Note that in parabolic-shaped channels the original water depth would interfere when 3/4 of the bore depth, a figure corresponding with a ratio of 1.33.

With the exception of the Bay of Fundy, tides along the coastlines of North America are not particularly strong and thus not conducive to the formation of tidal bores. However, the bottom gradients of the Bay of Fundy are too steep to permit a large bore to develop. Thus, a tide with an amplitude of 5 m can only develop a bore 0.6 m high after travelling 6.2 km over a horizontal plane. With a tide of 2.6 m amplitude it requires a distance of 12 km, and with an amplitude of 1.5 m, a distance of 21 km is needed in order to develop a 0.6 m high bore. No bore will form with a tide of 5 m amplitude if the estuary gradient is steeper than 0.000193. A tide of 2.6 m amplitude cannot form a bore in a channel with a gradient more than 0.0001. In short, channels of certain dimensions can discharge quantities of water up to a certain limit, at which point maximum flow conditions prevail.

The global catalogue of tidal bores (Bartsch-Winkler and Lynch 1989) includes five of the main bores from the Maritime Provinces, all in the Bay of Fundy area. These are in the Hebert, Maccan, Salmon and Shubenacadie rivers (Nova Scotia) and in the Petitcodiac River (N.B.). Each constitutes an important tourist attraction at several localities. The rivers Hebert and Maccan empty into Cumberland Basin upstream from Fort Beausejour. Like all Fundy examples, the height of bores developed in these two rivers is variable, being dependent upon predictable astronomical conditions. The chief controlling factor is the rate of rise of water in Cumberland Basin, which of course peaks during either the new moon or the full moon phase, whichever is closest to perigee. Compared to predicted arrival times of tides using Saint John tidal data, the arrival time of the bores in these two rivers is generally earlier at spring tides than at neap tides. The same is true of the Shubenacadie and Salmon River bores near the head of the Minas Basin.

In contrast to the Hebert, Maccan, and Salmon rivers, the (observed) arrival times of the bore in the Petitcodiac River at Moncton is delayed more at spring tides than at neap tides. A check on the difference between predicted and observed tides at Saint John, and the prevalent weather conditions at the time, rules out weather as a cause of the delay. The explanation may lie in the fact that the Petitcodiac River bottom is below mean sea level, whereas at the other localities, river bottoms are well above mean sea level.

Unfortunately, a causeway built in 1968 upstream from Moncton shortened the estuary by about 21 km and decreased its tidal prism by about 20 · 10⁶ m³. A similar story applies to numerous estuaries: when estuary equilibrium is disturbed, channel readjustment follows. In the Petitcodiac, the decreased outgoing tidal flow was no longer able to remove as much silt as previously (Thurston 1990). Consequently the channel lost 450 m of its width and the bottom gradient became steeper because of a tremendously increased rate of silt accumulation. For example, between 1968 and 1971, an estimated 10 · 10⁶ m³ of sediment accumulated in the estuary.

Change in shape of the channel has greatly affected the size of the bore (Figs. 31, 32), which now must follow a steeper, more tortuous channel. As a result the bore falls far short of the 1.2 m height that it had prior to causeway construction. Further, the arrival time of this tidal bore has also changed during the last several decades. In 1920 Dr. W. Bell Dawson determined that the Petitcodiac bore occurred 2 hours and 22 minutes before High Water (spring tide) in Saint John. The present daytime difference between the arrival of the bore and Saint John High Water of 1 hour and 38 minutes indicates that the causeway and resultant change in the shape of the Petitcodiac River estuary has retarded the arrival of the bore by almost one hour since 1920. (The Bell Dawson difference was used and mentioned as a footnote in the tide tables between 1922 and 1983 inclusive. In the 1984 tables a shift was made to the new difference of 1 hour and 38 minutes).

The decline of the Petitcodiac bore disappoints local tourists and tourism promoters alike. In curious contrast, in China there is an ambitious program to suppress the bore in the Quiantang River. This bore and those of the Amazon delta (McIntyre 1972) are among of the most spectacular in the world. In the case of the Quiantang River, the drainage area is about 50 000 km^{2 2}(roughly equivalent to the watershed of the Saint John River, N.B.). Its tidal section starts at Fuyang, which is occasionally reached by a bore. However, the estuarine portion commences 18 km downstream of Fuyang, at Wenjiayan, and continues 101 km further downstream to Ganpu (Fig. 33). The main tidal channel shifts frequently, sometimes 245 m a day. Near Hanghzou the width of the estuary is about 1 km, and 20 km at Ganpu. Beside the tidal channels are vast tidal flats raised to about 1.2 m below the highest tide levels. During low tide the channel is shallow, many sections having a depth of 1 to 2 m, but even here where flow is rapid, ships of 100 tonnes or less can navigate.

Beyond Ganpu is the flaring Hangzou Bay, which runs 100 km to its 95 km wide mouth, open to the East China Sea. The tidal range in Hangzhou Bay doubles over this distance of 100 km to a maximum range of 8.93 m at Ganpu. This rate of increase is 0.7% per kilometre, twice the rate of the Bay of Fundy. Upstream from Ganpu, the river bed rises abruptly and the channel becomes shallow. According to Dai Zeheng (1982), the bore begins to form near Jianshan 19 km upstream from Ganpu, and reaches its maximum height in the vicinity of Baboa (33 km) and Haining (40 km) (see Figs. 34 and 35). The height of the bore is normally 1 to 2 m, but reaches a maximum height of 3 m. It usually travels 5– 7 ms^-1 (18– 25 km/hr) but has been recorded at 12 ms^-1 (43 km/hr) (van der Oord 1951). Stretched across the 3 to 10 km wide channel, it moves upriver like a silver curtain with a roar that can be heard over a distance of 20 km.

Although there are few accounts of the Qiantang bore in western literature (see Darwin 1898), the bore must have been present in the river for at least the last 1000 years. Long standing doubts of its existence probably arose because Marco Polo never mentioned the bore in his detailed account of the city of Hangzhou (which he called Kinsai). According to Masefield (1926), Marco Polo probably left this tale untold lest he lose all credibility. Unfortunately the original manuscript is lost, and the oldest written account of his story dates from 1559. The tale of a metres-high wall of water entering the river at regular intervals was probably demanding too much from his audience.

Because tides are variable in strength, the tidal prism and tidal volume in an estuary are also variable as is the wet cross-sectional area, which is that part of the total cross-sectional area filled with water. As discussed in section 6.3, it is through this wet area that the water which fills the tidal prism must flow. The upper parts of an estuary may become virtually empty during Low Water. Water levels in the estuary are only able to match the rising tide if enough water flows through the available cross-sectional area to fill the tidal prism to the level reached by the tide (Lauff 1967; Carter 1998). In many cases this is physically impossible as the maximum current speed possible in a cross-sectional area is limited, being proportional to the square root of the water depth (Kjerfve 1988). Rectangular cross-sectional areas are able to discharge much more water than either parabolic or triangular cross-sectional areas because the larger cross-sections allow higher current velocities.

In channels that empty, or nearly empty, during Low Water, the first water to return after the turn of the tide is very much restricted moving into the estuary. This is so even discounting friction, because the water depth and cross-sectional area are very small or zero. However the following water finds a partly filled channel and can thus move faster and at a greater discharge rate. As long as the discharge is inadequate to fill the tidal prism at the same rate of rise of water surface as the outside water, the result is that the seaward gradient of the incoming water will initially have a strong upward curvature. This curvature may become so strong that a bore will form at the bottom of the incoming wave.

What happens when, at a certain location in the estuary, the water reaches its highest level? Certainly the more upstream parts of the estuary remain unfilled with water. Seaward of this point the water level is already falling. Thus water continues moving landward on its own inertia when the high tide level is reached at that point. As a result, the water surface has already dropped before the direction of flow reverses. Thus the time of peak elevation and zero velocity are not coincident. When the water ceases its advance into the estuary it has already a considerable downward slope seaward. Soon a sizable seaward current develops, occupying a smaller cross-sectional area than when it moved into the estuary on the rising tide. The falling water level may at first follow the outside tide rather well because of the large cross-sectional areas, but progressively the reduced depth restricts water outflow. Consequently the seaward gradient of the water surface during ebb flow now has a downward curvature.

In terms of sediment transport this means that while the incoming tide is at high levels and gradually reducing in speed, its sediment carrying capacity is reduced (Amos and Tee 1989; Amos 1995a). Much of the released sediment is spread over the full height of the channel banks, and during very high tides, over the salt marshes also (Postma 1967). However the outgoing tide is concentrated toward the bottom of the channel, having at times higher speeds than the incoming tide. This situation promotes erosion through undercutting of the channel banks. Excess material high on the banks will slide down, resulting in unstable soil conditions. In testimony, the banks of tidal creeks and rivers in the Bay of Fundy region are characteristically bare of any vegetation except perhaps for some being transported downslope by undercutting (Scott 1980). Only where currents are slight is vegetation able to take root. Overall the intertidal flats of estuaries may vary seasonally, storing clays and silts during summers and removing them during heavy freshwater run-offs.

The Reversing Falls at the mouth of the Saint John River, N.B. is a perfect example of what happens when an estuary of limited cross-sectional area must serve a tidal prism with a large surface area. It was on Jean Baptiste day, 24 June, 1604, that the first Europeans in the region, Pierre de Gua, the Sieur de Monts, a Huguenot merchant, and Samuel de Champlain, Royal Geographer, discovered the mouth of one of the largest rivers on the eastern seaboard of North America. These pioneers and others who came later described its mouth in their logbooks and reports. According to the Jesuit missionary Pierre Baird (1611), "The entrance to this river is very narrow and dangerous, for a ship has to pass between two rocks, where the current is tossed from one side to the other, flashing between them as an arrow. Upstream from the rocks is a frightful and horrible precipice, and if you do not pass it at the proper moment, and when the water is smoothly heaped up, of a hundred thousand barques not an atom would escape, but men and goods would all be lost." (Raymond 1910). Doubtless Baird was well aware of the Indian legend concerning the hazardous entrance to Saint John Harbour. Here the Indian cultural hero Glooscap is said to have created the Reversing Falls when he destroyed a large beaver dam built across the river's mouth. (The remains of Glooscap's beaver dam support a popular restaurant!) The natural feature responsible for the legend lies at the downstream end of a 4 km long, tortuous river section. At Indiantown, water surface levels can vary between 0.3 m above Geodetic Survey of Canada Datum (GD) during periods of small river discharges, to 5.2 m above GD during extremely large runoffs. Two kilometres downstream of Indiantown the river suddenly narrows to 215 m, and a sill occurs about 4.5 m below GD (see Fig. 40). Downstream of this sill the channel deepens to 45 m below GD and steep rocks conspire to form a gorge 106 m wide, with the bottom 40 m below High Tide level and a cross-sectional area of 2125 m². During Low Tide the river at the reversing Falls is only 78 m wide and the cross-sectional area reduced to a mere 1400 m² (Fig. 41).

The average freshwater discharge of the Saint John River is approximately 1100 ms^-1. In spring, the melting snowpack in the watershed, combined with precipitation, can generate freshets of more than 10 000 m³/s. The cross-sections over the sill and in the gorge are far too small to allow these storm discharges to pass at normal levels, which are little above MSL. Therefore at times the water has to rise over 5 m, even during low water, before such freshets can fully discharge into the sea. Even then, tides restrict the outflow twice a day during High Water. The average range of tide in Saint John Harbour is 6.7 m, reaching an elevation of 3.4 m + (above) GD. Large tides have a range of 9.1 m and reach 5 m + GD.

Upstream of the falls a system of drowned valleys (Grand Bay, Kennebecasis Bay, Long Reach and Belleisle Bay) provides drainage channels for the river and some of its tributaries. They have a combined surface area of 200 km² at the present sea level, and 250 km² at flood levels about 5 m above GD. Due to sea-level rise this system has gradually been filling over the past several thousand years. The tide has a range of about 0.5 m, but in the section of river upstream from Evansdale it is reduced to 0.3 m, and half that again 10 km further upstream. Upstream of the Falls the tidal prism is approximately 0.09 km³, and can be filled within 6.2 hours by an average flow of 4000 m³s^-1 with an average velocity of 5.5 ms^-1 through the area of 730 km² over the sill, requiring a drop of 1.5 m (Hansen 1970). When the river is low and with only a small gradient in the water surface, backwater effects due to the tides are observed near Fredericton, New Brunswick's capital, 125 km upstream of the falls.

In Saint John Harbour the tide ranges from – 2.0 m to +2.0 m with respect to MSL during small tides and from – 4.5 m to +4.5 m during large tides. Upstream of the falls the range is at most between 0.0 m and 0.6 m above MSL. This means that at High Water in the Harbour, the water will flow through the gorge and over the sill in a landward direction, dropping 1.75 m to 4.2 m depending on the strength of the tide. But at Low Water in the Harbour, the water will flow seaward dropping over a short distance 2.3 m to 4.8 m. Tidal water, augmented by the river discharge, gains a considerable velocity when moving over the sill where the available wet cross-sectional area is smallest. Rapids result in this section as the water moves turbulently through the gorge into the Harbour.

Navigation under these conditions clearly poses problems. Since tidal current and river flow are always present, there is no appreciable slack water in the Reversing Falls section. The ideal time to navigate the section is as close as possible to the time of slack water as indicated in the tide tables. Slack water at the end of the inward current is 2.4 hours after the time of High Water predicted for Saint John. The tide has then dropped sufficiently to be close to the same level as it is in the river upstream of the falls. Another slack time occurs 3.8 hours after Low Water, when the tide has reached the river level. Although slack water lasts less than 10 minutes, the section can be navigated for half an hour before and after this slack. Variations in meteorological conditions can alter the time of slack water by up to about an hour. At times it is impossible for ships to move upstream. An example is during spring runoff when river levels upstream of the falls are so high that they are not even reached at High Water levels of the tide. Under these conditions a steady seaward current will persist throughout the whole tidal cycle.

Flooding of the several lakes upstream on the Saint John River has troubled settlers on the fertile freshwater delta since 1694. In April 1987 heavy flooding associated with an ice jam caused water levels to rise over 5 m. The basic problem was recognized as early as 1693 by Lamothe Cadillac who, while conceding that the Saint John is the most beautiful, most navigable, and the most favoured river of Acadia, was all for blowing up the rock on which the Reversing Falls restaurant is situated. Indeed, the idea has been revived on more than one occasion since. Fortunately there has been strong opposition. People perceive that such a project would change the Saint John River into a tidal river with unsightly mud flats and alternating currents of turbid water. The remedy for protecting the low-lying lands would be worse than the problem because the annual inundation with fresh water would be replaced with inundations of salt or brackish water at every set of high tides. Moreover the Harbour would be ruined because tidal currents would become much stronger than they are at present, increasing the navigational hazards. For these good reasons the Falls remain a unique estuarine feature of the Fundy landscape.

Tidal energy, exploited in Europe over 900 years ago, is highly predictable and is harnessed in much the same way as hydropower; all that is needed is a barrage across a suitable estuary that allows the bay behind it to alternately empty and fill with the tides. The greatest amount of energy available from falling water, at least in theory, is obtained by dropping the largest amount possible over the greatest vertical distance (Daborn 1977).

The construction and operation of a tidal power plant is however much more complex than a hydro plant on a river. At a given site, the amount of energy available depends on the range of the tides and the area of the enclosed bay. Since the head can never exceed the tidal range, unless water is pumped from the sea into the reservoir using energy generated elsewhere, the capacity of the plant can only be increased by moving more water through the turbines. This water must move by tidal action, within the period of 6 hours or less, from the sea into a reservoir when the sea levels exceed the reservoir levels. This water must then be temporarily stored in the reservoir until the tide drops enough to create a head that will drive the turbines. The larger the reservoir, the more power the plant can extract from the water. In principle it would be best to build the barrage at a point in the estuary close to the sea. Most estuaries flare and deepen toward to sea, requiring the barrage to be more voluminous and therefore more expensive and challenging to build. No matter what the location the control of tidal waters through the channel, particularly during the late stages of construction of the barrage, is extremely difficult. High tides during this construction period force large amounts of water through the closure gap. Nevertheless, in order to make a tidal power plant energy effective, an estuary is needed where large tides prevail. To illustrate (Bray et al. 1982; Gordon 1984; Gordon and Dadswell 1984), large sections of estuaries in the Bay of Fundy fall dry at Low Water. When the tide rises, the surface area of the water gradually increases. In most cases the water surface area for a trapezoid is approximated as follows:

-> See the list of figures

where M = surface area in km² , at MWL, N = increase in surface area in km²/m, y = the height above MWL in metres. The volume R of a tidal section between levels a and b (see Fig. 42) is approximately:

-> See the list of figures

whence

-> See the list of figures

or where b is negative,

-> See the list of figures

The tidal prism P, with a total amplitude of A metres then becomes:

-> See the list of figures

In the upper reaches of the Bay, the volumes of R and P can be large. Examples are given in Table 16 together with the values of potential energy Ep, during MHW.

The energy in a body of tidal water can be compared to the energy in the pendulum of a grandfather clock. The weight, or bob of this instrument, with mass m, is suspended a distance L from a rigid support. If the system is at rest, the force of gravitation will keep the bob at its lowest position, exactly vertical, below the support. However when the bob is drawn aside a distance x, it has to be lifted the vertical distance y. The relationship between y, x, and L can be expressed as:

-> See the list of figures

The potential energy Ep is equal to 0.5 m · g · x²/L, and the kinetic energy Ek can be given as 0.5 m (dx/dt)² , when x is small compared with L. The total energy is then the sum of Ep and Ek, and can be expressed as:

-> See the list of figures

indicating a simple harmonic motion. When the bob, with a mass of 1 kg, and suspended 1 m below its support, is moved aside horizontally 0.44 m from its equilibrium position, it must be lifted 0.1 m, thus being supplied with an energy of 0.1 · g Joules. When released it will be accelerated, reaching its maximum velocity of 1.4 ms^-1 when passing its equilibrium position. This velocity enables the bob to be lifted 0.1 m at the other side of the equilibrium position. In the clock, the energy lost due to friction with the support and the surrounding air, is replenished by a dropping weight via the escapement mechanism. When the motion of the bob is interrupted at the equilibrium point, the system is released of its total energy. In order to start the bob moving again, the same energy must be resupplied.

A body of tidal water can be likened to a pendulum lying on its side. At Mean Sea Level, its equilibrium position, the strongest currents occur and thus the maximum amount of kinetic energy (Godin 1990). The highest potential energy occurs when the surface is close to its High Water or Low Water positions. In order to reach its Low Water position, the water must be evacuated toward the sea by means of its current velocity. The water then has to move uphill, decelerate, and eventually come to a halt, thus limiting the volume that can be evacuated near the head of the estuary. Because this movement causes eddies and friction, the oscillating tidal movement can only be maintained when fresh supplies of energy are introduced into the system by the oceanic tides. Nor is all of the energy dissipated during each tide, otherwise the outgoing water would not have the energy to move uphill during the latter half of the ebb cycle and a noticeable imbalance between the High Water and Low Water amplitudes would result. There should also be a relatively large dissipation of energy otherwise the tide would not respond as fast as it does to the changing gravitational influences of the Moon and Sun.

The energy E1, of a layer of water with velocity v, density D, a surface area Sy km², dy metres thick, at elevation y, with the capacity of dropping a vertical distance h, and subjected to an atmospheric pressure p kPa, can be described as:

-> See the list of figures

When the water is moving from a deep reservoir through the turbines, the original velocity v₁, of the water in the reservoir, is negligible in relation to the current velocity v₂ that it has when leaving the turbines after the h metre drop. The atmospheric pressure p₁, affecting the water in the reservoir, will be on average around 101 kPa, with possible deviations of 4 kPa. However the pressure p2, in the turbines can be much lower because of the Venturi-shaped passage. The drop in pressure can account for the measured discharge coefficients of such orifices, which can considerably exceed unity. Generally velocities in such passages can not be higher than (2g · h)^0.5, but certain shapes allow underpressures and higher velocities.

After the water has dropped the distance h, the energy changes in character. The potential energy Ep, represented by the factor D · g · h, is transformed into other energy forms, such as kinetic energy Ek, or heat energy Eh; it can also be extracted as electric energy Ee. Consequently the energy distribution E2 after the drop h, becomes:

-> See the list of figures

Thus, the potential for electrical extraction will be:

-> See the list of figures

Neglecting energy gain due to pressure differences, and assuming negligible initial velocity v1, the energy that can be extracted annually (706 tides) with a single-effect operation can be set at:

-> See the list of figures

Note that no energy extraction is possible when v2 = (2g · h)^0.5. To extract the largest amount of electric power, the value of v2 must be kept as small as possible. If the combined working cross-sectional area of all turbines is X m^{2 2}, and extraction occurs during t seconds, the velocity , v2 will be:

-> See the list of figures

The velocity will decrease as the values of X and t increase. Because it is not possible to extract all of the energy during the relatively short interval of slack Low Water, sufficiently large discharge channels must be available. Obviously, it is not practicable to employ the total tidal range as an energy head.

Another interesting and important consideration concerns the value of N · y in (41) for Sy. When N · y is large relative to M, the potential for extraction of energy on the ebb flow becomes much more attractive than during the incoming flood flow because of the larger surface areas at higher levels of the reservoir. Double-effect extraction is less attractive (Larsen and Topinka 1984; Charlier 1982) because, although more time is available for extraction, the available head will be smaller, and very large sluiceways are needed in order to fill and void the reservoir of large volumes of water during shorter intervals.

Promoters of tidal power usually use the following simplified equations for calculating the potential energy extraction. Here it is assumed that the base level to which the water in a reservoir can fall is at elevation z, and that the total energy E_e can be derived by the integration of the following equation:

-> See the list of figures

where K = D · g · 10⁹ = 10.05 · 10⁹. When y drops from level a to level b, the value of Ee becomes:

-> See the list of figures

If the range of the reservoir is Y and a = Y/2, and b = z =– Y/2, then (51) becomes:

-> See the list of figures

For a single-effect operation during 706 tides, the annual output can be estimated at:

-> See the list of figures

Promoters may replace Y by the tidal range, and use twice the value, indicating a double-effect operation in which the reservoir is emptied simultaneously at Low Water, and filled again instantaneously at High Water. This yields a highly inflated value of the potential power extraction.

Greenberg (1987) has calculated the mean potential energy of the Minas Basin at 1.15 · 10¹⁴ J per tide. For a smaller area (850 km², east of Cape Split), Godin (1990) estimated energy output at 2.657 · 10^{14 14}J. The theoretical potential energy for this area, calculated using (52), and Y as the local amplitude of the mean tide, is 1.8 · 10¹⁴ J per tide. However, Charlier (1982) claims that a total annual energy of 169.5 billion kWh can be generated (this is equivalent to about 8.65 · 10¹⁴ J per tide) with a double-effect unit.

A single-effect power station lets the water flow through the turbines in one direction only, generally from the basin into the sea, thus utilizing the greater basin storage at the higher levels. Since the amplitude of the tide can vary between 60 and 140%, its mean potential value can be even higher. As the output is proportional to the square of the amplitude, the importance of a large tidal range becomes obvious (see Table 16). Tidal range is even more crucial in areas where the water surface area increases significantly. In practice of course all this energy cannot be tapped. This would require the reservoir to be filled to high tide level, and then almost instantaneously released when the tide is out.

Experts on tidal power development estimate that only about 25 to 30% of theoretical capacity can be realized. For example, the French tidal power station on the Rance River has a basin with a surface area of 22 km² and a mean tidal amplitude of 4.25 m. The theoretical annual energy output should thus be 1567 GWh. However with an annual output of 544 GWh, the efficiency is close to 35% thanks to refined computer software, which optimizes plant operation, including pumping and double-effect power generation. A similar system is operated by the Russians in Kislaya Bay, near Murmansk. It has a 1.1 km² basin and a theoretical annual output of 6775 MWh of which only 2300 MWh is realized. The Russian and French tidal power stations operate with reservoirs used as holding ponds to create large heads, and turbines to extract the energy.

Alternate Fundy tidal power proposals call for placing paddle wheels or egg-beater-type propellers in the flowing water, without confining the water to either reservoir or restricted channel. One possible location is the Minas Channel, where current speeds can reach 7 to 8 knots. A current speed of 8 knots can be generated by a head of 0.86 m. Theoretically, by building a dam in the Minas Channel, a head of 10 m can be created. This is 11.5 times the head that the paddle wheel can be subjected to under the fastest currents. Nevertheless, in 1916 the president of Acadia University, together with two members of its engineering department and the head of its business academy, formed the Cape Split Development Company. During summer, a survey was made of the topo-graphical and hydrographical conditions between Cape Split and Squaw Cap Rock (just offshore from the Cape). The maximum tidal current was clocked at 11 mph (4.92– 5.66 m/sec). A model was constructed of a tidal power contraption consisting of pairs of endless chains linked to concave vanes and led over sprocket wheels that would drive pumps. The idea was that the pumps would continually top up two 250 000 m³ storage tanks erected on the 100 m high Cape Split. Water would generate electric power utilizing machinery built in four open sluices and housed in the 120 m-wide gap between Cape Split and Squaw Cap Rock. Although shares were quickly issued, the company dissolved in 1929.

There were a few early success stories relating to tidal power worldwide. Records show that tide mills existed on both sides of the English Channel as far back as the 12th Century. Similar mills were also in operation along the coastline of eastern North America shortly after the first settlers arrived. Some of the mills served several purposes. In 1634, a cove and marshland north of Boston were dyked off. A three metre-wide sluice admitted tide water, which eventually powered two grist mills, a saw mill, and a chocolate mill. Most of these plants were abandoned when more convenient sources of energy became available.

A modest proposal for Fundy tidal power was advanced by K.E. Whitman in 1944 for a double-basin scheme using the Maccan River estuary as a head-water pond and the Hebert River estuary as the tail-water basin. Hicks (1965) identified seven possible sites between Cape Split and the mouth of the Bay of Fundy, and between Long Island, N.S., and Cutler, Maine. That same year it was seriously proposed to build a barrage through the 5 km-wide Minas Channel. The channel is 100 m deep at low water, and a suggested dam at this site would have a volume of over 75 million cubic metres! At that date, only the Fort Peck dam in the U.S.A. was more voluminous. To handle the volume of material needed in construction would require exclusive use for two years of loading facilities equivalent to all those available at the world's largest shipping centre, Rotterdam.

There have also been proposals based on the difference in tidal range on opposite sides of the Chignecto Isthmus. Tides in Cumberland Basin have an average range of 10 m, while 23 km away, in the Northumberland Strait near Baie Verte the average range is less than 5 m. Many people have had the mistaken idea that the Low Waters of these tides are at the same level as the Fundy tides, tides on both sides of the isthmus being measured from Chart Datum. The fact is that each tidal station has its own Chart Datum set at some particular elevation below Mean Sea Level. (Chart Datums are set at elevations so low that the tide at any given location will seldom if ever fall below it). Thus, and contrary to some proposals that have actually been advanced (see for example, Pogany 1958), the perennial economic problems of the Maritime Provinces can not be solved by simply digging a canal through the isthmus and installing a tidal power plant on it.

A more thoroughly researched report was prepared in 1945 for the Government of Canada (H.G. Acres and Co. 1946) concerning tidal power development in the Petitcodiac and Memramcook River estuaries. However this report, like that prepared by the Atlantic Tidal Power Programming Board (1969) for the Federal and Provincial governments, concluded that development could not be justified under the prevailing economic circumstances. A different conclusion was reached in the Bay of Fundy Tidal Power Review Board (1977). Out of 37 possible sites, the three most promising were in the Minas Basin, Cumberland Basin, and Shepody Bay. According to the report (which really focused on the results of market research), the largest, in the Minas Basin, would be capable of generating about 5000 MW; it was given preference over runner-up Cumberland Basin.

Many proposals have been advanced to try and tap the power of the Fundy tides (for details see Desplanque and Mossman 1998a). To date, however, only a modest experimental plant exists in the Bay of Fundy region. Installed in 1984 and located in the Annapolis River estuary, this small unit (20 MW) uses a "Straflo" combined turbine-generator to extract energy on an ebb tide, using a tidal range of 4.5 to 10.0 m. It may ultimately be the pilot for a vastly more ambitious undertaking to harness the great tidal ranges of the Minas Basin. In any event the construction of barrages in tidal waters is a formidable assignment. At La Rance, the French built the power units (total 250 MW) for the plant under dry conditions by placing cofferdams part way across the Rance estuary (Table 17). The placement of the cofferdams at La Rance succeeded, but not without a few tense moments. The relatively small power unit (800 kW) of the Russian plant was built into a floating caisson, 36 m long and 15.35 m high, which was then towed to the site and placed on a prepared bed. To smooth this bed the Russians sent some divers to the bottom equipped with hand rakes. It is extremely doubtful that such a procedure could be used in Fundy waters. A submerged tide gauge held down by pieces of railway track was placed in the upper portion of Cumberland Basin on 21

May, 1978, in 12 m of water. The current carried it away, and it was never relocated despite a thorough search.

Closing off parts of estuaries in the upper reaches of the Bay of Fundy would be a complicated undertaking, comparable but more difficult than the massive works carried out for the Deltaworks in The Netherlands. The latter are carried out through the cooperation of government personnel working together with contractors and a labour force experienced in this sort of work for many decades. The Deltaworks actually began with the construction of the enclosure dam and the first polder of the Zuiderzeeworks in 1927. Before that, the best scientific minds (among them H.A. Lorentz, Einstein's mentor) studied the implications of the closure of the Zuiderzee. The practical knowledge gathered during operations on both the Zuiderzee-and Deltaworks have established that undertakings of such magnitude require a sound logistical base in order to guarantee success. Planning a tidal power facility is not just a matter of placing a line on often outdated maps, and launching a public relations campaign. It involves plans that incorporate basic facilities such as sheltering harbours, and requires a sound infrastructure for constructing the facilities. Within the rigorous timetable ordained by the unforgiving tides, many new techniques would need to be developed and perfected.

"Fundy Tidal Power – Update ‘ 82" (Baker 1982) stated, among other conclusions, that: "Significant reductions in overall cost of a tidal power plant can be achieved by shortening the construction period and that should be one of the objectives of definitive design" (Baker 1982). In order to reach this objective it would be imperative that labour peace is guaranteed during the construction period. The "Update" also calls for the construction of 50 sluiceways and 64 turbine caissons, each 59 m long, 39 m wide and 46.25 m high, to be towed and placed on mattresses. This latter manoeuvre is an extremely delicate operation to be carried out during the periods of slack tide with split second precision, and in very close proximity of previously placed caissons. Any error of judgement will result in damaging collisions or time-consuming strandings of caissons. The construction and transportation of these 16-storey-high structures would require the ultimate of technical know-how. From the human perspective, these structures would be enormous, but in fact they are as delicate as oversized aquariums. They will have to be placed on platforms that will support the caissons evenly; otherwise internal stresses can play havoc with the structures and their contained expensive turbines or sluice gates. The successful construction and emplacement of such platforms would be a colossal engineering feat.

The transition from fall to winter in the Bay of Fundy is a time of stark contrasts. Huge amounts of ice can be formed in a few days of heavy frost. As we shall see, this can rapidly bring about dramatic change in the character of tidal estuaries in the Bay of Fundy (Jennings et al. 1993). On Wednesday, 10 December, 1980, the Sun was out, the wind was light, and the temperature hovered about freezing on the marshes bordering Cumberland basin (Fig. 43). No snow was on the ground and not a speck of ice in the estuary. A survey of ice conditions in the area, planned for the coming winter, was about to begin. The plan called for surveys on foot, on skis, and by helicopter throughout the period that ice would be present in the Bay. That day scientists from the Bedford Institute of Oceanography would be inspecting sites that the senior author had volunteered to visit during the winter.

By the following morning, as the temperature fell, the weather had changed significantly. On Saturday, 13 December, it was so cold and windy that it was dangerous to be out on the marsh alone. However on Sunday, 14 December, the Sun appeared again and conditions moderated. Five centimetres of snow covered the ground. In the morning, as the tide ebbed past Lusby marsh (LM on Fig. 43), one third of the ebb channel was covered with ice moving in the outside bend of the basin. Already formed were most types of ice common here in winter.

The drifting ice was mainly slush, or frazil ice, formed in open water areas termed "ice factories" (Knight and Dalrymple 1976), and made up of an unconsolidated mixture of needle-like ice crystals and sediment-laden water. There was also a good representation of pan ice, and cake ice. Pan ice, also present, is formed from accumulations of slush ice, frozen together in flat slabs up to 15 cm thick. Cake ice is considerably thicker, and probably forms as pan ice is jostled in fast-moving, ice-packed water. On its perimeter, cake ice picks up slush ice, creating an elevated ridge, or levee. The resulting basin-shaped central part can at least temporarily, hold silt-laden water. However the water usually seeps through the porous ridge leaving the silt behind. Collisions with other ice cakes evidently round the undersides of the cakes. Floating by too were a few scattered blocks of composite ice protruding 0.5 m and higher above the surrounding ice. On 14 December the only type of ice not yet in evidence was floe ice, normally formed where salinity and tidal energy are greatly reduced. Floe ice consists of frozen assemblages of all other types of floating ice, 50 to 100 m in diameter, that move restlessly with the tides, up and down the estuaries.

Following Desplanque and Bray (1986), the foreshore is here divided into three subzones (see also Fig. 44). The upper subzone is the vegetated high marsh that, in the upper reaches of the Bay of Fundy, is approximately 1.2 m below the highest levels that tides normally reach. Dominated by river-related processes, this subzone is not often covered with tidal water. During some winters, tides do not reach this level, in which case the high marsh escapes ice deposition (Dionne 1989). However, frozen crust, another major type of ice also present on 14 December, usually forms on the surface of the intertidal sediment. This "shorefast ice" results from the combined action of downward-freezing pore water, upward accretion of precipitation, run-off and, depending on the season, sea water (see also Desplanque and Mossman 1998b).

The landward limit of the high marsh is formed by dykes, uplands, or in the case of extensive salt marshes, freshwater bogs. On the seaward side a vertical scarp 1 m or more high separates high marsh from the middle subzone. In sheltered sections of the shoreline the upper portion of the middle sub-zone may be covered during the summer by Spartina alternifl ora (low marsh) and algae. Where rock is present seaweed may cling to it. Slippery mud, underlain by semi-consolidated material of similar grain size, covers the lower sections. The middle subzone has a relatively gentle gradient (approximately a 1 m drop in about 50 m), and its lower limit is also a scarp (Fig. 44) although not as high as the upper scarp. The lower subzone slopes toward the edge of the ebb channel, forming the thalweg (the median line of the channel) in Cumberland Basin.

On 14 December, 1980, the lower half of the middle zone was completely covered with cake and pan ice. Several centimetres thickness of silt had accumulated on some of the cake ice. On the upper half, a scattering of ice indicated the height of the tide during the preceding few days. The foot of the upper scarp had not been reached by the ice. In some places the lower subzone was covered with a crust of glazed ice beneath which water flowed toward the lower edge of the zone; in others, the ground was bare and unfrozen.

Several ice blocks stranded in the middle zone were constructed of 4 or 5 layers of ice of differing structure. Most layers were parallel, but some were oriented up to 45° relative to the others. The blocks, from 1 to 1.5 m high and 3 m square, contained much silt. However their buoyancy was very high because of the numerous air pockets. These sizeable blocks must have formed during the preceding three days because, as noted, ice was not present on 10 December.

During this three day period, the tides had not been large by Bay of Fundy standards; they ranged from a predicted 7.8 m + CD at Saint John on 10 December, to 7.3 m + CD on 14 and 15 December. However from this day on the tides became stronger, reaching a predicted height of 8.5 m + CD on 22 December. This semi-monthly inequality of 1.2 m at Saint John translates into an inequality of 2.0 m or more in Cumberland basin.

The weather remained cold. On 21 December it was unsafe to visit the site, but on 23 December we observed the effect of the increasing strength of the tides on the accumulation of shorefast ice forming along the shoreline in the zone between neap and spring High Waters. The middle subzone of the fore-shore was now filled with a thick layer of chocolate-coloured ice to 0.3 m below the level of the high marsh. It was impossible to reach the lower subzone because ice-covered crevasses between the cakes made the going extremely hazardous. Gradually the rising tide submerged the lower subzone and all types of floating ice covered the basin. Pan ice penetrated marsh creeks, becoming stranded on their banks.

Flat pan ice is commonly deposited on earlier-formed pan ice ferried in on weaker tides to freeze onto the banks. On higher parts of the foreshore stranded ice is left longer exposed to freezing air temperatures, and is more likely to become anchored to the banks (Sweet 1967). Commonly, ice may become bonded so strongly to a clay or silt substrate that it will not refloat when covered by a higher tide. In contrast, the connection between ice stranded on gravel or loose rock is rather more fragile. Sheet ice formed in the lower subzone of the foreshore moves up and down with the tides because the substrate cannot freeze during the shorter exposure to super-cooled air. This vertical movement of the sheet ice creates a bellows-like action between the ice and soft mud, promoting vigorous erosion.

In the upper reaches of tidal estuaries, sea water can bring in ice that gradually builds up high vertical walls. Forced into the ever-narrowing sections of the estuary by a sort of ratchet movement, the ice is unable to exit. Initially, ice walls are rather porous and irregularly-shaped accumulations of ice. However, passing tides charged with floating ice cakes smooth off the rough edges, filling the pores and leaving a film of silt-laden water to freeze into a smooth surface layer of ice. Before long the trapezoidal cross-sections of tidal creeks, with side slopes of approximately 1:3.5, are transformed into rectangular channels of much smaller cross-sectional area (Gordon and Desplanque 1981). This in turn reduces the tidal prism and eventually also the flow rate of tide in the estuary (see Fig. 45).

During winter, ice walls 5 m high can build up in a week or less. Typically, ice forms a levee that is higher beside open water than toward the bank of a tidal creek (see Figs. 46, 47, and 48). These figures illustrate that impressive ice walls can form even in years of moderate tides. The foot of an ice wall tends to be slightly lower than the level reached by the lowest High Waters occurring during the frost period. Most of the ice becomes stranded shortly after High Water. Slack water intervals in tidal creeks are brief, to the point of non-existence. They occur after the water level has dropped following High Water. At High Water the water continues to move into the estuary on its own inertia. This process delays filling of the upper parts of the estuary with tidal water. Thus at High Water the ice is still moving in, to become stranded shortly afterward. Ice that is lifted onto the top of the banks or onto previously stranded ice will likewise be left high and dry when the ebb sets in. Before the next High Water the chances are good that this ice will become solidly frozen to the underlying base and will not be refloated. If it does manage to break loose, this will occur on the rising tide and stranding will occur at some point further into the estuary. The process sounds mundane perhaps, but in actual fact it is a most remarkable exercise in ice block gymnastics. The performance seems surreal because the delayed release of ice blocks from the substrate beneath the rising tidal waters causes them to be released by buoyancy, as if possessed of life.

Ice build-up will be heavier during some winters than in others because of tidal conditions (see Fig. 49, Table 18). Peaks of perigean tides and spring tides coincide in cycles of 206 days. Two of these cycles last 412 days with the result that from year to year the especially strong tides occur 47 days later than in the previous year (412 – 365). During the first half of the 206-day cycle, the difference in height between neap tide and spring tide is decreasing, or below average, allowing a lesser ice wall build-up. Conversely, during the second half of the cycle, tides will gradually rise to higher levels during the week before the perigean tides. One of the key ingredients in heavy ice build-up is thus the timing of the greatest difference between neap tide levels and spring tide levels. This occurs one or two months before the perigean and spring tides combine to form the strongest tide of the cycle. Thus, when the latter half of the 206-day cycle occurs during the frost period between December and the end of March, one can expect the greatest build-up of ice walls. The formation of high ice walls in the Bay of Fundy was first recorded by Henry Y. Hind on or near 25 April, 1875 (Hind 1875). One hundred days earlier, the spring tide will have been at its minimum. Thus, the latter half of the 206-day cycle fell during the first period allowing ice to accumulate on the strand a month early, with accumulation probably peaking toward the end of April.

Naturally, the ice walls on either side of an estuary converge inland. Near the upper end of the estuary they will almost touch, leaving a strongly reduced channel in which freshwater runoff takes over the main role in channel-shaping. The drastic reduction in cross-sectional area can lead to flooding during a sudden winter thaw or spring break-up. The channel may then become choked with freshwater runoff, and/or ice. Certain rivers, such as the Salmon River near Truro, N.S., seem more prone than others to such flooding.

The development of ice walls can contribute to disastrous problems in tidal waters. In 1965 a fertilizer plant was built at Dorchester Cape, N.B., near the confluence of the Memramcook and Petitcodiac River estuaries (see Fig. 43). In order to handle the bulk material to and from the plant, a wharf was constructed nearby. Because of the large local tides, a $2.5 million floating dock was constructed in the form of a 25 m × 90 m × 7 m concrete caisson connected to a wharf and concrete platform on shore by means of a bridge and connecting arms. At this location (see also Fig. 3) the average range of the tide is over 11 m, increased during large tides to more than 15 m. Since trucks had to move over the bridge the slope needed to be maintained within certain limits. Hence the distance between platform and caisson was substantial. Also, the tidal channel was dredged, allowing the caisson and ships tied to it to move freely up and down with the tide.

The facility, named the Port of Moncton, was officially opened on 24 November. Winter followed the departure of the first ship. An ice wall developed along the shoreline, and worse, between the platform and the caisson beneath the caisson arms. When this wall became sufficiently high and strong, it blocked the free movement of the connecting arm and bridge. The bridge lifted out of the water, the connectors buckled and broke, and the caisson drifted away, stranding on a nearby silt flat. The entire operation was subsequently shut down. Ironically, predicted and observed tides for November, 1965 were low, so that ice conditions might have been a lot worse!

What are the ecological implications of ice in the upper reaches of the Bay? At one time it was believed that ice moving from the high marsh during winter was the main agent responsible for transporting organic matter from the marsh into the Bay. However during most winters, ice is unable to move onto the high marsh. Thus the organic detritus of high marsh vegetation survives until spring only to disappear beneath fresh vegetation in the same manner as non-cut grasses on upland meadows, presumably by microbial decomposition.

Decaying high marsh vegetation is commonly deposited along the dykes in thick accumulations (Nova Scotia Department of Agriculture and Marketing 1987). Hydraulic conditions on the marsh are such that they inhibit most if not all seaward movement of organic debris; flow is too gradual at these shallow depths to accomplish transport.

On the low marsh, exposed vegetation is crushed and frozen into blocks of ice. Subsequently these blocks may be lifted by the spring tide, pulling the encased vegetation from the marsh. Removed to higher elevations they will remain stranded until the ice thaws, leaving the vegetation behind. In tidal rivers, ice walls begin to collapse by the end of March and most ice disappears except for shorefast ice that may persist until late April (see Figs. 50, 51). Stranding of large blocks of composite ice on the mud flats, acting in concert with the bellows action of sheet ice hinged to the shoreline, may cause the complete reworking of the soil under the sheet ice during the late winter and early spring. In late spring some tidal flats resemble plowed fields with great scars left by chunks of moving ice. Any form of life in this mudflat ecosystem survives only under a great deal of stress (Gordon et al. 1985). Macrofaunal diversity is very low. In addition to organic secretions of diatoms, most of the productive biomass derives from three species (Hicklin et al. 1980): a bivalve (Macoma balthica), a polychaete (Heteromastus fi liformis) and an amphipod (Corophium volutator). The amphipod, so important to migrating bird life in the Bay, is widely distributed over the mudflats. Its survival in winter is doubtless predicated by its ability to burrow well beneath the zone of scour and erosion. Since ice scouring is most intense on the outer portions of mudflats, the role of shorefast ice may afford an important protection to these species in the inner several hundred metres of mudflats at the mouths of estuaries. Corophium volutator may also occur on exposed tidal flats with no ice.

With approaching spring, flows of freshwater and tidal water progressively undercut the basal portion of ice walls and ice blocks. In places the ice is sculpted into dark-coloured, ephemeral toadstool-shaped formations charged with concentrations of estuarine mud as high as 18g per kg (i.e., 18 wt %) mud. By early April the thinner ice will have disappeared, but ice walls may persist until month's end (Fig. 52). At their base, huge accumulations of silt occur in sections sheltered from strong currents. It is remarkable that the ice fields are in the same locations where silt built up after construction of the Petitcodiac River causeway in 1968. Causeway construction promotes silt accumulation, with concomitant depth reduction of tidal rivers, and dire results for much aquatic life (Daborn and Dadswell 1988). With hindsight it can be seen that the river and its ice accumulations act like an enormous hydraulic sluice. Thus, whereas ice walls and associated phenomena may prove hazardous to man-made constructions they serve very important natural purposes. In the case of ice walls, their smoothness resists further narrowing of the channel, increases the net ebb and river flushing currents, maintains river depth, increases the surface slope seaward, forces the salinity intrusion seaward, and obstructs massive influx of ice from the estuary mouth.

There are many lessons to be learned from a consideration of winter conditions in a tidal regime as complex as that of the Bay of Fundy. Intertidal ice is an extremely active agent in diverse environmental processes, both physical and biological, in this macrotidal region.

From a strictly engineering perspective, the sequence of tides, temperatures, and wind velocities must be carefully evaluated in estuaries in northern regions subjected to a large tidal range. Floating structures attached to an estuary bank are at risk due to build-up of ice walls. So too, are bridges, where ice build-up increases the size of bridge piers, resulting in reduced cross-sectional area at the crossing (Figs. 53, 54). Construction designed to control flooding of marshlands must consider the downstream channel changes from trapezoidal to rectangular as ice factories swing into production. Shorefast ice can greatly inhibit the workings of one-way drainage devices, whether mechanical (flap gates) or electrical (gate slots), on hydraulic structures that are installed without due regard to winter conditions.

Concerning megaprojects like tidal power production, it is quite certain that a tidal power project would result in major changes in the dynamics and distribution of ice in an estuary. Reduction of tidal energy would promote water column stratification, which would extend the ice season. Sheet ice could expand near tidal rivers and by excluding extreme tides, the development of shorefast ice would be retarded. Not least, accumulation of drift ice at the barrage would need to be considered at the early design stage.

February is usually the time of heaviest ice build-up. Ice jams (dams) are most likely to occur in the upper part of an estuary during a period of low temperatures and spring tides. Conditions in the Bay of Fundy will be greatly exacerbated by strong, prevailing southerly to southeasterly winds that might coincide with a rapid thaw and heavy rains. Doubtless the most important factor affecting intertidal ice conditions in the Bay of Fundy is the unusually high tides. This results in several conditions not commonly encountered in other regions. Thick ice walls may be unique to Fundy estuaries due to the pronounced variation in the elevation of High Water during the spring/neap cycle, and the prolific High Water stranding of drift ice. Also unique, as we have seen, is the substantial variation in elevation of extreme (extraordinary) tides over longer periods than the spring/neap cycle. This factor affects the extent to which shorefast ice develops, and the degree to which the high marsh areas are influenced by sea ice. Hind (1875, p. 193) elegantly described the behaviour of ice in the lower zone of a Fundy estuary: "The appearance of an estuary in the Bay of Fundy at any time in midwinter presents some singular and striking phenomena, which may contribute to our knowledge of the manner in which different agents have assisted in excavating this extraordinary bay, and are now engaged in extending its domains in some directions and reducing it in others." At the time, his concern was with the potential impact of ice on the Baie Verte canal, proposed to link Cumberland Basin to the Northumberland Strait. Although this project was never completed, various large scale construction projects and a multitude of smaller scale coastal management and development schemes will continue to merit quality time applied to the task of understanding the dynamics and environmental effects of intertidal ice.

In the Bay of Fundy region changes to shorelines are due primarily to tidally-driven processes of erosion, and to changing sea level through geologic time (Shaw et al. 1994; Stea et al. 1998). Because of its economic importance, the rise in sea level since initiation of the last deglaciation about 15 000 years ago, is subject to intensive study. In eastern North America generally, and in the Bay of Fundy in particular, abundant and diverse lines of evidence indicate that regional differences in sea level reflect postglacial isostatic compensation. A vast literature exists on this subject and debate continues concerning a link between a possible accelerated sea-level rise in response to a global warming (Peltier 1999). Here we focus on the nature of the dynamic interaction between land and sea as evidenced by erosion and the Holocene history of sea-level change in the Bay of Fundy.

Topographical maps give the heights of mountains as measured from mean sea level (MSL). The same maps are also quite definite in showing the location of shorelines. This assumes that MSL is permanent and known with a great degree of accuracy. Unfortunately MSL is a rather elusive concept. It certainly is not at a permanent elevation. The landmasses are ever so slowly bobbing up and down, and the oceans are not always filled with the same volume of water. Tectonic and isostatic processes contribute to an uneven sea-level rise along coastlines.

The future duration of the current interglacial is uncertain. During glacial maxima, large portions of the continents ringing the Arctic Ocean were covered with massive sheets of ice. The largest accumulation of ice during the last glacial advance occurred 20 000 to 18 000 years ago. In North America this last glacial is called the Wisconsin glaciation, and in Europe, the Weichselian (or Würm). At that time, so much water was locked up in ice that sea level globally was 100– 130 m below present MSL (Fig. 55). Across the Scotian Shelf, for example, this is indicated by a well-developed submarine terrace at a depth of 115– 120 m (Fader et al. 1977). In the Bay of Fundy, however, the terrace occurs at a depth of 37 m, a feature of fundamental importance to the distribution of surficial sediments in the Bay (Fader 1989).

Centered over Hudson Bay, the Laurentide ice sheet blanketed most of Canada and extended far south of the Great Lakes. It crossed the Bay of Fundy and Gulf of Maine, approaching the edge of the continental shelf (van de Plassche 1991). End moraines and glacial outwash peripheral to the ice sheet, and even some drumlins that formed beneath it, now form many of the banks and shoals along the Maritime and New England coastlines (Grant 1989). The rate at which the ice has retreated since the last deglaciation began is quite astonishing considering that the nearest remnant is now located on Baffin Island about 3000 km north of the Bay of Fundy. Most melting was over by about 5000 years ago (Lambeck et al. 1990). As discussed below, the rate of retreat was irregular rather than constant, due in part to eustatic change and spatially varying crustal motions, even during the last 1000 years or so (Shaw and Ceman 1999; Shaw et al. 2002). The current rate of rise of relative sea level is therefore not necessarily tied to an anthropogenic greenhouse effect. In any case, the sea was able to return in the western sections of the Bay of Fundy about 13 000 years ago.

The weight of an ice sheet is sufficient to depress the Earth's crust significantly. Compensating for this downward movement, the land and sea bottom in front of the Laurentide ice sheet was forced up into a bulge more than 100 m high and 60 km wide. With retreat of the ice, recovery takes place and the peripheral bulges gradually disappear. However, this process requires thousands of years, and in the Maritimes the recovery is still going on (Fig. 55). Thus, bulges that may have formed small islands off the Nova Scotian coast and a land bridge between southwestern Nova Scotia and New England, have gradually undergone submergence together with large portions of the mainland itself. For example, at the peak of glaciation, Georges Bank was 40 to 50 m above sea level. Yet 11 000 years ago conditions were such as allowed growth of salt marsh upon it (Stea et al. 1994; 1998). About 12 500 years ago the area of Portland, Maine was at least 49 m below sea level, and the sea covered nearly 23% of that State as far as Bingham, presently 100 km from the sea. Coastal areas along the Bay of Fundy also emerged as the ice retreated. Well defined former sea level stillstand indicators, such as raised beaches and nicks eroded in shore-facing cliffs, set the marine limits about 42 m above present MSL along Digby Neck and numerous other Bay of Fundy locations.

From 11 000 to 5000 years ago, the land rebounded from the depressing effect of the ice (Grant 1989), and rose faster than sea level. Although over several intervals the rebounding effect reportedly slowed (Scott and Collins 1996; Stea et al. 1998), allowing reversals to occur, the tendency has been for the rate of sea-level rise to overtake that of the land. Consequently, submergence has occurred at a faster rate than that due solely to eustatic rise. Eustatic rise in sea level amounts to about 10 m in the past 7000 years. It is important to note that the inferred eustatic rate of sea-level rise of about 1.3 mm/yr is occurring globally, regardless of the vertical motion of any particular shoreline (Schneider 1997). This is the direct consequence of thermal expansion of ocean water combined with additional water mass derived from shrinking glaciers. Thermal expansion of the oceans is in itself is a very important factor, and although estimates in the literature range between 0.5 and 2 m/°C, the best estimates indicate that a 1°C rise in ocean temperature would result in a rise in MSL of about 60 cm.

Presently, sea level is rising all along the Atlantic coast of North America, but nowhere more rapidly than in Nova Scotia (Scott and Stea 2000; Fig. 56). This rise has been continuous for the last 7000 years at a rate of 20– 30 cm per century, accelerated between 5000 to 4000 yr BP when rates reached up to a metre per century (see Scott and Greenberg 1983). Tidal observations over the past 60 years confirm that the rate of rise remains about 3 mm/yr. For example, at Saint John, N.B., the measured rate is 3.6 mm/yr. At Halifax, sea level has risen 35 cm in the last 100 years; 20 cm is attributed to crustal subsidence, the balance is due to eustatic rise. Comparable rates occur along the coastline of the U.S.A. down to Florida. Along the northeast coast of Maine the rate exceeds 1 cm/yr. At places along the St Lawrence River estuary and the north shore of the Gulf of St Lawrence, the sea levels are not changing at present, although north of this line, for instance at Churchill, Manitoba, on Hudson Bay, land continues to rise relative to sea level (Hill et al. 1996).

While some areas were rebounding upward, other areas on the peripheral bulge were compensating for this movement and dropping in elevation (Scott et al. 1987). For example, at Lunenburg, N.S. sea level 7000 years ago was about 30 m lower than present, whereas peat that formed at the same time on what is now Georges Bank, can be dredged up from depths of 45 m.

During the last glacial maximum a glacier is believed to have occupied the Northeast Channel in the Gulf of Maine (see also section 4.3). The terminal moraine of this glacier may have served as a land bridge for the numerous species of plants that are native to southwestern Nova Scotia and southern New England but which are not found in the intervening coastal areas. Perhaps the Paleolithic people who settled in Nova Scotia 10 500 years ago used this land bridge to reach their hunting camps, the remains of one of which have been found at Debert, near Truro, N.S. (Grant 1975). Doubtless the existence of a land bridge restricted the free exchange of water between the Atlantic and the Gulf of Maine. The latter probably had a larger surface area than now because a large part of New England and the Maritimes was still covered with water. Today, with an average tide, about 300 km³ of water flows in and out of the Gulf of Maine during a 12.4 hour period. If sea level was 30 m lower than today, the Northeast Channel in its present configuration would still have more than 75% of its cross-sectional area available for this flow. From a hydraulics point of view this restriction would have made little difference in the tidal characteristics of the area. However, if the channel was restricted by glacial material, and the surface area of the Gulf of Maine was much larger, tides of present strength could not possibly have been generated in the Gulf of Maine and the Bay of Fundy (Grant 1970).

Some very interesting features of tides and sea-level history in the Bay of Fundy, as discussed earlier (section 5.2), are substantiated in the geological record of tidal marshes. Marshes began to build up when tidal waters first covered coastal lands (Shaw et al. 1993). In the upper reaches of Cumberland Basin nearly 20 m of red clastic marine sediment have accumulated over the basal glacial till. Near Aulac, N.B., the marine sediments overlie a 6 m-thick layer of peat (Chalmers 1895). Given a compression ratio of about 3:1 or more for the vegetation to peat transition, the peat represents a formerly much thicker unit. The overlying silt will therefore also have been deposited at a higher elevation than it is now. Information from the drilling of wells has established that the lower limit of the clastic marine material is at a depth of 17– 18 m below the present land surface. This sets the lower limit of that material at about 12 m below present MSL. Thus constrained, the marshes could have begun to form 6000 to 7000 years ago when tidal ranges and currents became large enough to move material from eroding coastlines and deposit it in the upper reaches of the Bay.

During the summers of 1980 and 1981 students of the Free University of Amsterdam studied the development of marshes in Cumberland Basin (Dekker and Van Huissteden 1982). About 600 boreholes were made with an average depth of 6 m and a maximum of 15.1 m (Noordijk and Pronk 1981). One borehole, reaching 8.9 m below present MSL, was still in red marine sediment. The texture and structure of this material are similar to those of recently deposited clastic layers. No indication was found that the tidal regime at the time of deposition was any less vigorous than now. Marshes that are formed in areas with weaker tides have generally a much greater content of organic material. The same is true of tidal marshes less often inundated with salt water.

As we shall see (section 9.6), tidal marshes along the upper reaches of the Bay of Fundy are generally built up to a level about 1.2 m lower than the level reached by the highest astronomically-caused tides. At this level, relatively few tides are able to cover the marsh with silt-laden water. Each additional layer of silt progressively decreases the potential for more buildup. Combined with tidal cycles, marsh build-up is intimately related to the rate of coastal submergence. When sea-level rise is slow, marsh build-up is also slow, resulting in fewer inundations and prevalence of freshwater conditions on the marsh, especially in areas farthest removed from the coast and main creeks. Plant growth in basin areas on the marsh will resemble fresh marsh vegetation and will leave more organic peat-like material (Shaw and Ceman 1999).

Studies of soil genesis in the marshes reveals that there are 5 to 7 layers of organic (plant) material at particular stratigraphic levels throughout the marsh areas. This has been interpreted in terms of successive periods of increased and decreased rates of relative sea-level rise (Lammers and De Haan 1980). This phenomenon indicates intermittent temporary slowdowns in the process of submergence. Similar observations have been made in coastal areas of the northeastern United States, and of France, Belgium and The Netherlands along the North Sea. If, as suspected, intermittent submergence is a world-wide phenomenon, postglacial sea-level rise is a punctuated process, not necessarily linked to an anthropogenic greenhouse effect as suggested by Van de Plassche et al. (1998).

Temporary slowdowns of coastal submergence means that certain sections of the coast with exposed bedrock were exposed longer than others to wave action and consequent erosion. Tidal action during such periods will have remained the same except that the eroded material could not have been deposited on the marshes to the same extent as previously, but must have been choking tidal estuaries and creeks. This would lead to even less tidal water reaching the upper marsh areas. However, when coastal submergence was renewed, this material was readily available to be moved on the marsh, leading to an interval of renewed and vigorous marsh growth (Grant 1975). The geological record preserved in the tidal marshes is quite explicit in this respect (Figs. 57A, 57B).

Initial rise in postglacial sea level is signalled by the presence of fossil forests at several locations in the upper reaches of the Bay of Fundy. Typically, fossil tree stumps are rooted in glacial till at the base of the marine sequence; remnants of younger fossil forests also exist at higher levels. Among the intertwined network of roots, the peat-like consistency of the original forest floor is commonly well preserved. A good example occurs on the shore near low tide at Fort Beausejour on the Nova Scotia-New Brunswick border (see Fig. 3). The levels at which these stumps occur is approximately 1.2 m below present MSL. In 1983 more than 70 stumps were exposed, some with a diameter of over 1.5 m. Although most disappeared during winter of that year, fresh examples, including several blow-downs, have since been exposed. Pollen analysis of the peat indicates that the forest was a moist, rather patchy mixed-forest with hemlock, pine and red maple, which until its destruction by the rising waters was not disturbed by salt water (Grant 1985; 1989). The tree stumps have been radiocarbon dated at 3300– 3700 years BP. The maximum tidal range at Fort Beausejour is now about 15.6 m. Allowing 1.2 m for extraordinary storm tides, the stumps are approximately 10 m below the reach of a storm tide. This is compatible with a rate of submergence of about 3 mm/yr.

Although the section at Fort Beausejour is slumped, the relative sea-level curve obtained is reportedly comparable to those documented from several other locations around the Bay of Fundy (Scott and Greenberg 1983, p. 1556; see Fig. 57B). Most show a distinct break in the rate of sea-level rise at about 2500 BP.

The more than 11 000 km coastline of the Maritime Provinces exceeds a quarter the circumference of the Earth. Of this, the coastline along the Bay of Fundy totals about 608 km in New Brunswick, and 805 km in Nova Scotia. Along much of this length, erosion-prone sandstones and conglomerates are exposed; along other stretches, "harder" rocks, like the basalt along much of the Nova Scotia coastline adjacent to North Mountain, offer stiffer resistance. Most resistant are massive igneous rocks, metamorphic gneiss and quartzite, and limestone. All are present in this geologically complex region. Erosion-resistant rocks form the headlands and capes, and each coastal section responds in its own way to the erosive powers of the tidal waters. Thus, one encounters 100 m-high cliffs, low bluffs, beaches, mud flats, tidal marshes and sand spits at various places along the coast, depending on local geological and hydrographic conditions. The difference in tidal ranges also leads to significant differences in the erosion of littoral materials.

In general, beach slope gradients depend upon the cohesion of the shore and bank materials and on local wave energy levels. Coastlines consisting of hard rock, such as the North Mountain basalt, are not easily broken down by action of wave and weather. However where bedrock consists of relatively friable material such as sandstone, the slope of the foreshore is formed of debris dislodged by wave action from banks and cliffs along the shore. Here the slope of the foreshore is uniform and gently dropping from 1:10 to 1:200, depending on local conditions.

Due to continuing submergence, a small section of back-shore, previously too high to have been exposed, becomes subject to wave action. Furthermore, the entire foreshore will be deepened so that wave energy will not be dissipated to the same extent in passing over it. Consequently wave attack on the shoreline becomes more vigorous. Equilibrium can be restored only if the foreshore is raised by debris removed from the freshly exposed backshore. Given the prevailing rate of coastal submergence and the evolution of the foreshore, the overall result is that the shoreline will have a tendency to retreat at a rate of 0.06 m to 0.8 m/yr depending on the slope of the foreshore and the resistance to erosion of shoreline materials. If the prevailing rate of submergence has been constant over the past 10 000 years, then sea level will have risen approximately 40 m since people first arrived on the shores of the Maritime Provinces, and 1.2 m since Père Pierre Baird described Saint John Harbour (see section 7.3).

The level of tidal waters lapping the shores of the Maritime Provinces is quite variable. Outside the Fundy region, tides are small, as in the western section of the Northumberland Strait (mean tidal range 0.6 m), or in the Bras d'Or Lakes, where the mean tidal range is 0.1 m. (The smallness of the range in the Bras d'Or Lakes can be brought home by the observation that changes in barometric pressure can cause differences in water levels of up to 0.3 m – Desplanque 1980; Petrie 1999.)

Wave energy is concentrated near the surface of the water. During high water the foreshore is covered by a significant depth of water and a far larger percentage of wave energy reaches the shoreline than when the tide is at low water. As noted earlier (see section 5.4.1), the zone of the shoreline near the water surface will be the most heavily subjected to wave action. Where water levels are rather constant, efforts at shore protection can be focussed in a limited zone. Conversely where macro-tides prevail, the problem of protection is complicated, and never more so than during perigean spring tides (Wood 1976).

In the Bay of Fundy tidal range in absolute figures is high. In the Minas Basin, tides have a mean range of 12 m and vary between 7 m and 16 m. Thus the zone in which the water surface meets the shoreline nearly 30% of the time varies between 2.8 m and 3.5 m above MSL during small tides and between 6.4 m and 8.0 m during large tides (Figs. 25, 26 and 27). This condition is clearly shown in terms of the amount of material eroded over a given tidal range. The exposed conglomeratic sedimentary rocks at Hopewell Cape, N.B., (see Fig. 29) provide an excellent example (Trenhaile et al. 1998; Desplanque and Mossman 2001a, 2001b). If any attempt is made to protect this sort of shoreline, then the zone requiring reinforcing needs to be at least 5 m high to have any effect. Attempts to protect shorelines are diverse: in many instances these prove to be expensive undertakings. Here we consider briefly one particular geological process all too often overlooked: the hydraulics of groundwater flow.

Due to density differences, fresh water can only move beneath the shoreline if it has an overpressure, because it needs to displace the heavier salt water. Should this overpressure dissipate because of friction and the pressures between salt and fresh waters become equal, then no transfer of water will take place. The plane where this situation exists is called the interface and its location is determined by the Ghyben-Herzberg ratio, which describes the static relationship of fresh ground-water and sea water in coastal areas (see Fig. 58).

Fresh water prevails on the land side of the interface, salt water on the sea side. Because fresh water loses pressure gradually, the interface rises in a seaward direction and interacts with the sea floor at the elevation where the column of salt water exerts a little less pressure than that of the fresh water moving along the interface. This means that the excess groundwater has a smaller zone of exit into a saltwater body than into a freshwater body, and is much more concentrated, typically in wet areas near the shoreline. This concentrated outflow may cause quicksand and other soil/rock- weakening conditions near the shoreline where wave erosion is focussed (Figs. 59A, 59B). Furthermore, the fresh water exiting near the shoreline will expand upon freezing, promoting more rapid weathering conditions than might otherwise be the case. Also, since this groundwater is likely to be warmer than the air temperature during winter, many freeze-thaw cycles are possible during that season. In order to inhibit erosion of a particular section of shoreline, care needs to be taken to avoid local build-ups of freshwater pressure. Protective cover employed to this end should not be so permeable as to allow material to be easily dislodged. Incorporation of an artificial filter beneath the protective layer should be considered.

Protection against coastal erosion is not only an expensive undertaking, but also involves high risks. The motto of a wise coastal engineer should read: "Be sure to put off to the future what you do not absolutely have to do tomorrow". The fight against coastal erosion is a struggle against strong and persistent forces in nature. Interaction of land and sea in the Bay of Fundy is a dynamic relationship. It is most obvious in recent cliff falls and in rapidly retreating bluffs as the sea advances inexorably and, at times of storm surge, disastrously. However, even on a decadal scale, continued sea-level rise is not always obvious upon cursory inspection. Measured in differences of several mm/yr, it is revealed in various sorts of tidal records, in bore hole data, and from radiometric and relative age data. At the head of the Bay of Fundy it takes the form of a startling paradox seen in the continued build-up of tidal marshes, as coastal submergence and erosion strive toward an equilibrium. Here, despite extensive dykelands, the marsh grows upward as sea level rises, yet recedes landward in the face of relentless erosion.

On a geologic timescale estuaries are ephemeral. The same is true for tidal marshes, because an extensive marsh occupies a natural estuary that has largely been infilled with sediment. A legacy of the ups and downs of the land surface in response to regional tectonics and climate, the system of tidal marshes extends around the Fundy shore, from Argyle and Chebogue south of Yarmouth, N.S., to the Musquash and Manawagonish marshes southeast of Saint John, N.B. (Fig. 60). They formed after the latest ice sheet retreated and mean sea level rose about 100 m to its present level, allowing tides to develop ranges and currents that could transport the sediment onto the marsh surfaces. Dyking was undertaken in the early 17th Century by the first French settlers, reclamation focusing along Cumberland Basin at the head of the Bay of Fundy. Johnson (1925) recognized three distinctly types of marshes along the eastern seaboard of North America:

1. Coastal plain marsh. This type occurs south of New Jersey and is abundant from Virginia to Georgia. Related in structure to Fundy type marsh (below), it is developed in a regime of tides of much less amplitude. For this reason the marshes are flooded by each tide. This is because the marshes can only build up to a certain vertical elevation below the highest local tides and the difference in height is large relative to the variation in the magnitudes of the local tides. However the soils have a higher organic matter content and a different surface aspect than those of the Fundy marsh.

2. New England marsh. Found from New Jersey to Maine, it differs from the Fundy marsh in that it is essentially a deposit of salt marsh peat with variable amounts of silt. The tides in this region too have a much smaller amplitude than those associated with the Fundy type. The Nova Scotia marshes in Yarmouth and Digby counties are of this type.

3. Fundy type marsh. These marshes are built in a regime of tides with large amplitude and strong tidal currents. The high silt-carrying capacity of the Bay of Fundy currents cause marshes to build up with mineral matter, which is removed from the bottom, the banks, and the exposed shorelines of the Bay. Incoming tide water has a reddish colour and a silt content up to 2%.

We define a salt marsh as a low-lying flat area of vegetated marine soils that is periodically flooded by saltwater inundations, generally caused by tides, to such a degree that only plants adapted to saline conditions can exist. This definition excludes: a) unvegetated tidal mud, sand flats and gravel bars; b) sand and shingle beaches; c) dunes; d) dykelands, being former salt marshes, presently protected from tidal inundations by dykes and aboiteaux; e) freshwater marshes that exist between the larger salt marshes and the upland areas. The distance from the sea is so large that salt water cannot reach this zone within the limited time period available even during the highest tides.

Since a marsh creek needs a certain area of marsh to maintain itself against sedimentation, there will be margins of salt marsh free of creeks, along uplands and dykes, and between creeks.

The Fundy-type marsh, although built in a regime of tides with large amplitude and strong tidal currents, rarely experiences overflow. Tides that rise high enough to flood marshes in the upper reaches of the Bay, move larger volumes of water at higher velocities and greater eroding and sediment-carrying capacities. During large tides, High Water at the mouth of the Bay may extend 2 m higher than High Water during small tides. In the upper reaches, this difference can exceed 4.5 m. However, in the upper reaches of the Bay of Fundy the sea floor is too high to allow a Low Water. Therefore, it is better to employ the terms "large" and "small" tides rather than refer to tidal range, the difference between Low and High Water.

In general, tidal marshes in the Bay of Fundy area are built up to a level 1.2 m lower than the highest astronomical tides. Thus, only large tides can reach marsh levels and only very large tides are able to cover the marshes. For instance, on average, only 52 tides per year (standard deviation, SD, of the variation in tidal occurrences = 12) will be able to reach above the level of the remaining undyked tidal marshes around Cumberland Basin (see below). The number of tides that are able to cover the marsh to a depth of 0.5 m or more is 11 (SD = 5). These peak tides come in sets at monthly intervals, and a group of these sets appear at intervals of 7 months, 4.53 years and 18.03 years. Consequently, for long periods there will be no tidal flooding of the marshes. Thus Fundy tidal marshes can not be viewed as natural hatcheries or nursery areas for saltwater fish (Gordon et al. 1985). However when an overflow does occur, silt content of the water tends to be very high. Silt deposition on the marsh depends on the frequency of tidal flooding and the silt content of the tidal water. In the upper reaches of the Bay, the silt content is much greater than near the mouth of the Bay, although the frequency of tidal flooding is much less. Which factor is of greater consequence is open to debate. Sediment deposition upon the marsh causes a rise in elevation of the marsh in the same order as the rate of submergence of the region (Allen 1995).

In summer the soil can dry out, producing mudcracks up to 5 cm wide and 40 cm deep, a geological feature usually taken to signify prolonged aridity. The vegetation on such areas must be able to tolerate occasional flooding with silt-laden salt water, alternating with long dry spells (Gordon et al. 1985). Few species of plants are able to do this. Spartina alternifl ora, a pioneer in this environment, is able to withstand the effects of frequent flooding: however even this species has its limitations, and is able to grow in tidal creeks only to a level one metre below the marsh surface. The lower parts of the creek are bare of plant life. On the flat, high marsh another species, Spartina patens, is short and dense, promoting its requisite moist soil conditions over reasonably long intervals of time. Unfortunately neither species of Spartina contributes much fertility to the marsh, for they are short of plant nutrients such as calcium and phosphate. Nor is fertility enhanced by the silt-rich tidal marsh soil. These offer far less in terms of ion exchange capacity than, say, the calcium-rich clays of western Europe marshlands, or for that matter, New England marshlands (Desplanque 1952; 1985). As we shall see, intertidal flats of estuaries may vary seasonally, storing clay and silt during the summers and freeing them to winter waters (Amos 1995a, 1995b).

The apparent contradiction that marshes in areas with strong tides are less often covered with sea water than those where tidal amplitudes are low, is explained by reference to Table 19. Here, the height of High Water has been measured from Chart Datum, the average of the lowest predicted annual low water level over an 18.61 year period. Mean Sea Level (MSL) in the Maritimes has been rising relative to the landmass during the past several thousand years. Using all the hourly readings taken from 1927 to 1975, the variable distance between MSL and Chart Datum (CD) at Saint John, N.B., can be expressed as:

-> See the list of figures

where Z0 is the vertical distance that M.S.L. is higher than C.D. on 1 January of the year Y.

Using the above relationship, as of 1 January, 1956, the value of Z0 was 4.270 m or 14.010 ft (MMRA. 1950– 1965). In Table 19, this value is used to estimate tidal amplitudes over a ten year interval. These data were used to calculate the S-shaped curve of Fig. 61, which shows the number of tides per year that will reach or exceed certain levels above Mean Water Level. According to the Canadian Tide and Current Tables, Mean Water Level (MWL) is defined as the height above Chart Datum of the mean of all hourly observations used for the tidal analysis at that particular place. At Saint John, N.B., there is most likely a changing difference in elevation between the MWL and the Geodetic Datum; hence MWL is used because it is linked to the local tidal movements.

Figure 61 is an example of what is known as a combination chart or nomograph, a type of chart widely used in engineering (Davis 1943). It does not have a horizontal axis (abscissa), employs a natural scale (rather than logarithmic), and displays the results of calculations which combine various functions (i.e., multiplication or division with addition or subtraction) based on observed tides at Saint John, N.B. The strength of the tides at several other stations in the region are represented in Fig. 61 by lines, the slopes of which are determined by the distance from the reference point (0 km) of Bar Harbor, Maine (see also Desplanque and Mossman 1998b).

From Bar Harbor, Maine, to the head of the Bay of Fundy the range of the dominant semidiurnal tides increases in magnitude exponentially at the rate of about 0.36% per kilometre. This rate is equivalent to 3.66% per 10 km, 43.2% per 100 km, etc. Given this 0.36% exponential increase in tidal range, when the high water level at Bar Harbor is 1 m above MSL, the level 196 km distant at Saint John can be calculated in the same manner as compound interest. The type of calculation is of course based on an exponential growth, and in this instance yields about 1 × 1.0036¹⁹⁶ = 2.02 m. Similarly, the ratio between Bar Harbor water levels and those 366 km distant at Burntcoat Head, N.S., is 1 × 1.0036³⁶⁶ = 3.73 m.

Most salt marshes in the Bay of Fundy are raised to the level of the average tide, that is the MWL, in the 18-year cycle. Also, the number of high tides per year will vary considerably, depending on the phases of the three main tide-generating astronomical factors. As shown in Table 20, these coinciding peaks run in cycles of 0.53, 4.53 and 18.03 years.

Applied to variations in the levels of Bay of Fundy tides, the High Marsh Curve on Fig. 61 follows places where the local marsh level is assumed to be 1.2 m below the high water level during extreme high tides. It simply shows that the frequency of tidal flooding is much less in the upper reaches of the Bay of Fundy than in cases near its mouth. Drawn from a large data base, the S-curve in Fig. 61 shows the range in number of tides that can exceed a certain level. Taking the marsh level as an example, at Saint John it is approximately 3.5 m above MSL. The intersection with the S-curve shows that at Saint John the marsh will be flooded between 130 to 230 (C-B) times per year, depending upon when it happens during the 18 year cycle. At Bar Harbor, the number of annual floodings is between 660 and 700 (F-G), whereas the Cumberland marshes can expect only about 15 to 60 per year. (D-E).

Note also that the number of tides/year that will exceed a certain level at a given location is readily determined from Fig. 61. For example, with reference to Mean High Water (line X– W on Fig. 61), a minimum of 254 tides/year, or a mean of 354 tides/year, or a maximum of 433 tides/year (read off the ordinate), will exceed an elevation of 1.55 m (above MSL) at Bar Harbor, 3.10 m at Saint John, and 5.75 at Burntcoat Head (and 7.65 m at Truro). Although the data for Truro are not included, the last number illustrates the point that although the tide level at Truro exceeds that at Burntcoat Head, the tidal range is less at Truro because the floor of the estuary is higher and falls dry during ebb tides. The Shubenacadie River estuary, roughly midway between Truro and Burntcoat Head, illustrates the intermediate case: whereas the tide reaches its highest level about 9 km upstream on the Shubenacadie, MWL occurs at the river's mouth (see Fig. 36, curve 2, and contrast points L and B respectively).

In 1632, after about 25 years of exploration, colonization by Europeans of the lands surrounding the Bay of Fundy began in earnest. The settlers came from the Loire region of France, attracted by the coastal marshes. They were well acquainted with the techniques of dyking and draining in and around coastal marshes in their homeland as a result of activity by Dutch engineers invited to France in 1599 by King Henry IV. First to be reclaimed under this scheme were the Petit Poitou marsh along the Loire (1599– 1642) and the Petit Flandre marsh in the Saintonge district near Rochefort (1607– 1639) (Montbarbut 1985; Griffiths 1992). Early Acadian settlers applied the same techniques to tidal marshes in the New World. It was far easier to prepare this land than to clear thickly forested uplands (Fig. 62). At first the process proceeded peacefully. However, the struggle for political hegemony resulted in the forceful removal from the region in 1755 of approximately 10 000 settlers of French descent ("Acadians"). They were replaced by settlers from the British Isles and New England. The French had exploited to the limit the potential for dyking and draining the existing vegetated high salt marsh to form polders (van Veen 1939). Today 36 000 hectares of dykelands form a significant percentage of land suitable for agriculture in the Maritime Provinces of New Brunswick and Nova Scotia.

Elsewhere in the world are similar landscapes: the polder landscape of The Netherlands is often used as the prime example of this cultural environment. There, over 350 000 hect-ares of land have been reclaimed from the sea and another 1 300 000 hectares are protected against occasional storm tides. In Bangladesh 1 500 000 hectares of land are protected against tidal flooding, although 30– 40% of its land area (144 000 km²) is flooded by river water during 4 to 6 months of the year. The technique of protecting such land is everywhere identical, with flood-prone areas protected against high sea levels by dykes. The dyked land is drained by means of gravity sluices or pumps. In the Bay of Fundy area, the tidal range is so high that during much of the day the sea level is much lower than the dyked land, no matter how strong the tide. This means that gravity drainage is possible almost all of the time. For smaller tracts of protected lands, sluices with clapper gates are used, while in larger projects, gates are operated electrically. Clapper gates allow fresh upland water to flow into the sea while blocking seawater from flowing inland over the marsh. Gated drainage sluices within the dyke are called aboiteaux. Although there are many versions of this term, the original word may have been abat-eau (protector against the water), similar to abat-vent (windshield), abat-jour (Sun shade) abat-son and abat-voix (sound boards). The version abateau remains in use in the Saintonge district of France.

Most of the dyked marshes in the Bay of Fundy occur at the head of Cumberland Basin, the central of three main arms of the Bay. Along the Basin, which has a surface area of 118 km², 147 km² of dykelands have been created. The Tantramar marshes east of Sackville, N.B., along the estuaries of the Tantramar and Aulac Rivers, form the largest block at (72 km²), according to Griffiths (1992) "… the largest salt-marsh lands in the world…" (see Fig. 61).

The surface of most of the dyked marshes is 2 m or more below the highest level reached by the tides, while the salt marshes have an average elevation of 1.2 m below that level. Dyke elevation is of course partly attributable to the geologically recent submergence of the landmass relative to sea level in the Maritimes. Another important factor is the length of time that the dykelands have been deprived of additional sediment. For example near Amherst, N.S., dykes protecting the John Lusby marsh were breached during storms in late 1949 and early 1950. The damage was so extensive that no efforts were made to reclaim the land. The following tides built the marsh up a metre or more, burying fence posts in the process, the tops of which are now barely visible above the ground. Another important variable is the rate of erosion. Thus, although tides are higher in the Minas Basin, the coastline along Chignecto Bay is more easily eroded, and this factor may favour the greater abundance of marshes around Cumberland Basin.

The early French settlers, who were mainly subsistence farmers, used the dykelands for a variety of grain, oilseed and fibre crops, shifting later to beef production. The dykelands became the producers of the hay required to feed their livestock during the long winters. Centuries later, when means of transportation were improved and large population centres had developed along the eastern seaboard of North America, the excess hay was exported as fuel for the horse-drawn forms of city transportation. This lucrative enterprise stimulated the urge to increase the acreage of productive land. Most of the sediment carried onto the marsh is deposited in natural levees along the coastline and the tidal creeks leaving the more inland areas with much less sediment cover. As a result, the marsh surface slopes landward towards the edge of the upland where, in saucer-like basins and low-lying areas distant from the coast, freshwater bogs are common. At these sites, bogs with cat-tails and sphagnum vegetation are developed in a zone between the marsh and upland and, in some cases, even in the centre of a marsh. An example of the latter situation is the so-called Sunken Island Bog on the Tantramar marsh near Sackville, N.B. Underlain by marine clays, the bog is a prime example of a floating sphagnum bog surrounded by a fringe of vegetation, characteristic of wet conditions.

It may be no coincidence that after an article on "warping" or "tiding" was published in the British Farmer's Calendar of 1804 promoting the build-up of lands by tidal inundations in Lincolnshire and Yorkshire, a similar project was undertaken in the Tantramar region of New Brunswick. History records that many settlers from Yorkshire came to the Sackville area. Canals were dug between the Tantramar River and several nearby lakes, in the process transforming hundreds of hectares of freshwater peat bog into profitable grasslands.

At the mouth of the Tantramar River, tides have a maximum range of 14.5 m. As an employee of the Maritime Marshland Rehabilitation Administration (MMRA) in July 1952, the senior author made, with the aid of explosives, a 500 m long ditch from the Tantramar River to Long Lake, 22.5 km from the mouth of the river. Within about three years most of this 7.5 hectare lake was completely filled with silt for new farmland, and no longer appears on Topographical maps. The sets of very high tides in 1953 and 1954 contributed greatly to this last "tiding" carried out in the Bay of Fundy area.

Through vigorous efforts, the early French settlers proved the feasibility of dykeland development and utilization. Many remnants of old Acadian dykes exist on sections of marsh that have since been allowed to revert back to salt marsh. It appears that these settlers built their dykes at the very edge of the salt marsh, leaving only a very narrow safety margin for waves to dissipate most of their energy during High Water (Baird 1954). In addition to earth, the settlers also used pole and plank facing. This same method had been used centuries ago in Europe to protect the face of dykes, but was abandoned after shipworm (Teredo navalis) invaded the waters of northwestern Europe in the first half of the 18^th Century. The teredo consumed all the wooden parts of dykes and sluices and generally promoted wholesale rotting of poles and planking. Weaknesses thus created became foci for wave action, which continued the degradation until eventually holes developed in the dykes; the process is similar to that which results in sea caves and blow holes. Fortunately, teredo is not found in the Bay of Fundy, although it is present elsewhere along the North American coastline. Only a limited number of mollusc species are able to thrive in the upper region of the Bay of Fundy, the common periwinkle (Littorina litorea L.) being a recent newcomer. The fact that plank facing is still used in Fundy dykes attracted the German geographer Carl Schott to the area in 1952 to study the application of this ancient form of dyke protection. His work (Schott 1955) remains the best comprehensive report on the Bay of Fundy marshes.

After World War II economic studies indicated that rehabilitation of deteriorated dykes and aboiteaux would be beneficial. Accordingly, the Federal Government financed and carried out the building or rehabilitation of 400 km of dykes and new sluices, while provincial governments and landowners agreed to share the cost of associated drainage facilities. Several former dykelands were in such a state that no rehabilitation was attempted. Of these lands, a few have been acquired by the Canadian Wildlife Service to serve as National Wildlife areas, some of which are important stopover and feeding areas for migrating waterfowl. Low-lying sections of the dykelands, especially those fringing the uplands, were dyked off internally and transformed into waterfowl habitat (work that is financed and carried out by that uniquely North American organization, "Ducks Unlimited"). Some tracts of dykeland are small parcels of land sandwiched between the upland and meandering tidal estuaries. The required height of the dykes is only about 1.5 to 2.0 m above the original marsh level.

Present-day dyke construction is carried out using bulldozers and draglines. Topsoil is first removed from the projected dyke site and stockpiled for later re-installation. The bulk of the dyke material is taken from the drainage ditch running inside and parallel to the dyke. Any necessary additional material is taken from borrow pits outside the dyke. The outside face on exposed sections of dyke is covered with quarried rock trucked in during the winter when the dykes are able to withstand heavy traffic. These rocks dissipate wave energy generated during High Water.

The dimensions of the aboiteaux or discharge sluices are determined by a number of equations that take into account potential storm runoff from the watershed to be discharged through the sluice, the ranges of the local tides, the elevation of the dykeland, and the sluice invert elevation allowed by the receiving tidal channel. For sluices built in larger creeks and estuaries the available storage volume of basins upstream of the sluices is taken into account, as well as the probable hydro-graphs of the storm runoffs.

The large tidal ranges in the Bay of Fundy permit exclusive use of gravity drainage. Pumped drainage is uneconomical and would only be required for short periods during High Water should tide water exceed the water levels upstream of the sluice, and only then if the storage volume was insufficient to hold the watershed runoff during that period. Such occasions would be so rare that the large capacity pumps would hardly ever be used.

Some 435 small aboiteaux were built by the Maritime Marshland Rehabilitation Administration in the 1950s and 1960s. Dimensions ranged from 0.3 × 0.3 m boxes to multiples of 1.2 × 1.5 m boxes, set side by side and made from chemically treated lumber. More recently, sluices have been built with asphalt-coated corrugated steel pipe. The flapgates are made of bronze or steel and hung by horizontal hinges or steel chains. The latter method appears superior for it allows debris to pass through the sluices with less damage to the gates.

To eliminate the need to rehabilitate tens of kilometres of dykes and numerous small aboiteaux, six large tidal sluices were built in larger tidal estuaries (those of the Annapolis, Avon, Memramcook, Petitcodiac, Shepody and Tantramar rivers). The largest of these structures are the five 8.8 × 8.8 m sluices (with inverts at 1.5 m below MSL) set in the Petitcodiac River estuary near Moncton, N.B. Two slightly smaller sluices in the Annapolis River estuary were converted in 1982 into an experimental tidal power station (see section 7.5). The large tidal sluices are steel-reinforced concrete structures with flared entrances and are equipped with electrically hoisted gates which can, if necessary, be operated manually. These gates can also be electrically heated for use during periods of heavy frost. In the larger estuaries the sluices were not able to pass enough water during the time of dam construction. Therefore auxiliary sluices were built to keep tidal currents in the closure gap at manageable speeds. Later on these sluices were buried within the dam itself. These dams also serve as causeways for highway and rail traffic, shortening the connecting links among the provinces to a significant degree.

The operation and management of dykelands are carried out under the direction of "marsh bodies", elected by landowners and assisted by officers from the provincial departments of agriculture. There are 86 such "marsh bodies" in Nova Scotia and 39 in New Brunswick.

The reclamation of Fundy tidal marshes has been regarded by some people as ecologically disastrous in light of recently publicized figures of salt marsh production. For example, in semi-tropical Georgia, Spartina alterniflora has been measured at 2883 gm^-2/yr, a figure greatly exceeding that of other forms of natural vegetation or agriculture (see Pomeroy and Weigert 1981). In Virginia, the measured rate of production is 1143 grams, in Rhode Island 668 grams, and on high marshes around Cumberland Basin a mere 400 gm^-2/yr. Factors responsible for the relatively low production of salt marsh vegetation in Cumberland Basin include: the shortness of the growing season, nutrient limitation, inundation characteristics, sediment aeration and, last but not least, tidal stress.

Unfortunately, the Georgia figures are most often used in ecological arguments naming the Bay of Fundy a disaster area. Simply put, Fundy tidal marshes can remain dry for months on end and thus cannot be construed as hatcheries or nursery areas for saltwater fish. The dyking of the high marsh in the upper reaches of the Bay will not have had the same biological impact as results from utilization of marshes in more southern latitudes, or in regions with smaller tides. Nor is there much basis for stating that reclamation of the salt marshes has been detrimental to the region's bird population. When the French settlers arrived in the upper reaches of the Bay, they called the area "Tintamarre" because of the birdcalls that filled the air. The present name of the Tantramar marshes derives from this word, which means din, racket, or noise. With this explanation comes the idea that the original undyked salt marshes had a rich birdlife. It is interesting to examine this proposition.

In The Netherlands, until recently, every hectare of reclaimed or unreclaimed marshland seemed to have its own resident couple of lapwings, godwits, and skylarks, and every acre of water its pair of mallards. This was the case despite the dense human population and intense agricultural use of the polderland. More intense agricultural practices and use of pesticides have recently changed this picture drastically. In contrast, Maritime marshes are almost devoid of birdlife, especially the out-to-sea marshlands. The French may possibly have named the marshes "Tintamarre" during the spring migration period when the marshes are staging areas for geese and ducks, or during the late summer period when hundreds of thousands of waders congregate at selected areas.

Overall, the Fundy marshes have very little bird life. Indeed, the salt marshes could never have provided suitable nesting habitats for birds because there is a total lack of trees. Furthermore, the infrequent monthly flooding of the marsh with salt water, which occurs in sets that arrive about 47 days later from year to year, makes the marshes unsuited for ground breeders. Since these floodings are more intense every 4 to 5 years, it seems improbable that bird species could have become used to monthly floods and incorporated them into their genetic programs. This leaves only breeding areas for birds like the raucous redwing blackbirds that can multiply among the cat-tails in transition areas between the salt and freshwater marshes. This may well have been the "tintamarre" that so impressed the French settlers.

The dyking of marshes has had another, less remarked upon effect (Teal and Teal 1969). Before dyke construction, tidal marshes acted as sinks for excess suspended sediment in the system. Following dyke construction, this suspended material may have resulted in changes to the habitats not only of plankton, but also of fish like shad and gaspereaux. Historical records show that in 1837, vessels from 50 to 150 tons were able to sail almost daily up Cobequid Bay and the Salmon River to receive and discharge cargo near the now defunct Board Landing Bridge near Truro (see Fig. 60). After 1867, however, siltation of the river prevented further shipping access. Similarly in England by about 1888 all the harbours on the southeast coast of England were silted up due to dyking of marshes, requiring a vast expenditure for constant dredging to keep harbours operational.

Very little use is being made of salt marshes for socio-economic purposes unless the terrain is transformed to dykeland. Dykelands provide pasture, especially for sheep, and a source of "wild" hay. Some plants such as Salicornia and Plantago were eaten as a vegetable on a minor scale (Trueman 1896). The marsh is generally inaccessible and unattractive to conventional recreation seekers. Fundy salt marshes are cold in winter, wet in the spring, mosquito ridden and covered with nearly impassable vegetation during the summer and early fall, and muddy throughout the year. In fact, smells emanating from marshes are usually so obnoxious that they have a serious effect on the recreational usefulness of adjoining shore properties. The only "recreational use" of salt marshes is the realization of their importance within ecosystems.

Today, although an estimated 90% of the original salt marshes at the head of the Bay of Fundy has been reclaimed by dyking, much of the rest remains idle. The once lucrative hay market has disappeared. Few if any people actually live in the tidal marsh area itself. Farmers use the land around their farmsteads with the most effect and the greatest care. However the further the land is removed from their daily supervision the less attention it receives. An enduring myth is that marshlands are very fertile, whereas in fact they are limited in some plant nutrients such as calcium and phosphate. The silty soils also offer far less in terms of ion exchange capacity than, for example, the calcium-rich clays of west European marshlands. One can argue whether the present-day utilization of the Bay of Fundy dykelands is what it should be, but it appears that the benefits derived from them over the centuries far outweigh any disadvantages caused by dyking the high marsh areas.

Storm tides (here taken as synonymous with storm surges) generally spell trouble. A positive storm tide is a large rise in water level accompanying a coastal storm, the water rise caused by violent winds and low atmospheric pressure. Conversely, with high atmospheric pressure, a negative storm tide may lower sea level from its predicted value (Wells 1986; Parkes et al. 1997). When a storm tide coincides with an exceptionally high astronomical tide and shallow water depths, the results may be little short of catastrophic. The record of storm tides is knowable, and they may be predictable in the Gulf of Maine-Bay of Fundy region. Inevitably they are linked to pressure changes in the atmosphere, and to wind set-up.

Atmospheric pressures deviating from the normal can cause appreciable changes in sea level. Where pressure is high, sea level will drop, while the excess water moves to the areas where the pressure is below normal. A difference of one kiloPascal (1 kPa = 10 millibars) can effect a 0.0975 m change in sea level. Such differences are more easily detected in areas of very small tides, an example being the Bras d'Or Lakes on Cape Breton Island, N.S., which are connected to the ocean by small channels restricting the daily tidal movements of water. In part, the salinity levels in these lakes are maintained because of changes in water level caused by variable barometric pressures responding to cycles longer than the semidiurnal and diurnal tide cycles.

In Maritime Canada as a rule the largest deviations from normal of atmospheric pressure occur during fall and winter (see Table 21). A possible deviation from the mean of 5 kPa means that sea level at times can be raised nearly 0.5 m. This is not so great as to create flood conditions, unless coincident with the High Water of extremely high tides. Unfortunately, statistics on air pressures are not readily available. However the Monthly Meteorological Summaries of the Truro station in Nova Scotia indicate that a deviation of 2 kPa or more occurs during three or four months of each year, and that a deviation of 4 kPa may occur more than once a year.

Sea level also may be raised above normal due to a wind/ storm "set-up" (see Fig. 63). This is a flow of air that drives the water toward the shore line. Here the gradient of the water is balanced by the shear stress of the wind blowing over the water. Quite apart from the set-up which propagates into the Bay due to wind over the Gulf of Maine and Scotian Shelf, in the upper reaches of the Bay of Fundy the tides (here, in reference to total water level; i.e., tide plus storm surge) are large and at times the currents are very strong. Currents in the Minas Channel can run at the rate of 7 to 8 knots (3.5 to 4 ms^-1), while in the Minas Basin and Chignecto Bay the rates drop to 2.5 to 3 knots, increasing again in Cobequid Bay and Cumberland Basin to over 4 knots. Winds blowing against the current will, in effect, lessen the height of the tide and the rate of the current, but they may create tidal rips in which the surface waves are steepened to the point of breaking, causing violent seas.

The wind set-up is estimated by an equation which takes into account wind speed, fetch, angle between wind direction and direction for which the set-up is calculated, and the depth of the section of water. For each section with a different depth a separate calculation is made. After passing over all sections, the total wind effect will depend upon the fetch (distance over water that the wind is active). The following equation can be used to estimate the wind set-up (Sx) along the shoreline:

-> See the list of figures

Where Sx = wind set up in m, over the fetch section; K=stress coefficient [Lorentz used (3.09 ± 0.4) · 10^{– 5}; U.S. Corps of Engineers uses 2.36 · 10^{– 5}]; U = wind speed in km/hr; A = angle between wind and fetch direction; X = length in km, of fetch section with depth (D) in metres.

When the set-up becomes 10% of the depth, which happens when the wind speed exceeds approximately 60 times the depth in m, (55) results in too large values of Sx. It is better replaced by:

-> See the list of figures

Note that the set-up is smaller when the depth is larger. One would thus expect that during High Water the wind set-up would be less than with Low Water. However in tidal waters where the large sections fall dry during Low Water the increased fetch during High Water can more than offset the influence of increased depth.

In areas where the tidal range is small, storm conditions can raise water levels somewhat above that of ordinary tides. In the upper reaches of the Bay of Fundy the difference between levels reached during large tides and small tides may exceed 4 m, and wind set-ups occurring during larger tides may cause flood conditions. Severe flood conditions are virtually guaranteed when adverse weather conditions coincide with the High Water of extremely large, astronomically-caused tides. The probability of such a coincidence is small and very difficult to estimate. However, indications are that in the Maritime Provinces of eastern Canada such conditions have occurred at intervals of more than a century. Historical accounts document two occasions when great storms coincided with a very high tide. These occasions were October, 1869, and November, 1759.

On the afternoon of 4 October, 1869, Thomas Earle tried to leave Beacon Light at Sand Point in Saint John Harbour (Fig. 64), but heavy seas drove him back. As wind and waves continued to rise, water flooded the first floor of the lighthouse, and threatened to rise higher. Waves broke over the top of the lighthouse, carrying away the bell and smashing the windows and the globe protecting the gas light. Retreating to the middle story, the keeper closed the hatchway behind him. The gas light continued to burn, and fanned by the storm, ignited the superstructure. In desperation Earle managed to fit a spare globe around the light, and was somehow able to contain the fire. Three and a half hours later the tide receded sufficiently to allow him to gain the safety of higher ground.

Thomas Earle, like most coastal folk along the Bay of Fundy, was used to high tides. He would have known the difference between "spring" and "neap" tides. He knew too that storms can cause unusually high water levels, especially when coincident with an otherwise normal high tide. Normally he would have relied on the Farmers' Almanac to predict tidal conditions throughout the year. But conditions in the Bay of Fundy on 4– 5 October, 1869 were not normal.

Earle may never have heard of Pliny the Elder, or of Sir Thomas Herschel, who in 1860 helped sustain the legend of 120 foot-high tides by repeating the myth of the 4th Century BCE navigator, Pytheas. However he probably had read in Maritime newspapers about S.M. Saxby's dire prediction of the great storm from which he was to only narrowly escape. The prediction had been published the previous November in the British newspapers London Standard and London Press (Saxby 1868). It read:

"I now beg to state with regard to 1869 at 7:00 a.m., October 5, the Moon will be at the part of her orbit which is nearest the Earth. Her attraction, will be therefore at its maximum force. At noon of the same day the Moon will be on the Earth's equator, a circumstance which never occurs without marked atmospheric disturbance, and at 2 p.m. of the same day lines drawn from the Earth's centre would cut the Sun and Moon in the same arc of right ascension (the Moon's attraction and the Sun's attraction will therefore be acting in the same direction); in other words, the New Moon will be on the Earth's equator when in perigee, and nothing more threatening can, I say, occur without miracle. With your permission, I will, during September next (1869) for the safety of mariners, briefly remind your readers of this warning. In the meantime, there will be time for repair of unsafe sea walls, and for the circulation of this notice throughout the world."

Saxby, a civilian instructor of Naval Engineers of the British Royal Navy, published books and almanacs of weather predictions, claiming a relationship between stormy weather and the Moon crossing through the plane of the Earth's equator. As this occurs every two weeks it could be said that Saxby gave himself a better than 50 percent chance of claiming reliable forecasts. Note that Saxby did not specify the location where his threatening tide would occur. Astronomic conditions were to be right for higher than normal tides, world-wide, on 5 October, 1869.

The prediction was fulfilled in the Bay of Fundy by what was to become known as the "Saxby Tide" or "Saxby Gale". A storm hit the North American eastern seaboard on 4 October, 1869. All along the New England coast severe flooding and wind damage occurred. Between Washington and the upper reaches of the Bay of Fundy, more than 150 vessels sank or were blown ashore, 121 of them between St. Andrews, N.B., and Machias, Maine. Near Point Lepreau, N.B., the barque Genii was wrecked and eleven lives lost (Ganong 1911; Tibbets 1967).

At 17:00 hours, the wind increased to a gale, and an hour later rain began. By the evening of 4 October, Saint John streets were littered with debris torn from buildings. The gale continued from the south, reaching hurricane force about 21:00 hours by which time the rain had stopped. Around 22:00 hours the wind shifted to the southwest and subsided. Fully an hour and a half before the tide ordinarily would have reached its peak, waves from the Bay of Fundy were breaking over every wharf in Saint John Harbour. Ships parted from their moorings; some were driven ashore, others were badly damaged. Dozens of wharves, fish shacks, and abutments were washed away (MacLean 1979).

About 160 km northeast, at the head of the Bay of Fundy, the Saxby Tide occurred at 01:00 hours, 5 October, overtopping the dykes by at least 0.9 m. In Cumberland Basin, two fishing schooners were lifted over the dykes of the Tantramar marshes and deposited five kilometres from the shoreline. At Moncton, N.B., the water rose nearly 2 m above the next highest tide on record. On the marshes west of Amherst, N.S., a barn in which some people had been sleeping, was despatched well over a kilometre up the marsh. During the course of the storm two men were carried out to sea and drowned. Three others were able to cling to timbers and were eventually washed ashore. Much of the hay stored on the marsh was swept out to sea and the remainder was scattered over the area. Farmers drew lots in order to divide the little hay that was salvaged. In and around the Minas Basin, the gale was less severe, although the rainfall was heavy. Everywhere dykes were breached, cattle and sheep were drowned, lengthy portions of railroad beds were washed away, and in many areas travel became impossible. Communications likewise were shut down by the weather to a degree difficult to imagine today.

A year after the disaster, a survey was made between 13 August and 31 December for a proposed canal across the Chignecto Isthmus from the Bay of Fundy to the Northumberland Strait. To simplify calculations, by avoiding negative values, the surveyors assumed a datum (reference elevation) of 100 feet (30.48 m) below the average top elevation of the dykes. By this reckoning the Saxby Tide was reported to have reached levels of 100 feet or more. These figures of course have no direct relationship to mean sea level. But they have, unfortunately, helped perpetuate the erroneous notion of enormous tides.

During the 1870 survey, extremely high astronomical tides (i.e. tides not associated with storms) occurred on 23 September (28.83 m), 26 October (28.83 m), and 23 November (28.80 m) for an average of 28.82 m. Comparing this elevation with the average top elevation of the dykes yields (30.48 m– 28.82 m) a difference of 1.79 m. Thus although the level of the dykes may have been raised slightly in the years since 1870, the Saxby Tide was at least 1.5 m higher than astronomically caused high tides.

Only sparse accounts survive of a probable precedent, over 100 years earlier, of the Saxby Tide. The nameless historical storm tide battered the Bay of Fundy region on 3 and 4 November, 1759. In Beamish Murdock's "History of Nova Scotia" contained in the Gentleman's Magazine of 1760, it is recorded that at Fort Cumberland (earlier and later named Fort Beausejour) 700 chords of firewood were swept away from a woodyard that was at least 3 m above the protecting dyke. In Saint John Harbour water reportedly rose 1.8 m higher than was usual for large tides. Storm waves broke on the terraces of Fort Frederick, located well inside the Harbour, demolishing a store house and spilling provisions into the sea.

It is unknown whether anyone foretold the historical storm tide of November 1759. The regionality of storms and storm tides makes predictions like Saxby's a rather unscientific exercise, however much they may impress the local populace. Even today, meteorologists, assisted by a world-wide computerized network of data-gathering instruments on the ground and in satellites, are faced with more variables than they are able to handle (Wells 1986). The behaviour and frequency of complex extraordinary weather systems, their whereabouts, strength, rate and direction of movement, are beyond their grasp.

Whereas the Saxby Gale was predicted a year before it happened, the "Groundhog Day" storm of 2 February, 1976, was forecast only hours before it broke. Two days earlier a weak low pressure area hovered over Alabama and Texas. Subsequently, this system met a small high pressure system from western Ontario. Small craft warnings were issued on 1 February, advising of strong southerly winds. Then around noon that day, gale force winds were predicted. During the night, gale warnings were changed into severe storm warnings. By 8:00 hours on 2 February, the barometric pressure had dropped precipitously over the Gulf of Maine, signalling the likelihood of a storm accompanied by higher than normal tides (Amirault and Gates 1976).

The storm hit the coast of Maine hard (Morrill et al. 1979). In places, the tide rose more than 2.5 m above the predicted level, heavily eroding the coastline. Waves hammered coastal installations. A freighter anchored in Penobscot Bay was blown aground. The strong south-southeasterly winds, which had been blowing for 5 to 6 hours over the open water along the major axis of the Penobscot Bay, resulted in a storm surge in the Bay and up the Penobscot River. Much of Bangor was flooded. In less than 15 minutes the water reached its maximum depth of 3.7 m in the river, 3.2 m above the predicted tide level.

By the afternoon of Groundhog Day (2 February, 1976), the storm was raging along the eastern seaboard (Fig. 65). Intermittent power failures and curious sparking effects were the result of short circuits caused as winds swept seawater across the countryside. Fortunately, the tide was an apogean spring tide (Conkling 1995). In Saint John, the high tide was expected to be only 7.7 m (25.2 ft) above CD at 13:10 hours. Note that for comparative purposes, this is almost exactly equal to mean Higher High Water (25.3 ft) at Saint John. However, on the afternoon of Groundhog Day the tide rose to 9.16 m (29.65 ft) above CD, fully 1.46 m higher than expected. The damage would have been simply enormous had the storm occurred on the perigean spring tides: sixteen days later on 18 February, the tide at Saint John was predicted to reach a height of 8.4 m above CD; a month and a half later, on 18 March, the tide was predicted to rise 8.66 m above CD; two and a half months later on 16 April, the tide was predicted to be 8.84 m above CD, 1.07 m higher than the predicted tide on Groundhog Day. In short, a storm on 16 April, 1976, would have had the potential of causing calamity on the scale of the Saxby Tide.

In light of the above considerations it seems a distinct possibility that another storm tide of the magnitude of the Saxby Tide of 1869 will occur in the Bay of Fundy region. The big question is when will this storm likely occur? Until then, successor dykes to those erected during the 17^th Century along the Bay of Fundy marshlands, continue to protect agricultural lands.

These agricultural lands are shrinking before the onslaught of modern highway and suburban development schemes, not to mention the continued transgression of the sea. Aboiteaux draining these lands are of course unable to discharge when the tide outside is higher than the water upstream. With a heavy runoff, the potential therefore exists for the land to be flooded with fresh water. This is realized and tolerated by farmers of course, but the situation is rather less stoically accepted by developers.

Generally, the stronger tides happen during spring and summer around midnight, and during fall and winter near noon. But the fact that the Saxby Tide occurred during the night and in the fall shows that even this rule is not valid for extraordinarily severe storm tides. Nor does the fact that the landmass of the upper Bay of Fundy area is submerging at a rate of about 3 mm per year improve the situation. The dykes can be made higher, but if they are not raised to Saxby-type levels, their overflow will raise the flood levels in the dyked areas that much higher. The developed areas near Truro, Weymouth and Advocate in Nova Scotia, and Moncton in New Brunswick (see Fig. 60), are among potential victims of such storm tide events.

Pytheas, a navigator from the Greek colony of Massalia (modern Marseilles, France), explored the northern Atlantic Ocean in the Fourth Century B.C.E. Proceeding north after passing between the Pillars of Hercules, he noticed the lengthening of the summer days and observed the midnight Sun in Thule, a six day voyage north of Britain. He was aware too, of the relationship between tides and the Moon's motion along its orbit.

Pliny the Elder (23– 70 A.D.) mentioned in his Historia Naturalis that, according to Pytheas, the tides north of Britain rise to heights of 120 feet ("octogenis cubitis"). The notion of the existence of such enormous tides persists. For example, the eminent 19th Century scientists Sir John Herschel and Sir Oliver Lodge repeated the belief that 120 foot-high tides occur in the Bay of Fundy. Only a few years ago the same misconception was linked to the extraordinarily high tides associated with the so-called "Saxby Gale".

One might wonder where the extraordinary high tides occurred that so impressed Pytheas. He must have been familiar with the 40 foot-high (12.19 m) tides that regularly occur along the Brittany coast on the French side of the English Channel, or in the Bristol Channel south of Wales. According to the Roman writer Festus Avienus, these waters were visited a century before by Himilco, a famous Phoenician explorer, and by merchantmen trading for tin. The only other tides in the North Atlantic of the same order of magnitude are the 30 ft (9.14 m) tides in the White Sea portion of the Arctic Ocean, and the more than 40 ft (12.19 m) tides in Ungava Bay, Quebec and the Bay of Fundy. The White Sea can only be reached by waters lying north of the Arctic circle where the midnight Sun can be observed in midsummer. Could Phoenician seafarers have reached the western side of the Atlantic and, after visiting the Bay of Fundy, passed to Pytheas the notion of 120 foot-high tides?

As we have seen, tides are caused by the attraction of the Moon and Sun on water particles near the surface of the Earth. Since the orbits of the Moon around the Earth, and of the Earth around the Sun, are elliptical, the effects are variable in strength, like the resulting tides. The redeeming feature is that every aspect of each motion has a corresponding periodicity to which tidal variations can be related. The pages-long equation describing the paths of celestial bodies, a masterpiece of human ingenuity, was first set out by Louis Lagrange (1736– 1813) and Pierre Simon de Laplace (1749– 1827). Yet even the ancients possessed considerable knowledge concerning these matters.

Nearly four centuries before Pytheas described the influence that the Moon's motions have on the tides, Chaldean priests in the Middle East were able to predict recurrence of eclipses. This was because of their knowledge of the Saros, a Babylonian name adopted by modern astronomers for a cycle with a period of 18 years, 11 days and 8 hours. In this 18.03-year cycle, the Moon, Sun, and Earth return to almost identical relative positions to each other. This is the cycle in which similar solar and lunar eclipses repeat themselves. Eclipses result when the Sun, Moon, and Earth are in, or almost in, one straight line.

The paths of the solar eclipses over the Earth's surface are almost identical in shape, but are located 110° to 130° west of the path of the eclipse of 18 years previous (Abell et al. 1988). For example, one series of solar eclipses began on 17 May, 1501 (Julian calendar), as a partial eclipse. (At present, 12 such series producing total solar eclipses occur during a Saros cycle of 18.03 years). After 15 Saros cycles with partial and annular eclipses, the eclipse became total on 6 November, 1771. The total eclipse seen on 7 March, 1970, over Mexico, the USA and Canada was one of the series; likewise that of 18 March, 1988 as seen over the Pacific, Sumatra and Borneo (see Fig. 66). After this, there will be 35 more total eclipses followed by half a dozen partial ones. All told, the Saros prediction is valid for 1226 years [(1988– 1501) + (35+6) · 18.03 = 1226.23 years)] for this particular series.

Because the astronomical conditions conducive to generating large tides match the Saros cycle, their recurrence at 18.03 year intervals is expected. Could it be that this particular timing is closely linked to the occurrence of exceptionally high tides like the Saxby Tide? There is good suggestive evidence that this is the case, but first we need to look further into the causes of tides and their variations. Only then will we better appreciate the implications of the Saros for a future tide comparable to the Saxby.

Orbital forcing of tidal cycles is only a small portion of the spectrum (Fig. 67) of astronomically-driven periods which exert gravitational effects on Earth and the affairs of humankind (Rampino et al. 1987). The same periodic behaviour governs changes over time in the energy distribution reaching Earth's atmosphere, and must have done so throughout geologic time. We are concerned here only with tidal phenomena within a small portion of the calendar and solar frequency bands.

S.M.Saxby might have added to his prediction that the great storm tide destined to immortalize him, especially in the view of many Maritimers, would coincide with the Saros (Desplanque 1974). In this 18.03 year cycle, Moon, Sun and Earth return to almost identical relative positions. It is astounding to realize that by 800 B.C., Chaldean priests knew the Saros well enough to accurately predict eclipses. Could it be that eclipses, or more particularly the time interval between eclipses, might be associated with far higher than normal tides?

In general, the average gravitational effect of the Sun is about 46 percent that of the Moon; however, in the Bay of Fundy the effect of the Sun is only about 15 percent that of the Moon. Further, because the orbits of the Earth around the Sun and the Moon around the Earth are elliptical, and their paths influenced by many factors (Fig. 67), the gravitational effects are variable in strength, like tides (House 1995).

Normal tides are termed astronomical tides because their main variations are generated by three astronomical phenomena as noted earlier: the variable distance between the Moon and Earth; the variable positions of the Moon, Sun and Earth relative to each other; and the declination of the Moon and Sun relative to the Earth's equator. Additional astronomical factors that influence tides include: the Earth's rotation; the eccentricity of the Moon's orbit, which varies depending on the Sun's position in relation to the longest axis of the Moon's orbit.

Non-astronomical phenomena include: the possible increasing tidal range in places like the Bay of Fundy due to deepening waters (Godin 1992); atmospheric disturbances; the geometric shape of inlets, bays and ocean basins; the postglacial rise in sea level (see Fig. 68). This last factor, by no means trivial, translates to about a world-wide 2 mm/yr submergence of the land in relation to sea level (Schneider 1997).

As seen from Earth, the Moon and Sun seem to move within two imaginary rings around Earth's centre (Fig. 69). The ring in which the Moon's motions are confined has an outside diameter of 813 000 km and a maximum thickness of 50 000 km, while its width varies during an 18.61 year period between 410 000 km and 255 000 km. The Sun appears to move within a similar ring, with an outside diameter of 3.04 · 10⁸ km, a maximum thickness of 5 · 10⁶ km, while its width remains constant at 1.25 · 10⁸ km. At perigee the distance to the Moon is 3.57 · 10⁵ km, and at apogee the distance is 4.07 · 10⁵ km.

The positions of the Sun and Moon in their respective (elliptical) orbits, and in relation to each other and the Earth, are only occasionally repeated. When one position is inducive to generating large tides, an approximate date for a repeat performance can be determined by matching the several types of astronomical months. Such months are designated as synodic, anomalistic, tropical, nodical or evectional, according to whether the revolution of the Moon around the Earth is relative to the Sun's position, the shortest distance to the Earth, its passing through the Earth's equator, its passing through the ecliptic, or the variation in the eccentricity of its orbit. For some months, such as the synodical (full and new moon), tropical, and nodical, the characteristics influencing the tides occur half-monthly. Since the synodical conditions provide the dominant tidal conditions, one can expect two sets of spring tides during one synodical month. But when one of these sets coincides with the Moon's closest approach to the Earth, extra high (perigean) spring tides will occur.

It takes 29.531 solar days between one new moon and the next. However, only 27.555 solar days (anomalistic month) elapse from the time that the Moon is closest to the Earth, to the next such occasion during the Moon's elliptical orbit around Earth. When the Moon is in perigee, and its phase is either full moon or new moon, one can expect the strongest tides. However, as the periods of both movements are not the same, the coincidence of such occurrences is only periodic.

Imagine a racetrack. On this track are two cars, marked N and F. They always move half a track apart around the raceway. Let the raceway be 360° long. Thus, each day, these cars move 360°/29.531 = 12.19°/day (V). Another car, marked P, starts at the same time beside car N, but its velocity is 360°/ 27.55 = 13.06°/day (W), thus somewhat faster than car N. In time, car P will overtake car F, which was 180° in front of car N. The time it will take to close this gap can be calculated (approximately) as 180/(W– V) = 205.892 days.

The same applies to lunar movements. After about 206 days, the conditions for stronger than average tides recur, perigee coinciding either with new moon or full moon. Two of these periods are 411.78 days long. Thus, each year one can expect the conditions for stronger tides to be (411.78– 365) = 47 or 46 days (leap year) later than in the previous year.

The declination of the Moon also has an influence on the strength of the tides on particular days. This declination has its strongest values twice in a period of 27.321 solar days (tropical month) with a velocity of 360°/27.32 = 13.18°/day. This is somewhat faster than the velocity of the perigee. Therefore in order to have the coincidence of a similar combination between perigee and declination of the Moon, it will take 1615.75 days, or 4.42 years. Should a particular part of the declination cycle cause somewhat higher tides then the coincidence with perigee will repeat after 4.42 years. During the Saros cycle of 6585.32 days, there will be 238.997 perigee cycles, 446.01 (full moon-new moon) cycles, and 482.07 declination cycles.

The most favourable combination of factors to produce strong tides in the Bay of Fundy occurs when perigee coincides with spring tide at the very time that anomalistic, synodic and tropical months peak simultaneously. As it happens (see Table 22), the best match occurs after a period of 6585.3 days (18.03 years). The driving mechanism of this cyclic phenomenon is the same one that orders the timing of eclipses – the Saros.

Given the clockwork precision of astronomical conditions and their absolute control over normal tide variations it is reasonable now for us to enquire whether the Saxby Tide, and for that matter other historical storm tides in the Bay of Fundy, coincided with the Saros.

What are the chances of a periodic storm system on the scale of the Saxby tide? The only certainty is that the relationships between the Moon and the Sun that produce the highest tides on Earth are repeated in the same periods as those that create solar and lunar eclipses (Abell et al.1988), namely the Saros cycle of 18.03 years. Therefore, to check the position of an historical high tide in the Saros, one need only add to the tide's date the appropriate multiple of the Saros (Table 23) to reach a particular time interval for which the tidal record is well known. Detailed tidal records in Canada were first kept about 1894. So as a reference point let's choose a date close to the end of 1958, which is eleven Saros after the Saxby Tide. The tidal levels referred to are those measured at Saint John, where the average high tide is 7.7 m (25.2 feet) above CD. To test the reliability of the method of prediction, our examples of historical storm tides can now be checked against predicted tides, n Saros cycles later.

Checking the multiples of the Saros against storm tides, we discover that the storm tides of 1759 and 1869 correlate very closely with predicted high tides of the Saros cycle (Table 24). So do the 1976 Groundhog Day storm, and the exceptionally High Water of 12 October, 1887, experienced in Moncton, and the storm tides of 20– 22 December, 1995 (Taylor et al. 1996). However, it is important to bear in mind through any exercise of this type that Saros cycles are long term harmonic motions. This means that near the top or bottom of the cycle the rate of change with time is relatively small. Thus, the "peaks" of Saros cycles are not confined to points in time, but to rather short intervals of time.

Whether property owners along the Bay of Fundy should be reminded of their next appointment with the Saros in 2012– 2013 AD is not trivial question (Fig. 70). With increasing encroachment of people in coastal zones, the risk of loss of life and major property damage in the Gulf of Maine-Bay of Fundy region is substantial in the event of a tide like the Saxby Tide (Shaw et al. 1994). Simply put, what is the probability of a storm tide coinciding with a large astronomical tide? We can conveniently address this question using the above examples. We know that a "peak" of the Saros occurred between 6 March, 1958, and 31 December, 1959. During this time, there were 1288 tides at Saint John, 37 of which were extreme astronomical high tides (28.5 feet or higher). Thus, the chances that an historically memorable storm tide coincided with one of the 1288 tides, is slightly less than 3 percent. This is assuming that the occurrence of stormy weather conditions is spread evenly throughout the year. The increased incidence of high winds during late spring and fall probably favours the odds of gale force conditions coincident with high tides slightly above 3 percent.

Most assuredly, postglacial sea-level rise is a significant factor in all this. With each and every repeat of the Saros, an increase of the high tide mark of at least 3.6 cm (2 mm/year for 18 years) can be expected. Thus, since the Saxby Tide more than seven Saros ago, sea level has risen eustatically nearly 25 cm. Added to the minimum 1.5 m by which the Saxby Tide exceeded high astronomical tides, a height is calculated that that is more than sufficient to overrun the present dyke system.

It seems likely that tides like the Saxby might be recurrent, although one wishes for a larger database. The clockwork precision of astronomical conditions exerts absolute control over normal tidal variations. But there remains much to learn about long term periodic events associated with tides and the weather. We have seen that only significant storms coincident with large tides, or extraordinarily severe storms coincident with medium tides, can result in higher tide marks than are reached by astronomical tides alone. Detailed tidal records over several decades show that in the Bay of Fundy there is a tendency for slightly higher maximum monthly High Water marks in a 4.5 year cycle, examples being the peaks that occurred in 1998 and 2002. Indeed, in this region, high perigean tides levels can be anticipated at intervals of 1 month, 7 months, 4.5 years and 18 years. When such high tide levels coincide with severe atmospheric disturbances, exceptionally High Water surfaces can be expected. However, short of an extraterrestrial catastrophe, they are not likely to attain the 120-foot (36.6 m) height of legend. Property owners along the Bay of Fundy should nevertheless keep in mind their next appointment with the Saros in 2012– 2013.

The need for precise determination of tidal water boundaries stems from numerous concerns, including cadastral surveying, coastal property evaluation, development of offshore resources, protection of fisheries, and ownership of the foreshore and sea bed (Nichols 1983; Daborn and Dadswell 1988). Historically, the dividing line between wet and dry land, or as far as the tide ebbs and flows, has been critical in resolving tidal boundary problems in the coastal zone (Ketchum 1972). However, the location of the mean high water line has been a matter of considerable litigation (Greulich 1979; Desplanque 1977). What exactly is this dividing line, and how can levels like Mean High Water and Mean Low Water be most accurately defined?

Presumably, Mean High Water is reached under mean astronomical conditions, with perhaps some long-term tectonic and climatological influences to be considered. Climatic influences in Atlantic Canada tend to raise the water during the winter months, so that in areas of relatively small astronomical tides along the Atlantic seaboard the frequency of extreme high observed water levels is higher during that season. However, astronomical influences are variable and follow cycles in which the magnitude of influences waxes and wanes (Schureman 1941). The longer the cycles, the greater the variation. In the Fundy region, as discussed earlier (section 11.0), distinct cycles are recognized (Desplanque and Bray 1986).

Meteorological influences can raise or lower water surface levels over a period of days, during which time High Waters and Low Waters are similarly affected. On a year long scale, climatic influences in the Atlantic region tend to raise the water level during winter months. Thus in areas of relatively small astronomical tides along the Atlantic coast of North America the frequency of extreme high observed water levels is higher during winter. The eustatic rise of sea level can not be neglected in these considerations: over the course of an 18 year cycle, accepting a global average 2 mm/yr eustatic rise in sea level (Schneider 1997), water level will have risen by about 3.5 cm. In view of the above considerations it appears that decisions handed down in courts of law on issues concerning tidal water boundaries are in many cases equivocal (Nichols 1983; Desplanque 1977).

Harvey et al. (1998) provided a very interesting analysis of the legal and policy framework concerning the restoration of the habitat of Fundy estuaries. Unfortunately, important overlapping boundaries of provincial and federal jurisdiction, in some cases unresolved, seem to have resulted in poorly regulated environmental protection. Details of the many issues of territorial and legislative jurisdiction, as they generally concern Bay of Fundy waters and the coastal zone, are beyond the scope of this paper. Nevertheless many of the key issues center on the problem of establishing specific water level boundaries.

Concerning the tide levels and terms used in Canada (as set out in section 2.1), all tidal measurements are made from the local Chart Datum (CD). The International Hydrographic Bureau recommends that CD at a certain location should be at an elevation so low that the tide at that place will seldom if ever fall below it. The reason for this recommendation is that the soundings on hydrographic charts will show the minimum depth of water with which mariners will need to deal. The tidal range gives them an extra margin of safety. Generally tidal range is small and so is this factor in their margin of safety. However, on Bay of Fundy charts showing a number of tidal stations, the difference between CD and Mean Water Level (MWL) in one section of the charted area may be quite different than it is in other sections. The soundings on such charts do not allow one to construct a proper three-dimensional picture of the shape of the Bay.

The Canadian Marine Sciences Branch defines mean sea level (MSL) as the level that oceanic water would assume when no atmospheric, hydrologic or tidal influences act upon it. They also use the term Mean Water Level (MWL), which is the sea level resulting in the absence of tidal influences. Up to certain limits the base level of the tides is moving up and down with the water level caused by atmospheric and hydrologic conditions. Since it is easier in the field to establish a local MSL or MWL than a level indicating the Mean High Water Mark (MHW), it is recommended that MHW be the level reached by the M2 amplitude above MSL.

It is worth noting too that whereas the Canadian Tide and Current Tables give the values of tidal differences for Higher High Waters and Lower Low Waters, the U.S. Tide Tables give the values for Mean High Water (MHW) and Low High Water (LHW). Consequently, the values of the ranges for mean and large (spring) tides differ greatly in both tables for the same locations (see Table 25). The differences between these datums are significant to the problem of establishing tidal boundaries. Canada apparently lacks a definition of MHW, a level usually taken in surveying practice as equivalent to "Ordinary" High Water (OHW) (Nichols 1983). Britain's intertidal zone is also proving difficult to map due to use of different vertical scales by Ordnance Survey and the Admiralty's Hydrographic Office, and the fact that the latter organization uses "the lowest astronomical tide" as zero point for depth (Tickell 1995). The Americans take MHW as the average of all the High Water heights observed over the National Tidal Datum Epoch. For this reason American definitions are inappropriate for direct use in Canada.

Next let's examine how an accurate determination of the Mean High Water mark can be made.

For most locations (ports) along the Canadian sea coast, the M2 tidal constituent is dominant. In fact the amplitude of the M2 tide very closely represents the average tidal conditions. When measured above the local MSL, it will indicate the level of MHW. The value of this amplitude has been determined for a great many ports along the Canadian sea coast. As for MSL, it can be easily determined by taking the average of hourly readings in calm conditions over one lunar day. The mean of these readings should indicate the MWL of the day. In order to verify if this level is close to MSL one can check at the nearest tidal recording station if the observed tides for that day correspond closely to the predicted ones. If so, the MWL can be used as a substitution for MSL. Otherwise a correction can be made by applying the difference to the measured value of MWL. Generally these differences (when they occur) are in the same order of magnitude for a number of recording tidal stations.

Since the Higher High Water (HHW), the Lower High Water (LHW), and the MWL values (see Table 26) above local CD are determined for a great number of stations along the eastern seaboard, the Mean High Water mark for most locations can be accurately determined with very little effort. A large degree of accuracy is not warranted because the landmass of the southern part of the Atlantic Provinces is steadily submerging at a rate matched by the rise of the High Water mark.

As expressed in Canadian law, "The land on the seaward side of the high water mark is prima facie held by the sovereign in common law jurisdictions." In other words such land is generally owned by the state, i.e., in Canada, the provincial or federal government.

Some Nova Scotia acts, for example, reflect this principle (Kerr 1977). The 1967 Nova Scotia Beach Protection Act states that the Governor in Council, on recommendation of the Minister of Lands and Forests, may designate as protected beach, an area which "…may include the land extending seaward from mean high water mark and such land adjacent thereto … The Minister of Lands and Forests may post signs on or near land of the Crown extending seaward from mean high water mark, warning the public that the beach is protected under this Act".

The 1975 Nova Scotia Beaches Preservation and Protection Act declares that "beach" means that area of land on the coastline to the seaward of Mean High Water mark, and that land landward immediately adjacent thereto, to the distance determined by the Governor in Council.

The 1949 Nova Scotia Marshland Reclamation Act interprets as "marshland" the land lying upon the sea coast or upon the bank of a tidal river, and being below the "level of the highest tide". Legislative practice assumes that the vertical and horizontal location of the Mean High Water mark, and the level of the highest tide are established and available all along the coastline. Unfortunately, confusion abounds in legal circles about the characteristics of the tides and the terminology used to describe them. For instance, in the upper reaches of the Bay of Fundy the tides can reach more than 8 m above MSL during strong tides, whereas during weak tides the High Water is scarcely 3 m above that level. Approximately 50% of the tides reach above the 5.5 m mark. Furthermore, the range varies from year to year and from location to location along the coast. Complicating matters is the fact that MSL changes even in the short term (Fairbridge 1987), at different rates in relation to the landmass, depending upon the location. Indeed, the concept of MSL is like Earth's Magnetic North Pole, elusive. Thus, in tide tables, the term Mean Water Level is used.

The confusion in legal circles is exemplified in "Water Law in Canada-The Atlantic Provinces" (La Forest 1972) where the three types of tide of which the law takes cognizance are described:

• Tide type #1 – high spring tide, which occurs at the two equinoxes;

• Tide type #2 – spring tide, which happens at the full moon and the change of the Moon;

• Tide type #3 – the neap, or ordinary tide, which takes place between full moon and change of the Moon, twice every twenty-four hours.

Unfortunately, because of three reasons discussed in the ensuing paragraphs, it is doubtful whether these types of tide have any significance along the eastern and western seaboards of the North American continent.

Firstly, Pliny the Elder (23– 79 A.D.) observed that the tides appear to be the strongest in the periods close to the equinoxes, namely on 21 March and 23 September. This observation may well be true for tides along the eastern side of the North Atlantic Ocean. However, Pliny could not know that the conditions along the American coastline are different. Thus, Table 27 shows that although tides on the eastern side of the Atlantic are highest near the equinoxes, this is clearly not the case for tides along the eastern or western coastlines of North America. Note that the tides in 1953 and 1975 on the eastern seaboard of North America may reach their highest levels at any month of the year due to their advance by 47 or 46 days each year as a result of perigee coinciding with either new moon or full moon. The definition of tide type #1 implies that equinoctial tides are in a way special, but this rule is in no way universal.

Secondly, the term "change of the Moon" is archaic, and designates new moon. It means, simply, that tides occurring during the period that the Moon appears in its full moon and new moon phases are stronger than average tides. This is of course due to the fact that these phases of the Moon occur when the Earth, Sun and Moon are most closely aligned. At these times, the gravitational action of the Sun reinforces the action of the Moon, resulting in higher than average tides, just as is supposed to happen in type #1 tides. During its orbit, the Moon is in one of its two quarter phases when halfway between its new moon and full moon positions. At such times the gravitational action of the Sun counteracts the dominant Moon's action, resulting in neap tides.

Unfortunately, in legal circles European tidal conditions are taken as standard and assumed to be universal. Theoretically the Sun's action is close to 46% of the Moon's gravitational influence. However, when the tides are analyzed, the actual percentage can differ depending upon the locality. Some of the measured percentages on the east side of the Atlantic Ocean are: Casablanca 37%, Rabat 40%, Lisbon 39%, St. Nazaire 35%, Flushing 27%, Bremerhaven 25%, London 27%, Liverpool 32%, Kingstown 30%. These figures are the ratios between the local tidal constituents S2 and M2 (Schureman 1941). Note that these percentages are smaller than the theoretical one. Nevertheless, they are relatively strong compared to those along the North American eastern seaboard south of Sable Island. Here the percentages are 20% or even less (15%) in the Bay of Fundy. This diminished strength of the Sun-caused tides means that the term "spring tide" loses much of its significance. In the Bay of Fundy the varying distance between Earth and Moon is the most important factor in determining tidal strength. Perigean spring tides are outstanding, while apogean spring tides hardly differ from average tides. Thus, the importance otherwise given to the term "spring tide" shows that tidal phenomena around the North American continent are not very well understood in legal circles.

Thirdly, the term "neap, or ordinary tide … which takes place … twice every twenty-four hours …" is a very confusing expression. For example, the period of 24 hours is not very precise because in most cases, the tides occur on average twice in a period of 24.84 hours, the lunar day. However, it is the term "neap or ordinary" that presents difficulty with respect to definition.

The term "High Water of Ordinary Spring Tides" is used in some British publications. The U.S. Coast and Geodetic Survey Tide and Current Glossary (1999) mentions that the term "ordinary" is not used in a technical sense by the Survey, but that term, when applied to the tides may be taken as equivalent to the terms "mean" and "average" (ASCE 1962). Thus, from that service's point of view the "ordinary tide" would be the same as the "mean tide". However a neap tide is a tide weaker than a mean or average tide. Thus the term "… the neap or ordinary tide …" is a contradiction in terms.

Tide type #3 is correct insofar as this tide takes place between full moon and the "change of the Moon" (new moon), and presumably is the weakest possible. However, the term "ordinary", has various meanings, one being "of common or everyday occurrence", another being "average or mean" (La Forest 1972). However, when a condition is variable, it is not possible for that condition to also happen all the time, yet be the average condition. It is also true that all tides equal to or greater than the neap tide will attain a certain common level. It could be argued that a neap tide is an ordinary tide. However, it certainly will not be a mediocre, medium, mean, normal, or average tide. In this case the weakest possible tide is set as a standard. No mention is made of the variability of the distance between the Moon and Earth although this is an influential cause of tidal variability, and of high tides, and can equal or exceed the effect of the changing phases of the Moon.

In order for true diurnal tide conditions to exist, the amplitude of the diurnal tide must be two to four times larger than the semidiurnal tide, depending on the time relationship between the two. However, under certain astronomical and oceanographic conditions, tides may vary from diurnal to semidiurnal during a span of little over a day. Thus only one tidal oscillation sometimes occurs in the southern parts of the Gulf of St. Lawrence and in Northumberland Strait. This is due to an extreme case of diurnal inequality when the Moon has the greatest degree of declination and the local semidiurnal components of the tide are weak. As the Moon is in its greatest declination when the Sun is near the solstices, diurnal inequalities are the most prominent during the summer and winter months. Another expression of tidal diurnal inequality, prevalent in the Bay of Fundy (Desplanque and Mossman 1998a) is the sequence of Higher High Water– Lower Low Water– Lower High Water– Higher Low Water (see section 5.3.3). This sequence can also be reversed. Other conditions are two High Waters of equal elevations, and two unequal Low Waters (for example, Northumberland Strait) and of two equal Low Waters, but unequal High Waters as observed along the eastern shore of New Brunswick and the north shore of Prince Edward Island. These types of tides are evidently not recognized in legal circles.

Clearly the three types of tides, recognized by law, make little sense and can cause a great deal of confusion when applied in court decisions. The delineation and demarcation of boundaries in tidal zones becomes an impossible task when the courts assume conditions which are non-existent. Furthermore, the relative submergence or emergence of the landmass with respect to mean sea level is usually completely overlooked when decisions are made. Coastlines along the Bay of Fundy and southern Nova Scotia are generally submerging more rapidly than the global rise of sea level indicates (Shaw and Forbes 1990). Thus, MSL and also the high tide levels are rising at rates that vary 2 or 3 mm/yr to rates exceeding 8 mm/yr, as claimed for some coastal areas of Maine (Scott and Medioli 1979). In areas where beaches have a very gentle slope, such a rise can make a large annual horizontal shift of the High Water mark. Here the establishment of a permanent property boundary, based on the Mean High Water mark, is totally unrealistic.

The definition of the private-state boundary in common-law countries has its genesis in a 17th Century treatise, "De Jure Maris", by Sir Mathew Hale, chief justice of King's Bench, the highest court in England (Hale 1667). He devised the definitions of the three types of tides discussed above. Sir Mathew wrote his treatise in the same year that the 24-year old Isaac Newton conceived the idea of universal gravitation. At the time, Hale was the chief baron of the exchequer, and probably not yet knowledgeable of the new tidal theories that would follow from Newton's work.

Hale's private studies included investigations in classical law, history, the sciences and theology. He exercised considerable influence on subsequent legal thought. Small wonder, therefore, that his misconceptions on the nature of tides endure. A leading case in point was the precedent set in 1854 by the British decision "Attorney-General vs. Chambers". The facts of the case are not significant but the legal interpretation of tidal terms is particularly important for the Maritime Provinces of Canada. The court was asked to determine the legal rights of the parties, a matter which depended entirely upon the interpretation of the term "high water mark". In its decision the Court emphasized the significance of Hale's doctrine, noting that: "All the authorities concur in the conclusion that the right is confined to what is covered by "ordinary tides", whatever be the right interpretation of that word."

It is clear that the Lord Chancellor had problems with the term "ordinary". The Court defined the ordinary tides as: "… the medium tide between springs and neaps … It is true of the limit of the shore reached by these tides that it is more frequently reached and covered by the tide than left uncovered by it. For about three days it is left short, and on one day it is reached. This point of the shore therefore is about four days in every week, i.e. for most part of the year reached and covered by the tides … The average of the medium tides in each quarter of a lunar revolution during the year gives the limit of all usage, to the rights of the Crown on the seashore."

In other words, the High Water mark is calculated by averaging the medium high tide marks for each week in the lunar cycle during the year. It is clear that long term cycles such as an 18-year cycle, were not taken into consideration. Reference to Table 28 shows that this can result in different interpretations of the position of the High Water mark, depending on what year is taken into consideration.

In common law, private ownership of land ends at the mean high water mark, and title to the area between the MHW and low water marks (so-called tidelands) is held by the sovereign states. In a leading U.S. Federal case, the United States Supreme Court referred to: "the mean high tide line which is neither the spring tide nor the neap tide, but the mean of all the high tides." Unfortunately, in other cases a different line is used, for example the vegetation line, the highest winter tide, and the mean higher high tide; a few states use the low water line as boundary (Bostwick and Ketchum 1972). Quite apart from the choice of tidal cycle, this lack of standardization bedevils tidal boundary issues on national and international scales. A long-term cycle was taken into account in the U.S.A. in the so-called "Borax" decision (United Sates Supreme Court 1935). In this landmark case in 1935, the United States Supreme Court ruled that "an average 18.6 years of tidal observations should be used to determine the datum elevation". Promoted in the U.S.A., this has been described as a progressive decision which incorporates the most accurate methodology for determining tidal boundaries. However the definition does not deal with submergence of the landmass in relation to mean sea level, difficulties in areas with diurnal or strongly mixed diurnal and semidiurnal tides, and non-tidal influences. Furthermore the 18.6 year cycle is based on the period of revolution of the Moon's nodes and during this period the diurnal inequality of the tides varies in strength.

The 1949 "Tide and Current Glossary" prepared for the U.S. Coast and Geodetic Survey defines MHW as "the average height of the high water over a 19-year period". For shorter periods of observation, corrections are applied to eliminate known variations and reduce the result to the equivalent of a mean 19-year value. The 19-year cycle, the so-called "Metonic Cycle", was chosen because 235 lunations occur almost exactly in 19 mean solar years, and this is in step with the Julian calendar (Greulich 1979). Named for its discoverer, the Greek astronomer Meton (432 B.C.), this cycle was used by the Nicene Council in 325 A.D. to fix the date of Easter.

Unfortunately, in legal cases where tidal heights are important, the emphasis placed on cycles of 18.61 (and 19 years) may not be warranted. This is the case in regions like the Bay of Fundy where diurnal inequality of the tides is of such minor importance that it can be virtually ignored among variations caused by the coincidence of perigean and spring tides. Table 22 shows the situation with respect to long term cycles of conditions leading to stronger and weaker than normal tides. Note that if one used 246 anomalistic months (18.558 years), 247 anomalistic months (18.634 years), or 252 anomalistic months (19.011 years), the multiples of other types of months are not nearly as closely matched as with 18.03 years. Thus, the courts may easily be led astray by misapplying astronomical data.

In theory, the height reached at High Water is the full amplitude above MSL, which is defined as the level that oceanic water would assume if no tidal or atmospheric influences are acting upon it. On land, the datum used by geodesists, surveyors, and engineers, is the Geodetic Survey of Canada Datum (GSCD, or GD). This datum is based on the value of Mean Sea Level prior to 1910 as determined from a period of observations at tide stations at Halifax and Yarmouth, N.S., and Pointe au-Père, Québec, on the east coast, and Prince Rupert, Vancouver and Victoria, British Columbia, on the west coast. In 1922 this was adjusted in the Canadian levelling network. Because in most areas of the Maritime Provinces the landmass is submerging relative to MSL, geodetic datum drops gradually below Mean Sea Level. Thus, local MSL at present is about 280 mm (0.9 ft) higher than G.D. However, there is a dearth of data, and no one is certain what the exact difference is between GSCD and MWL at different stations. This situation is troublesome for engineers and biologists who need to know the proper relation between the two datums at particular places and it is no less likely to trouble legal minds.

A legal decision regarding tidal issues in the Bay of Fundy resulted from the 1962 to 1965 trial proceedings in the case of Irving Limited and the Municipality of the County of Saint John vs. Eastern Trust Company. This trial (for details see Desplanque and Mossman 1999a) highlighted a wholesale incorrect use of tidal data and definitions. Central to this case was the definition of Mean High Water (MHW). In brief, it transpires that no matter what is chosen as a MHW elevation, the value employed in the Irving case has no valid statistical basis.

In retrospect, it is instructive to consider the manner in which MHW (and HHW) vary according to the increased tidal range toward the head of the Bay of Fundy. This will be illustrated for the Minas Basin, where seven principal tidal constituents account for more than 90% of the total variability of the tides (see Table 3). Note that while the amplitudes of these semidiurnal tides increase toward the head of the Bay and Basin, the diurnal amplitudes remain virtually constant at 0.2 meters. The shallow-water tides are relatively small, but will certainly become large for locations within estuaries. No constituents are available for the tides within the estuaries. However the diurnal tides probably will not be altered when progressing into the estuaries. Semidiurnal amplitudes are clearly a function of the distance from the port of Saint John where the principal tidal hydrographic station is located. Thus as shown in section 5.3.2, the range of the dominant semidiurnal tides in the Bay of Fundy increases exponentially as they advance, at the rate of about 0.36% per kilometre. This allows local tidal range to be estimated very accurately, with reference to Geodetic Datum, from which follow realistic estimates of MWL and HHW.

The above relationship is, among other things, relevant to proposed tidal power schemes in the Bay of Fundy. These might very well modify the tidal regime (Greenberg 1987) and conceivably lead to international legal conflicts. Indeed a model used during the 1977 studies for Fundy tidal power development concluded that should a tidal power plant be built in the Minas Basin, the amplitude of average tides would increase by 0.15 m along the Gulf of Maine coastline of New England, and up to 0.25 m along New Brunswick coastline. Whatever the truth of this dire prediction, the fact is that coastal submergence is a reality that must be faced due to continuing sea-level rise. No protest, political or otherwise, can alter this situation. Further, if MHW is to be established, one has to make the choice between a permanent level linked to a certain year such as the Geodetic Datum of Canada, or a level which moves with changes in sea level due to geomorphological influences triggered by eustatic sea-level rise.

When the International Court of Justice set the boundary line between Canada and the United States through the Gulf of Maine, in October 1984, people on both sides of the border protested that the Court had decreased the area of the Gulf of Maine to such a degree that many jobs would be lost in both countries. If that same Court has to make a decision about changing characteristics of the tides and water levels in the Gulf of Maine-Bay of Fundy system, our knowledge of these characteristics had better be able to stand up to close examination.

An interdisciplinary approach is needed in the matter of tidal boundary delimitation. However, whatever the roles of lawyers, surveyors and scientists, the terms MHW and MLW need to be unambiguously defined in accordance with modern tidal and astronomical principles and terminology. European tidal conditions taken as standard in legal circles are not universal and consequently do not permit recognition of various common types of tides such as those governed by diurnal inequalities. Establishment of tidal boundaries is impossible when courts assume non-existent conditions.

Tidal datums are commonly misconceived to be fixed planar levels rather than undulating time and space-dependent surfaces. Yet, except for non-astronomical factors such as storm surges, tectonic activity and postglacial sea-level rise, tidal prediction with reference to specific tide levels can be made with a high degree of confidence. In fact there is a tendency to give greater credence to the predictions and the tidal constituents on which they are based than to the observed tides. However, if the procedure is correct, the average of the predicted and observed heights should be close together. The more accurate establishment of tidal datums hinges on improved prediction, which in turn requires more reference ports, updated tidal information and improved surveying techniques. Where numerous ports exist as along the eastern Canadian seaboard and the M2 tidal constituent is dominant, the amplitude of the M2 tide as measured above local MSL, closely approaches MHW. Also, the exponential increase in amplitude of the semidiurnal tides in the Bay of Fundy allows MWL, HHW and the local tidal ranges to be predicted quite accurately in this home to the world's highest tides.