The April 1996 Rockland Landslide

Landslide Potential along the North Shore of Rockland Harbor

While none of the north shore of Rockland Harbor is without risk of future landslides, our work suggests that some areas have more potential than others for a catastrophic slide of the type that occurred April 16, 1996. Although exact prediction of such landslides is not possible, the primary factors which bear on the potential for catastrophic landslides in the area are the following: (1) thick section of glacial marine clay (Presumpscot Formation), (2) high and steep bluffs lacking lateral support, and (3) water, especially ground-water conditions, but also precipitation and surface drainage. Other factors may only be minor contributors to increased potential for catastrophic landslides. However, the less dramatic but continual gradual erosion, small-scale slumping, and bluff recession that has occurred in the past, will continue over time just as it has been along the north shore of the harbor.

Thick Section of Clay

As the thickness of clay increases, the weight of this thickness, or overburden, may exceed the strength of the clay material itself. Technical staff from R. G. Gerber, Inc., made vane shear measurements in several boreholes around the April 1996 slide area. These tests used a Geonor 65 mm x 130 mm vane coupled with lightweight actuating rods and torqued with a calibrated torque wrench. The test is described in ASTM Special Technical Publication 1014 (Richards, 1988) and is a standard method for measuring the in situ undrained shear strength of silts and clays.

The results of the vane shear measurements show the strength of the clay at various depths (Figure 11). The line shown on the diagram shows the expected strength of normally consolidated clay of the Presumpscot Formation. At depths greater than 25 feet, the measured values are close to the expected values, indicating that the clay at depth is normally consolidated. At depths less than 25 feet, the measured shear strength is much greater than expected for normally consolidated clay, indicating the clay near the surface is "overconsolidated."

This might be explained by natural weathering of the clay surface over time, which can pull the clay structure together, giving it added strength. The brown color of the clay and the staining along vertical fractures in the clay which extend from the surface to about a 13 ft (4 m) depth are the result of normal weathering processes (oxidation and dessication). Hence, the apparent overconsolidation of the clay can be attributed to these weathering processes. Also, changes in depth of the water table in the clay may have occurred in the geologic past during dryer and wetter times thousands of years ago, and may, in part, have contributed to the overconsolidated nature of the upper part of the clay.

The expected strength has been calculated by multiplying the estimated overburden stress by a value of 0.2 (the undrained strength ratio). The estimated overburden stress, or pressure, is caused by the weight of the overlying material. This weight has been estimated by assuming an average density for the clay. Laboratory tests on Presumpscot clay in southern Maine typically give an undrained strength ratio of 0.2 (unpublished data, R. G. Gerber, Inc.). From this, we conclude that those areas underlain by 25 or more feet (8 m) of clay, when coupled with a steep bluff slope, have a higher risk for landslides than those areas underlain by thinner clay sections. Yet, localized factors such as bluff slope inclination and drainage do, however, preclude the use of 25 feet of clay as a hard and fast rule for landslide prediction. Also, subtle variations in the clay moisture content, strength, and other factors will determine whether a single block of clay will fail during the landslide, or whether there will be several blocks failing in a retrogressive manner.

Nonetheless, our seismic work along Samoset Road and Waldo Avenue show several areas beneath which the clay is 25 feet (8 m) or more thick. There are several areas of particular concern (Figure 12). One is along Waldo Avenue from just west of its intersection with Samoset Road and extending westward. Along this section the ground surface is at about 45 feet (14 m) above mean sea level while the bedrock surface deepens to 20 feet (6 m) below mean sea level for a total thickness of 60 to 70 feet (18 - 21 m) (also see Figure 8). Our seismic work does not extend westward beyond the intersection of Waldo Avenue with Glen Road, but at this intersection the thickness of the clay is greatest. It is likely that the thickness of the clay decreases westward from there to a place immediately north of the prominent rocky peninsula (Figure 12).

Other areas of concern are along Samoset Road (Figure 12). There are three embayments along the coast which correspond to troughs on the bedrock surface in which the thickness of the overlying clay is greater than 40 feet. The first is southwest of the intersection with Waldo Ave, the second is near the 1996 slide, and the third near the intersection with the old dirt access road to the Samoset resort. These embayments are indicated by arrows on Figure 12.

We can further generalize to say that areas landward of exposed ledge along the shore generally have thinner covers of clay than those areas which lack such exposures. Where detailed seismic data are lacking, such a generalization may be useful in assessing relative risk among neighboring areas.

High Bluffs Lacking Support

High and steep-sloped bluffs lacking lateral support at the base of the slope along the shore are a factor along the entire northern shore of Rockland Harbor. The bluff height and slope decrease approaching the southeastern end of Samoset Road near Jameson Point, so at present the relative risk of a catastrophic landslide decreases toward the breakwater. Slopes are more gradual there, and clay thickness and overall elevation of the land is less. However, small scale slumping, gradual erosion, and bluff recession will continue along the shore here. Potential for slides may increase over time, depending on erosion and slope steepness.

Water

The previous section cites several published references describing the role of water and precipitation in initiating ground failures. It is likely that the melting winter snowpack of 1996 followed by Spring precipitation may have resulted in high ground-water levels and may have contributed to the conditions leading to the April 1996 landslide. The weight of this water and its contribution to lowering the strength of the clay may be factors in the slide. A mid-winter thaw may have been a contributing factor to the 1973 slide along Waldo Avenue.

Also, it is suspected that an elevated ground-water head may have existed in glacial till or sandy deposits near the base of the slope at the time of the landslide. Whether such a sandy layer existed there is not known because the base of the bluff was disrupted and is now covered by debris. It was not found on the shore in the area of the slide, but sandy deposits were found about 20 feet below the surface of the landslide in some hand borings made where the old shore line would have been. This suggests that there may have been a sand layer several feet below the beach. Such a sand layer overlain by thick clay would be expected to develop a high pore pressure. An artesian condition in glacial till at the toe of the bluff slope would have reduced the lateral support for the bluff at the shoreline. This elevated head of water would have dissipated in the landslide area relatively soon after the slide occurred so it could not be measured after the fact. However, geologists from R. G. Gerber, Inc., did note slight upward pressures on the slope inclinometer casings at the top of the landslide during their installation, and slight artesian conditions in hand borings drilled in the 1996 landslide area. Of course, it is not known whether an excess head existed in the sand before the landslide, only that there was evidence for it after the slide.

From our seismic work, it is clear that modifications to drainage and ground water along Glen Road where major retail centers recently have been built were not factors in the 1996 landslide. The water table along Waldo Avenue at the time of our survey was lower than along Samoset Road in the area of the slide, and the water table gradient along Waldo Avenue (Appendix C; ROCK-41) is away from Samoset Road. These observations indicate that ground water in the Glen Road area moves down gradient toward the ocean and not toward the area of the 1996 landslide.

Modifications to drainage at the Samoset Resort are not likely to have contributed to the initiation of the 1996 landslide. This area was a stream and swamp prior to development into small ponds. The area lies on the north side of a minor drainage divide which generally directs water to the northern side of the resort property, away from Samoset Road. Furthermore, behind the houses on the northeast side of Samoset Road are several small intermittent wet areas which had standing water at the time of our investigations. These areas drain to the northeast. None of these water bodies showed sudden changes in water elevation in response to the landslide, as would be expected if there was a direct hydraulic link between them and the landslide failure surface. The water table does not change significantly in the vicinity of the slide, and there were only minor amounts of ground water seeping from the head scarp of the slide during our field surveys. In fact, in this area as in most areas of coastal Maine, the natural water table is within several feet of the ground surface for most of the year so that most of the clay deposits are saturated most of the time.

Apparently, the two houses involved in the 1996 slide were still using private septic systems. The addition of effluent from these systems may have increased the load on and reduced the strength of the underlying materials. Also, it should be noted that in the area of the February, 1995 landslide (adjacent to the April 1996 slide), ceramic drainage pipes protrude from the top of the bluff. One of these has a black plastic drainage pipe attached to it and is visible on the surface of the 1995 slide in Figure 2B, Figure 2C, and Figure 2D. Water was noted to be flowing at a very low rate from the black plastic pipe during a field visit to the slide the day of April 24, 1996. The landowner (William Eaton, personal communication to Stephen Dickson, April 1996) stated that his residence is currently on town sewer, but that with a dye test he determined that one of the ceramic pipes is from a basement drain in his home. The origin of the other drain is unknown. Drainage easements between the Eaton and Gerrish homes, and on the east side of the Smalley home are recorded on the City of Rockland sewer map for Samoset Road (Rockland Wastewater Treatment Facility). These septic and drainage systems provide potential pathways for increasing the amount of groundwater to the February 1995 and April 1996 landslide areas, and therefore may have been contributing factors.

Other Factors

Several residents indicated that fill had been used along some areas of Samoset Road and placed in the area of the 1996 landslide in order to eliminate wetlands and otherwise improve the land for building construction. Artificial fill can certainly be a contributing factor in some slides. However, the borehole lithologic logs for the April 1996 site (Appendix A) show about 3 feet (1 m) of fill material only in the driveway area of the Gerrish house. The other borings show natural soils grading down into natural geological materials. In our investigations of the site we reviewed the head scarp and sidewalls of the slide and found only natural geological materials, with the exception of a 3 - 4 foot (~1 m) wide section of fill immediately adjacent to the foundation of the Gerrish house. Similar fill was noted around what remained of the Smalley foundation, and neighbors stated that a small amount of fill had been added to her property for gardening purposes. In the toe sections of the slide investigated, no fill material was found. Based on this information, we conclude that artificial fill had no or only limited effect in causing the landslide.

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

Last updated on October 6, 2005