Black Mallet Landslide
Preliminary Geotechnical Evaluation

Prepared by
Derek Gay Ph.D.
Cassandra Rogers Ph.D.
Department of Civil Engineering
The University of The West Indies
Saint Augustin
Trinidad and Tobago

 

CONTENTS

EXECUTIVE SUMMARY

A preliminary geotechnical evaluation of the Black Mallet landslide, Castries, Saint Lucia, has been carried out. This was based on data gathered through a reconnaissance survey (Gay and Rogers, 1999), a detailed topographic site survey (The Chief Surveyor, Saint Lucia, 1999) and borehole and inclinometer data. (Strata Engineering, 1999) The following inferences were made:

  1. Approximately 2.2 hectares of land at the base of the northeast slope of Parker Hill are actively moving downslope.

  2. Within the landslide zone colluvial soils, 4m to 6m thick, lie over weathered andesite bedrock. In the lower third of the slide zone, a sand lens, 2.5m to 3.5 m thick, occurs between the colluvium and andesite bedrock.

  3. Inclinometer data indicate that the failure plane lies at the andesite/soil (sand and colluvium) interface.

  4. Two vectors of movement have been identified. The dominant direction of movement of the upper slide zone is approximately toward the east whereas the lower slide mass is moving approximately toward N65O E. The non-alignment of these movement vectors gives rise to a rotatory motion between the two masses in plan.

  5. Direct measurements of shear strength for both the colluvium and sand layers were not available at the time of this investigation. Thus only preliminary stability analyses, based on conservative strength estimates, were conducted. These analyses also suggest that only 0.8 m of an additional mean rise in piezometric head is required to initiate further movement. (or limiting equilibrium). This could be achieved with a rainfall event.

  6. The slope movement is thought to have occurred as a result of the development of high artesian pressures in the weathered andesite bedrock, leading to failure of the overlying colluvium and sand lens. These elevated pressures are likely to have developed as a result of an unusually high rate of water recharge. Residents report that the Black Mallet area is continually wet, even in the dry season and that cracks in structure are commonplace (though not to the level currently experienced).

  7. The topography of the area suggests that the recharge sources that are most likely to influence the slide zone emanate within the Parker Hill and Maynard Hill watersheds.

  8. It was reported that deforestation at the summit of Parker Hill was carred out in early 1999 to facilitate land development. Although such activity can lead to increases in the rates and volumes of recharge into the permeable weathered andesite, the role of this phenomenon in triggering the landslide could not be conclusively established at this time. This demands further investigation.

  9. It is not clear whether the seismic event of June 16th, 1999 (Magnitude 5.7, Martinique) could have contributed to the sliding, as no records of accelerations or Intensities were available for the island. This activity is also worthy of further investigation, as seismic activity has been known to trigger increased flows and pressure in aquifers.

  10. Construction activity at the Marchand River (toe of slope) does not appear to be directly related the observed mass movements.

  11. Based on the above, the following recommendations can be made:

There remains a high risk of continued large-scale movement following heavy rainfall. It is therefore recommended that the affected area be evacuated. The main concern at this time is to relieve the water pressures in the slide zone. The use of a well point system is expensive and not always reliable. Under the circumstances, it is suggested that a large sump be excavated, its water pressure monitored and pumped periodically.

More detailed data acquisition and analyses are required in order to advance the reliability of the preliminary results obtained to date. The following is recommended:

INTRODUCTION

This report describes the findings of a reconnaissance survey and preliminary analyses of the Black Mallet landslide in lower Castries Saint Lucia as indicated in Figure 1. Drs. Derek Gay and Cassandra Rogers of the Department of Civil Engineering, The University if the West Indies, carried out the survey over the period November 25th, 26th 1999 at the request of the Ministry of Works Saint Lucia.

In this report, the findings of the reconnaissance survey and data analysis to date are presented and interpreted in conjunction with the findings of the available borehole investigations and topographic surveys.

BACKGROUND

On October 10th 1999 the Ministry of Works Saint Lucia was officially apprised of the severity of landslide activity in the Black Mallet area of Castries, Saint Lucia. In response engineers from this Ministry, in collaboration with officials from the Planning Division, immediately carried out reconnaissance surveys of the area and conclude that damage to buildings and infrastructure was sufficiently severe to warrant the evacuation of residents and the demolition of some buildings. This initial slide was contained in the lower area of the slide zone and included an area of approximately 1.6 Ha, as illustrated in Figure 2.

Over the weeks that followed another slide scrap developed south ("uphill) of the first encompassing an additional 0.6Ha. This scrap crossed Maynard Hill Road at approximately 70 m from the top of the original scrap and aligned eastward along a paved stepped walkway that ascends toward Parker Hill, this illustrated in Figures 2 and 6.

As far back as June/July 1999 cracking was reported in a building at the corner of Black Mallet Road and Mauricette Gap. What could have been interpreted as local settlement of the corner column of this reinforced concrete and blockwork structure eventually developed into the scrap of a full-scale landslide, currently referred to as the Black Mallet Landslide. However, it was only around September 17th 1999 that reports of significant damage to buildings were received and structural damage consistent with mass movement on slopes became evident.

Oral history suggests that damage to buildings and infrastructure in the Black Mallet area has been occurring for many years (Henry, 1999) although not to the degree of that currently experienced. This was corroborated during the reconnaissance survey by the authors, as "aged" settlement related cracks were in evidence in many older buildings, which appeared.

To be unrelated to the slide in question. In addition, it was reported that there are many areas between Maynard Hill Road and the Marchand River (approximate borders of the landslide area) in which surface soils were perennially wet.

The field data, accounts of damage and records of related chronological events referred to in this report were gathered through interviews and the offices of the Ministries of Works and Communications and Planning and Development. The following persons were instrumental in this exercise

At the time of the visit, a topographical survey was also carried out within the area to identify slide scraps and the extent of ground movements. Geotechnical investgations comprising the drilling of four boreholes and the installation and monitoring of inclinometers and piezometers had previously been conducted by Strata Engineering Limited.

SITE DESCRIPTION

Location

The Black Mallet area of Saint Lucia is located to the southwest of Castries, on the lower northeastern slopes of Parker and Maynard Hill, as indicated in Figure 1. Topography

The topography of the site varies from moderately sloping (90 – 120) to steeply sloping (13o – 17o) as can be appreciated from the contour plan in Figure 2 and cross sections of Figures 4 and 5.

Existing Infrastructure

The area can be considered as part of an unplanned residential suburb adjacent to the downtown Castries city/business centre. Building can be typically described as belonging to

The category of "low cost" infrastructure, comprising typically wooden and/or reinforced concrete and blockwork construction. Many of these residence are not served by a central sewer system and employ traditional "outhouse" cess-pit and septic tank /soak-away systems.

Geology and Soils

The geology map of Saint Lucia suggests that volcanic andesite formations underline the site. The andesite bedrock is moderately to deeply weathered. In the vicinity of the uppermost slide scrap toward Parker Hill exposures of this andesite can be observed. On the lower slopes the bedrock is covered by slit to clay rich colluvial soils. The geology of the area is illustrated in Figure 3.

RECONNAISSANCE INVESTIGATIONS

Field Survey:

A field reconnaissance survey was carried on November 25th and 26th 1999. This survey comprised a walk/drive about survey of the affected areas including detailed inspections of the slide scraps and building damage.

At the time of the survey roadway areas, which had ruptured as result of the land slippage, had been repaired through back-filling and asphaltic concrete repaving. Houses that were deemed a threat to public safety had been demolished and more were earmarked for evacuation and demolition. Work was also in progress on the filling and diversion of the lower reaches of the Marchand River.

During the week prior to our site visit it was reported that the upper slide scrap (secondary scrap) had developed. This scrap aligned approximately east-west, crossing Maynard Hill Road approximately 70m south (uphill) of the primary lower slide. It was possible to follow this feature along much of its length as it aligned parallel and adjacent to the paved stepped pathway leading toward Parker Hill. The rupture zone along this path was principally a direct shear zone, transitioning through a compressive shear into a tensile zone at the extremity of the failure scrap. This mechanism suggesting movement downhill in the direction of steepest descent as illustrated in Figures 2 and 6. None of the boreholes carried out addressed this secondary scrap as the geotechnical investigation had been completed prior to the onset of its development.

Continuing our reconnaissance beyond the steps along a northward path along the 50m contour, lead to an outcrop/exposure of andesite bedrock. This bedrock was thought to be the surface expression of the weathered andesite founded at depth in the boreholes within the primary slide area some 40m below.

The observed housing density in the area was far greater than that suggested by the topographic survey as this survey typically included building which had suffered significant damage and those whose occupants had to be relocated and or compensated, see Figure 6.

It was clear from the survey that all buildings within the landslide area suffered some degree of damage as a result of mass soil movement. However, as expected, those located close to, or over the rupture zones suffered significantly higher levels of damage. In many cases rupture zones cut directly through buildings as illustrated in the edited photographic journal in Appendix A.

Evidence of recently repaired water mains were also observed in open trenches and at the ground level. It was reported that ground translation within the primary slide varied from 0.5 m to in excess of 1 m, causing the rupture of water mains and the dislocation of the bases of power lines.

Reported:

The exact sequence of movement is not clear. However it is reported that initial movement occurred between June and July, 1999 at a house at corner of the Black Mallet Road and Mauricette gap, in the eastern section of the slide zone (see Figures 2, 6 and Appendix A). This was indicated by the development of a 5m scarp and translation of an outer column of the structure. Movement was next observed at a building on the western section of the slide zone and along Maynard Hill Road. This was followed by large-scale movement of sections of the slope in late September, 1999.

GEOTECHNICAL INVESTIGATIONS

Field investigations were carried out by Strata Engineering Limited, and comprised 4 boreholes. Standpipe Casagrande type piezometers were installed in each of the boreholes, whereas additional holes were drilled at three of the boreholes locations to accommodate down-hole inclinometers. These boreholes logs and inclinometer results are given in Appendix B.

To date the laboratory test results are unavailable as these investigations are ongoing.

GEOTECHNICAL ANALYSIS

Soil Profiles and Properties

Boreholes log data in conjunction with topographic survey information was used to develop a longitudinal section through BHs 1, 2, 3, within the primary slide area, as illustrated in Figure 4.

This data suggests that a layer of colluvium of thickness varying between 4 – 6 m overlie the area. Between BH1 and BH2 a layer of sand 2.5 – 3.5 m thick was also encountered, however, this sand was not continuous to BHs 3 or 4. Weathered andesite completed the profile to the drilled depth of all boreholes. This is consistent with the geology of the area (as described in an earlier section) which indicates that a weathered andesitic formation underlise the Black Mallet, Marchand, Maynard Hill and Parker Hill regions.

Depth of shear failure

Inclinometers at BH locations 1 and 4 (viz. S199BM1and S199BM3, respectivley) indicate that shear failure and sliding is occurring at the andesite/soil interface. At BH1 8m of colluvium and sand slide over the weathered andesite. This information is of significance as it suggests that the weathered andesite (and not the sand) is more likely to be responsible for hydraulic continuity to artesian water sources.

Soil Strength Parameters

Shear Strength:

No direct measurements of shear strength have been reported to date. Hence, the best available estimates of strength would be obtained through empirical correlation with the Standard Penetration Test (SPT blows/ft.). SPT N values in the sand layer encountered in BHs 1 and 2 average, N = 10 blows/ft. and N = 44 blows/ft. respectively giving f = 30o and f > 40o respectively.

Pore-water Pressures:

Pore-water pressures measured 24 hrs after completion of the boreholes have been recorded in the boreholes logs and are plotted in Figure 4. These measurements suggest that a piezometric surface exist above the level of the sand layer but within the elevation of the

ground surface. It is not clear however, whether or not this piezometric surface is also a free water table within the colluvium deposits.

Analysis of Stability Lower (Primary) Slide

  1. If it assumed that conditions at the time of the geotechnical investigations (at BH 1 &2) are representative of limiting equilibrium, FS =1.0, then the mobilised un-drained shear strength at the andesite/sand is back calculated to be, Su = 31.02 kN/m2. This translates to drained strength parameters of f ’ = 18.5o, c’=0, for the andesite/sand interface. These results were obtained using the soil stratigraphy as illustrated in Figure 4 and by making the following additional assumptions:

  1. If it is assumed that the sand exists at its lowest angle of friction f ’ = 30o (based on SPT results at BH2). Then it can be shown that for failure to occur at the andesite/sand interface a piezometric pressure of 8.9m head of water would have had to exist at the base of the sand layer. The measured piezometric surface currently stands at approximately 5.0 m above the base of the sand layer and the ground surface at approximately 8.0 m above the base of the sand layer. This assumption inherently also suggest that if an angle of friction f ’ = 30o is indeed representative of the sand, then this slope currently stands at a Factor of Safety = 1.73 (assuming current conditions = conditions at time of the geotechnical investigations).

  2. If it is assumed that the conditions at BH3 (area where sand layer was not encountered) are representative of limiting equilibrium, FS = 1.0, then the mobilised un-drained shear strength at the andesite/colluvium interface is back calculated to be, Su = 14.95 kN/m2. This translates to drained strength parameters of f ’ = 13.3o, c’ = , for the andesite/colluvium interface.

Analysis of Stability of Upper (Secondary) Slide (toward Parker Hill)

  1. The angle of steepest descent in the secondary slide occurs along a line from Maynard Hill Road to the andesite rock out crop toward Parker Hill (parallel to the paved steps), a slope of approx. 1:3, b =18.4o. These topographic features can be appreciated from the NE-SW section taken through Parker Hill as illustrated in Figure 6. Under dry conditions at limiting equilibrium, FS = 1.0, the mobilised un-drained shear strength at the andesite/colluvium interface would be given by, Su = W.Sin(b ). For a mean thickness of colluvium of 3m, the mobilised un-drained shear strength at the andesite/colluvium interface would be 16.6 kN/m2.

CONCLUSIONS

  1. The Black Mallet landslide has occurred over an area whose mean slope is » 12.33o (the primary ‘lower’ slide). In engineering terms, this is a relatively shallow slope and one not normally associated with mass movements of the magnitude observed, unless special mitigating circumstances prevail. In this case, it appears that the special mitigating circumstances are associated with the high artesian pore-pressures introduced at the level of the underlying weathered andesite bedrock and sand lens above it.

  2. Preliminary analyses of stability suggest that better estimates of shear in both the colluvium and sand layers are required to facilitate more reliable assessments of stability. However, using relatively conservative estimates of strength based on empirical correlation with currently available data, the Factor of Safety associated with the lower primary slide is given by F.S. = 1.73. However, it should be noted that this analysis also suggest that a mean rise of only 0.8 m head over the current piezometric level is required to re-establish a condition of limiting equilibrium and sliding within this zone.

  3. A preliminary study of surface deformations and manifestations of sliding suggest that the mean movement vectors of the lower and upper sliding masses are not aligned. They appear to follow their direction of steepest descent. This non-alignment of movement vectors typically gives rise to a rotatory motion (in plan) between the two masses resulting in a far more stable configuration than that of two masses sharing parallel axes of movement. This latter condition being conducive to progressive failure with a high potential sudden and catastrophic energy release.

  4. The observed slope movement appears to have occurred as a result of the development of high artesian pressures in the weathered andesite bedrock, leading to mass movement of the overlying colluvium and sand lens. These elevated pore-pressures are likely to have developed as a result of an unusually high rate of water recharge into slope. The topographic configuration suggests that the primary recharge sources for the slide zone are Parker Hill to the east and Maynard Hill to the south. Residents report that the Black Mallet area is continually wet, even in the dry season. In addition, aged settlement cracks in buildings within the slide zone indicate that movement has occurred with some regularity in the past, though not catastrophically. Further it is reported that the summit of Parker Hill was deforested in early 1999 to create the Parker Hill land development. The collective data suggest that deforestation activity of the summit of Parker Hill may have led to increases in the rate and volume of recharge into the permeable weathered andesite, during the 1999 rainy season, leading to the development of high artesian pressures in the slope.

  5. Possible additional contributors to the observed mechanisms of failure would also include:

These events are sufficiently important so as to warrant further investigation. Of particular importance would be the characterisation of rainfall associated with the hurricane weather systems.

  1. Construction of retaining walls at the slope toe (along the Marchand River) does not appear to be a cause or trigger of the observed movement.

  2. It is imperative that ongoing analysis and monitoring of this site and adjacent sloping areas be undertaken, as failure of such a land mass constitutes a potential hazard not only its internal residents but also to all downstream infrastructure and population of the Castries area.

RECOMMENDATIONS

Immediate and Short Term

There remains a high risk of continued large-scale movement following a heavy rain event, it can therefore be recommended that the affected area is evacuated.

Remedial action

  1. Seal scarp exposures

  2. Implement permanent drainage infrastructure within the slide zone so as to control and contain surface runoff.

  3. The main concern at this time is to relieve the water pressures in the slide zone. The use of pumped wells to achieve this is not recommended, as it is expensive and not always reliable. Under the circumstances, it is suggested that a large sump be excavated in order to pump out the water (similar to deep construction de-watering).

  4. Additional boreholes and/or test pits are required to establish the extent of the sand lens, as described in the text.

  5. Obtain the results of shear strength testing of the colluvium and sand to enable more reliable stability analyses to be performed.

Medium to Long Term

  1. Acquire historical rainfall records for the region consistent with the estimation of event probabilities and associated intensities, which can then be associated with a likely level of mass movement (risk).

2. Conduct appropriate geotechnical and hydrogeolgical studies of the Maynard Hill/Parker Hill areas (the site of the current activity) and extend these to adjacent populated sites which surround the Castries area, so as to establish the risk and liklihood similar movements. This assessment would include an evaluation of the geology, soils, surface and sub-surface water conditions, which could influence the development of artesian conditions in the underlying weathered andesite.
Return to the top
You are visitor since February 23, 2000
http://lucianstuff.i.am
1