The April 1996 Rockland Landslide

Description and Geology of the April 16, 1996 Landslide

The pre-landslide bluff profile was relatively steep, dropping more than 40 feet (13 m) vertically over a horizontal distance of 80 feet (25 m) (Figure 3). Note that the pre-landslide bluff profile was never measured at the actual landslide site. The description here and in Figure 3 is for a nearby bluff about 165 ft (50 m) northwest of the landslide area. Prior to the April 16 event, a small landslide involving approximately 2,000 cu yds (1,530 cu m) had occurred on the upper part of the bluff on February 16, 1995 (reported to Stephen Dickson by William Eaton, next-door neighbor to the Gerrishes), remnants of which can be seen in Figure 2C. The remainder of the bluff was covered with oak and poplar trees and shrubs. The land surface above the bluff was relatively flat and covered by a lawn with houses (Figure 2A).

Following the event, the landslide site was surveyed with a total survey station (EDM) on April 24, 1996 (Figure 4), and a log of the stratigraphy of the exposed bluff was recorded (Figure 5). Numerous ground photographs were taken from around the landslide and on its surface on April 24 (Appendix A). Three test borings were drilled by Northeast Diamond Drilling and R. G. Gerber, Inc., around the landward edge of the slip face between April 26 and May 1, 1996 (Appendix B).

The landslide affected about 0.6 acres (0.25 hectares) of the upland, completely removing one house and largely destroying the second. The resulting headscarp was nearly semicircular, with the remnants of one house protruding from the northwest side (Figure 2D and Figure 4). The erosional scarp resulting from the landslide was about 210 feet (65 m) across and involved upland more than 130 feet (40 m) from the former edge of the bluff. The headscarp dropped almost vertically for 26 feet (8 m) near the road and on the east side, and for about 13 feet (4 m) on the northwest side.

The landslide deposit occupied 3.5 acres (1.42 hectares ) and projected 590 feet (180 m) seaward onto the tidal flat, or about 450 feet (140 m) seaward of the former top of the bluff. The surface of the landslide consisted of a series of intact blocks 3 - 10 feet (1-3 m) across, which near the head of the scarp were tilted landward. Trees and shrubs were still sitting in the blocks, even though they were tilted both landward and seaward as much as 45 degrees by the event. Many of the blocks were striated on their undersides. In the central part of the slide, the blocks were not arranged coherently across the slide deposit, but were disturbed in many places, so that no block was more than about 16 feet (5 m) long. Depressions, often filled with water, separated many of the blocks for the first 330 feet (100 m) from the headscarp.

The outer area of the slide deposit was more disturbed than the inner area. No large blocks were found here, and the surface of the slide consisted of a rubble of small blocks of mud 1.5 feet (0.5 m) long or less. The outer edge of the deposit formed a scarp up to 6 feet (2 m) high along the sides, but formed a more gradual, tapered shape at the toe (Figure 2C). Boulders were common around the periphery of the toe of the slide, as well as displaced blocks of salt marsh peat.

A description of the sedimentary section exposed in the northwest headscarp is shown in Figure 5. There is 1 foot (30 cm) of soil at the surface. Near the surface, the sediment displays obvious horizontal layering of interbedded clayey silt and 1 - 3 inch (2 - 5 cm) thick layers of coarse sand. Vertical fractures coated with purple-black (manganese oxide?) stains are found in the upper 6 feet (2 m). Below the upper 6 feet, the section is comprised of massive to faintly-bedded fine-grained sediment (predominantly clayey silt). Large stones are scattered throughout the material to the base of the section.

Bedrock Geology in the Area of the April 1996 Rockland Landslide

Rock Types. The bedrock, or ledge, which forms the natural, solid foundation of the northern side of Rockland Harbor consists of very old metamorphic rock. The bedrock has been divided according to rock type into two formations, the Ojier Point Formation and the North Haven Formation (Osberg and others, 1985). The Ojier Point Formation consists of andalusite-garnet-quartz-mica schist thinly interlayered with light gray biotitic quartzite. Both the schist and the quartzite contain a strong metamorphic foliation that was folded and contorted during formation of the Appalachians.

The North Haven Formation consists of layered greenstones and feldspar-rich gneisses. These rocks probably originated as mafic and felsic volcanic rocks or tuffs, but were recrystallized and deformed into metamorphic rocks at the same time as the Ojier Point Formation. The original rocks may be as old as Precambrian (Precambrian Z; 570 - 800 million years ago). They were metamorphosed and deformed along with other rocks of the Penobscot Bay region due to heat and pressure in the earth during a geologically active time, probably in the Silurian Period (about 430 million years ago).

Internal Bedrock Structure. The combination of ancient small folds and faults has produced a complex bedrock structure in which the rock foliation has different orientations in different places. The foliation strikes between northwest and northeast, and dips moderately steeply toward the east. Minor fold hinges trend toward the north-northeast (N20^oE to N30^oE) with shallow plunges. Predominant fracture sets in the bedrock of northern Rockland Harbor are oriented west-northwest (N70^oW) and north-northeast (N35^oE). Such fractures are common in the bedrock of New England, presumably related to the opening of the Atlantic Ocean during the Mesozoic Era. It should be emphasized that no active faults have been identified in Maine. Weston Geophysical Observatory, Massachusetts, which monitors seismic activity in New England, reported to the Maine Geological Survey that there was no earthquake or seismic activity in coastal Maine at the time of the Rockland landslide.

Shape of the Bedrock Surface. The depth to bedrock, commonly referred to as ledge, depends on the thickness of any overlying sediment and the shape of the bedrock surface. The bedrock surface has an irregular shape because it has been eroded unevenly through geologic time. Even where the earth's surface is fairly flat such as along Samoset Road, the depth to bedrock is not everywhere the same due to the irregular shape of the bedrock surface beneath the overlying sediment. There are places along the shore where the bedrock is exposed in outcrops at the earth's surface. These places are indicated on the map in Figure 6. Bedrock of the North Haven Formation is well exposed in the ledges toward the end of Jameson Point and along the shore near the breakwater. Farther to the northwest, where the bedrock belongs to the Ojier Point Formation, there are gaps where bedrock is not exposed, especially in the coves. In these coves, the bedrock surface was more deeply eroded at some time before the end of the last ice age. That surface is now at an elevation below sea level, and the eroded depressions on the bedrock surface in those areas are infilled by glacial and postglacial sediments.

The thickness of surficial deposits in northern Rockland Harbor and along Samoset Road varies in part due to the depth to bedrock. From the shoreline bedrock outcrops alone it is difficult to judge precisely the depth to bedrock inland from the shore. In general, bedrock is inferred to be close to the ground surface at the south end of Jameson Point because of the abundance of shoreline outcrops, the presence of a bedrock outcrop on the Samoset Golf Course (beside the 5th fairway), and a report by the Jameson Point groundskeeper that ledge was encountered during excavation for the two condominiums farthest toward the southeast end of Samoset Road. There are no bedrock outcrops farther to the northwest along Samoset Road or along Waldo Avenue to the west.

Seismic reflection profiles 1500 feet (460 m) offshore of the coast reveal a prominent sub-bottom acoustic reflector ranging from less than 16 feet (5 m) to greater than 65 feet (20 m) beneath the seafloor (Figure 7). This reflector is the acoustic basement and its high relief and strong, coherent return appears like a reflector recognized in previous work as bedrock (Belknap and others, 1989). The bedrock reflector is shallowest seaward of bedrock outcrops along the coast at Jameson Point, and descends to greater than 65 feet (20 m) beneath the seafloor seaward of the 1996 landslide. Bedrock consistently remains shallow seaward of coastal bedrock outcrops and is more deeply buried seaward of coastal embayments with eroding bluffs and past landslides.

In detail, the bedrock surface in this area is irregular due to the combination of a north-northeast trending bedrock foliation, northeasterly and northwesterly-trending fractures, and a southerly ice flow direction during the last glaciation. Any of these factors may affect the shape of the bedrock surface. In order to determine the depth to the bedrock surface inland from the shore, especially in the areas of the 1973 and 1996 landslides, the Maine Geological Survey conducted a seismic refraction survey in May, 1996. The results of this survey are found in Appendix C of this report. These data and the depth to bedrock from the boring logs (Appendix B), and records reported to the Maine Geological Survey from water well drillers provided information to prepare a bedrock surface contour map (Figure 8).

Figure 8 represents the elevation and shape of the bedrock surface in the landslide area along Waldo Avenue and Samoset Road. The areas of higher bedrock elevations generally correspond to the areas where bedrock outcrops occur along the coast (Figure 6). These areas have a shallow depth to bedrock, and hence a relatively thin cover of overlying sediment. Those areas where no outcrop is present along the coast correspond to low areas on the bedrock surface, and at the coast are where eroding embayments of thick sediment overlying the bedrock are found.

Surficial Geology of the Area Surrounding the April 1996 Rockland Landslide

Surficial Sediments in the Vicinity of the Landslide. Surfical geologic materials (such as sand, gravel, silt, and clay) are the unconsolidated sediments that lie on top of the bedrock over much of New England. In Maine, these sediments form deposits left by glacial ice, water, wind, and organic processes. Most of them were deposited within the past 25,000 years, during and after the most recent glacial episode.

The Maine Geological Survey has mapped the surficial geology for much of Maine. Current information is available at different levels of detail for different parts of the state. New maps and information become available as they are produced. The surficial geology of the Rockland quadrangle was mapped at a reconnaissance level by Smith (1974). Two principal types of glacial deposits are found at or near the landslide site: till and glacial-marine clayey silt.

Till is a heterogeneous mixture of rock debris that was deposited directly from glacial ice. It may contain sediment ranging from silt and clay-size particles to large boulders. Till is often the lowest and oldest surficial stratigraphic unit in any particular area, and commonly rests on top of the bedrock. It has been mapped at the ground surface on the higher terrain west of U.S. Route 1 in Rockland, usually at elevations above 100 feet (30) m; but closer to the coast, the till is largely concealed beneath the younger mud deposits. Shoreline exposures on the point just southeast of the slide show the clayey silt directly overlying bedrock, so till deposits are not present over the entire area. However, the three boring records (Appendix B) drilled near the head of the landslide showed variable thicknesses of till resting on bedrock, ranging from less than a foot (30 cm) to about 10 feet (3 m) thick. Based on field observations, the landslide itself appears to have occurred entirely within the glacial-marine clayey silt.

The glacial-marine clayey silt is part of a widespread geologic unit called the "Presumpscot Formation" (Bloom, 1960, 1963), which is very common in lowland areas and river valleys of coastal Maine and is named after exposures along the Presumpscot River in Portland. It is commonly called "clay" or "blue clay" by well drillers and the general public. Actually, this generally massive-appearing formation contains a mixture of clay, silt, and sand particles in varying proportions, and often in alternating layers of differing thickness and particle size range. Clay and silt size particles are typically dominant, being mixed together to texturally form a clayey silt or silty clay, or what might simply be called "mud." However, in geotechnical engineering, the term mud implies an extremely soft consistency (Sowers and Sowers, 1970). Hence, in keeping with popular and familiar usage of this term, the Presumpscot Formation will be called clay here.

The Presumpscot clay is made of finely pulverized rock material that settled to the sea floor from muddy plumes of glacial meltwater that flowed into the ocean (which flooded large areas of southern Maine during the recession of the last ice sheet). Thus, the mineral composition of most of this glacial sediment consists of common minerals such as quartz, feldspar, and mica (Kelley, 1989).

The Presumpscot Formation is limited to low-lying areas that experienced marine submergence in late-glacial time (described below), so the glacial-marine clay generally occurs only at elevations below 200 to 300 feet (60 - 90 m) in southern Maine (and at even lower elevations near the landslide, as noted above). Clay thickness is usually measured in tens of feet, but may exceed 100 feet (30 m) in some valleys and coastal areas. The clay deposits extend out under the ocean, where they are buried by younger marine sediments (Kelley and Belknap, 1991).

Where the landslide occurred , the clay thickness was 35 to 45 feet (10 - 14 m) thick. The upper part of the clay shows a typical brownish-gray color resulting from weathering near the original ground surface, while deeper parts of the deposit have the gray or blue-gray color of fresh, non-oxidized Presumpscot Formation. The transition from brownish-gray to blue-gray color occurred at a depth of 6 feet (2 m) in the exposed headscarp (Figure 5) and at depths of 8 to 10 feet (2.5 - 3 m) in the three borings near the head of the slide (Appendix B). Blocks of clay exposed in the lower part of the slide deposit contain many stones that originally were probably dropped from icebergs when the ice sheet was retreating from the area. These same blocks also contain fossil shells of a marine mollusk species called Portlandia arctica, which lived in cold muddy waters close to the ice margin.

Glacial-marine clays whose age and origin are similar to the Presumpscot Formation occur in the St. Lawrence Lowland of eastern Canada (Elson, 1988) and other recently glaciated parts of the world. These clays commonly are vulnerable to landslide hazards, particularly where local conditions of slope, relief, drainage, toe erosion, and other factors conspire to reduce the stability of the deposits (Amos and Sandford, 1987).

Glacial and Postglacial Geologic History of the Landslide Area. It is not certain how many glacial episodes occurred in Maine during the Pleistocene "Ice Age," because the last glaciation removed much of the evidence of earlier glacial action. The most recent ice sheet covered southern Maine between about 25,000 and 14,000 years ago. Rock debris that was eroded and dragged beneath the slowly flowing ice caused the bedrock surface to be ground smooth in many places. Glacial scratches and grooves are seen on the ledge surfaces along the shoreline near the landslide, with especially good examples on the point just south of the slide. In this area, the grooved rock surface records a complex history of ice-flow directions, resulting from shifting flow patterns over time and local deflections of ice flow by the knobby rock surface. Overall in the landslide area, the dominant flow direction in late-glacial time was apparently slightly west of south, averaging S5^oW (azimuth 185^o).

The weight of the ice sheet caused a depression of the earth's crust by hundreds of feet, and the land did not immediately rise back to its original elevation when the glacier margin withdrew from the coast. Therefore, the ocean flooded the coastal lowland (including the landslide site), and marine sediments were deposited in large areas where there is now dry land. Meltwater streams issuing from the ice deposited sand, gravel, and clay into the ocean in these areas. The water-laid sediments blanketed much of the layer of till that had been previously deposited beneath the ice. Periodically the edge of the glacier stood still or advanced slightly, and ridges of sediment (called "moraines") accumulated along the ice margin.

Soon after the Presumpscot clay was deposited, the rising land forced the sea to recede far offshore. As the world's glaciers continued to melt, sea level gradually rose over several thousand years, first more rapidly and later more slowly. Sea level is still rising to this day. During this period of ongoing sea-level rise, erosion of glacial sediments has been occurring, causing shoreline retreat and landslides along the Maine coast.

Seismic Refraction Data and Interpretation

Seismic Refraction Methods. During the period from May 6 - 16, 1996, the Maine Geological Survey conducted seismic refraction surveys in order to gather subsurface geologic and hydrogeologic information in the area of the Rockland landslides of 1996 and 1973 (Figure 9). Site survey lines were located roughly perpendicular as well as parallel to the directions of the slides. Actual locations selected were made on the basis of access and surface suitability.

Subsurface Profiles from Seismic Data. In total, eighteen 12-channel lines were seismically surveyed, yielding 5 single-line profiles and three multiple-line composite profiles. Interpreted hydrogeologic profiles resulting from the seismic-refraction surveys are presented in Appendix C. Distances shown on the x-axes are measured from geophone number 1 for each profile. The y-axes show the elevation or altitude above sea level. Note that the vertical scale of the diagrams has been expanded to 6 times the horizontal scale to better illustrate changes in elevation. The geologic materials at depth have been inferred from their measured seismic velocities. Since there is a greater concentration of redundant bedrock data in the middle of the lines than at the ends, the bedrock surfaces depicted in the central portions of the lines reflect subsurface conditions with a greater amount of certainty than do the ends of the lines. Abrupt changes in the interpreted bedrock surface at the extreme ends of the lines might not reflect actual bedrock topography. In most cases, overlapping and intersecting lines gave profiles in close agreement with each other. The exception is where the southeast end of profile ROCK-18 approaches the middle of the multiple-line profile ROCK-1517 along Waldo Avenue. Although these lines did not actually intersect, they were only separated by about 30 feet (10 m), the width of the road, yet the elevation of the bedrock surface shown on the two profiles differs by about 18 feet (6 m). This apparent discrepancy remains unresolved, but suggests that the simple interpretation presented for the south end of profile ROCK-18 may be incorrect due to some subsurface complication in that area. Profile ROCK-1517 is based on three lines that gave results in good agreement, so for purely statistical reasons the data near the middle of profile ROCK-1517 is more reliable than the data near the end of the single line ROCK-18.

Water Table. The water table is the underground surface that separates unsaturated material above, from water-saturated material below. Because seismic waves travel faster through saturated material, the water table depth was detected by the seismic survey. The water table is indicated on the profiles in Appendix C by a thin black line with a small black triangle pointing to it. Below the water table, the glacial deposits (mainly clay) are fully saturated with ground water.

The depth at which the change from saturated clay to unsaturated clay occurs is difficult to verify directly by sampling because water is transmitted so slowly through the clay. The seismic data is supported, however, by the observed features at the landslide site. Where the landslide occurred, the fresh exposure in the bluff face (Figure 3) and the observations from drilling (Appendix B) showed a distinct change between the upper oxidized clay with dessication cracks, and the underlying massive, blue-gray clay. This change is attributed to unsaturated versus saturated clay. The observed change from unsaturated to saturated clay occurred at 6 to 10 feet (2 - 3 m) depth, matching well the depth at which the measured seismic velocity changed. Therefore, the field observations confirm that the line depicted in the seismic profiles is the water table.

The depth to the water table shown in the seismic profiles is between 4 and 7 feet (1 - 2 m) below the ground surface. As is normally the case in New England, changes in the water table from place to place closely follow changes in the ground surface elevation. Over areas where the ground surface is flat, the water table is approximately horizontal. The seismic profile along Samoset Road that crossed the area of the 1996 landslide is shown in profiles ROCK-41 and ROCK-1567 (Appendix C). The head of the 1996 landslide is in the vicinity of the 100 foot distance mark on profile ROCK-1567. These profiles show nothing unusual in the water table where the landslide occurred.

No tests were done to assess movement of ground water. Even so, the seismic data alone provide important information about ground-water flow. For example, the ground-water table beneath Samoset Road (Appendix C; ROCK-41, ROCK-1567) is at higher elevation than the water table beneath Waldo Avenue (Appendix C; ROCK-1517). Hence, the ground-water flow will be along gradient, from higher to lower potential, or in this case from the higher elevation landward areas along the roads toward the lower elevation along the coast.

Marine Geology

Rockland Harbor occupies a curved embayment along the western shore of Penobscot Bay. This is within the Island-Bay coastal compartment (Kelley, 1987), an area defined by estuaries sheltered by numerous islands. Vinalhaven, North Haven, and Islesboro Islands protect Rockland Harbor from waves from the southeast, east, and northeast, respectively. Owls Head peninsula blocks waves approaching Rockland from due south. The mainland is rocky and steep on the southern side of the harbor, with up to 200 feet ( 60 m) of local relief. To the north and west the relief is subdued except on the shoreline itself where a 20 to 40 foot (6 - 12 m) scarp rims the harbor.

Tides in the area average 9.8 feet (3 m). The wind at the Rockland Coast Guard Station comes most often from the western quadrants; the strongest gusts are from the northwest and northeast. No studies have been made of wave heights in the harbor, but on the basis of the wind data, it appears that the largest waves would be derived from east-northeast directions. It was probably for that reason that the breakwater was built in 1904 along a generally north-south orientation from Jameson Point (Figure 1). The breakwater extends 4300 feet (1.3 km), almost halfway across the opening to Rockland Harbor, and may inhibit tidal exchange somewhat compared to conditions prior to its construction.

Sea level has risen in the outer Penobscot Bay area from around 10,500 years ago (Barnhardt and others, 1995). Early rates of rise were rapid, about an inch (2.5 cm) per year, but slowed appreciably in the middle to late Holocene. By 1,000 years ago, sea level had slowed its rate of rise in southern Maine to less than 1 mm/yr until this century (Kelley and others, 1995). The tide gauge in Bar Harbor has recorded a 2.6 mm/yr rate of sea-level rise there since 1947 (Lyles and others, 1988).

Coastal bluffs of glacial-marine sediment have retreated since sea level began to rise. The contemporary bluffs are all retreating through a variety of mechanisms, including landslides, in response to the long-term rise of the sea (Kelley and others, 1987). The rate of bluff retreat on the Maine coast varies due to local conditions such as wave exposure and the physical properties of the bluff, but probably ranges from about an inch to a foot (2 - 30 cm) per year. The northern shore of Rockland Harbor was identified as a hazard zone for bluff erosion by Kelley and others (1989) because of the high bluffs of glacial marine clay, scalloped shoreline shape, and past history of landslides there.

Contents Introduction Description Other slides Factors Potential Action Summary References Appendix A Appendix B Appendix C Plate 1

Last updated on October 6, 2005