Landslide initiation in saprolite and colluvium in southern Brazil: 3 Field and laboratory observations

1

2

3

4

56

7

8

9101112131415161718192021222324

2526

27

28

29

30

31

32

2006) xxx–xxx

+ MODEL

GEOMOR-02105; No of Pages 16

www.elsevier.com/locate/geomorph

ARTICLE IN PRESS

Geomorphology xx (

OF

Landslide initiation in saprolite and colluvium in southern Brazil:Field and laboratory observations

Willy A. Lacerda ⁎

Civil Engineering Program, COPPE – Instituto Alberto Luiz Coimbra de Pós-Graduação e Pesquisa,UFRJ – Federal University of Rio de Janeiro, Brazil

Received 25 September 2004; received in revised form 4 December 2005; accepted 23 March 2006
O
RECT

EDPRAbstract

The weathering of granitic and gneissic rocks in tropical regions can reach depths of more than 100 m. In southeast Brazil thereare situations where landslide initiation depends on the fluctuation of the groundwater level, on the impact of falling rocks and onintense rainfall, causing superficial slides. The fluctuation of groundwater induces cyclical variations of the pore water pressure,and consequently of the effective stresses. This variation causes cyclic expansion and contraction of the structure of the saproliticsoil, weakening the imbrication of grains and loss of the cementation that may exist. This could be called a “fatigue” phenomenon.The practical effect is the lowering of the Mohr shear strength envelope, and a sudden rupture of the soil at a lower groundwaterlevel than that which would be compatible with the intact soil strength properties, initiating a landslide.

Another situation arises during intense rains, when a rock slab or a rock block detaches from the uppermost parts of a slope.This occurs where steep rock outcrops exhibit relief joints or where residual blocks of granite roll down the slope, impacting thecompressible, saturated colluvial soil overlying the saprolitic soil. The sudden increase of pore pressure can liquefy the soil. Finally,another mechanism is that of the advance of a saturation front in a steep slope of unsaturated saprolitic soil, reaching a depth belowthe root zone. The loss of the cohesion due to suction, without the beneficial contribution of the roots to the shearing strength,causes a sudden slide. During extreme rain episodes literally hundreds of such superficial slides, reaching 1 to 3 m in depth, occurin a given basin. The concentrated runoff that flows along the surface of the thalweg of the basin carries this soil in a muddy state,and a debris flow ensues.© 2006 Elsevier B.V. All rights reserved.
RKeywords: Landslides; Debris flow; Tropical soils; Saprolite; Colluvium O
33

34

35

36

37

38
UNC1. Introduction
In Fig. 1 the southeastern coast of Brazil is shown. Ashaded portion lies between a massive mountain range(The “Serra do Mar”) and the Atlantic Ocean, betweenparallels 20 °S and 30 °S. The basal rock is part of the

39

40

41

⁎ Tel.: +55 21 25627200; fax: +55 21 2562 8464.E-mail address: [email protected].

0169-555X/$ - see front matter © 2006 Elsevier B.V. All rights reserved.doi:10.1016/j.geomorph.2006.03.037

Please cite this article as: Lacerda, W.A. Landslide initiation in saprolite aGeomorphology (2006), doi:10.1016/j.geomorph.2006.03.037.

coastal shield, and is composed mainly of gneiss andgranite, with intrusions of basic rocks (diabase dykes arecommon). From the highest elevation of the mountains(reaching little more than 2000 m in isolated peaks, withan average of 800 m) down to the coastal plains, sapro-litic and colluvial soils cover the rock mantle, until thealluvial plains are reached. It is along these slopes thatlandslide phenomena that affect engineering works areactive.

nd colluvium in southern Brazil: Field and laboratory observations.

mailto:[email protected]

http://dx.doi.org/10.1016/j.geomorph.2006.03.037


42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

2 W.A. Lacerda / Geomorphology xx (2006) xxx–xxx

ARTICLE IN PRESS

2. Soil nomenclature

Blight (1997) defines “residual soil” as the weatheredmaterial of the “in situ” rock, and which did not suffertransport since its formation. In Brazil it is common, ingeotechnical literature, to call saprolitic soil “residualsoil”, meaning that the igneous rock was decomposed “insitu” (Barata, 1969). A further distinction is made for thesuperficial part of the saprolitic soil, in general cementedby iron and aluminum oxides. It is usually called inBrazil “mature residual soil”, and the underlying sapro-litic soil, with no cementation, is referred to as “youngresidual soil”. Colluvial soil, being a superficial deposit,is also subject to this oxide forming process, and may bemistaken as “mature residual soil”. In this paper the termsaprolitic soil will be used meaning both young andmature residual soil. The term colluvium, in the presentpaper, refers to both practically intact soil material fromslides with little transport to completely remoldedmaterial of large and sudden landslides.

3. Some characteristics of saprolitic and colluvialsoil in a tropical environment

Ab'Saber (2003) describes the main characteristics ofsaprolitic and colluvial soils in Southeastern Brazil:

• Deep weathering of crystalline rocks (up to 100 m ormore);

UNCO

RREC

Fig. 1. The “Serra do Ma


TEDPR

OOF

• generalized occurrence of red yellow podzolic soils,or latossols, along the hill slopes, developed on col-luvial and eluvial deposits and on the top of roundedhills, and also on the saprolitic soil derived fromthese crystalline rocks;

• superposition of colluvial deposits, due to the cli-matic fluctuations of the Quaternary, often separatedby stone lines, burying paleo-thalwegs;

• presence of “sugar loaves” in areas where the tectonicjoint spacing is abnormally large, of the order ofhundreds of metres, and where the relief jointingsystem is more frequent;

• rivers and springs fed perennially by the phreaticwater;

• perched water tables along the humic layer of theforests.

These facts are also reported by Fookes et al. (2000),who, commenting on their geomorphological model forhot wet climate, state that the depth of weathering can bemany tens of metres, controlled by the distribution ofdiscontinuities in the rock mass. Also, duricrusts are for-med where groundwater concentrates soluble weatheringproducts (Thomas, 1994). Deeply weathered colluvialsoils are often difficult to distinguish from in-situ weathe-ring profiles. Laterites are hard subsurface deposit ofoxides of aluminum and iron found in tropical soils wherethe water table fluctuates with seasonal changes in preci-pitation. This process is called “laterization”, and when

r” region in Brazil.



OOF

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

Fig. 2. Formation or colluvium under a rock outcrop (Fookes, 1997).

3W.A. Lacerda / Geomorphology xx (2006) xxx–xxx

ARTICLE IN PRESS

UNCO

RREC

the laterite is not present, the soil is called “lateritic soil”by some authors, and this includes both colluvial andsaprolitic soil. These definitions were discussed in theProgress Report of the Committee on tropical soils of theISSMFE (Committee on tropical soils of the ISSMFE,1985). Fig. 2 shows the formation of colluvium in atropical environment.

4. Stress–strain behavior of saprolitic and colluvialsoils

A feature that shows the distinct behavior of sapro-litic and colluvial soils is their stress–strain curves intriaxial tests, or stress–displacement curves in directshear tests, as shown in Fig. 3 (Lacerda and Silveira,1992). It can be seen that the saprolitic gneissic soil (a)exhibits peaks up to 600 kPa, whereas the shear stress ofcolluvial soil (b) is strain hardening.

Both colluvial and saprolitic soils near the surfacecan form cementation due to iron and aluminum oxideformation. In this case, a true cohesion exists, and thestress–strain behavior of both soils is similar. In Brazilthe geotechnical engineers use “mature residual soil”, or“lateritic soil” for this condition. Sometimes lateriticcolluvial soil exhibits peak and dilatancy at low normalstresses, under 100 kPa.

Sandroni and Maccarini (1981) discussed the in-fluence of feldspar, quartz and mica on the strengthenvelope of drained direct shear of saprolitic soils. Foramounts of mica above 30% there is a tendency forlower shear strength envelopes, with effective frictionangles below 20°, while for low mica content and highfeldspar/quartz content friction angles are above 25°,typically 30°.


TEDPR5. Mechanisms of instabilization

Several authors (Barata, 1969; Costa Nunes, 1969;Wolle, 1985 and Lacerda, 1997, 1999, 2004) have dealtwith the causes of instabilization in tropical regions.

The main mechanisms of instabilization of saproliticand lateritic soils and rock slopes operative in the Serrado Mar range can be summarized as follows:

• Rupture by cyclic pore pressure in saprolitic soils• Loss of suction• Debris flows• Rupture induced by fall of rock slabs onto saturatedcolluvial or residual soils

• Rock slides along relief joints• Slide reactivation due to artesian aquifers• “Injection” of water from deeply seated permeablerock veins

• Obstacles to ground water flow

The main mechanisms of reactivation of colluviumslopes with a near surface ground water level are:

• Cuts made at the toe of the slope• Loads (usually road embankments) imposed fromtheir middle section up

• Elevation of the ground water table

6. Rupture of saprolitic soil by cyclic pore pressurevariation

The stress state of an element of soil in a slope hastwo components: the total stress state and the pore waterpressure (positive or negative) acting on it. The positive



TEDPR

OOF

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

198

Fig. 3. Normalized shear stress vs. displacement of residual (left column) and colluvial soils (right column) (Lacerda and Silveira, 1992).


ARTICLE IN PRESS

UNCO

RRECpore pressure variation is due to the variation in the

ground water level. In this case, the total stress remainsconstant, for practical purposes. Failure can then occur ifthe pore pressure is such that the effective Mohr strengthenvelope is reached. This mechanism of failure of satu-rated saprolitic soil with one cycle of pore pressureincrease was called to attention by Brand (1981).

Lacerda (1989) noted the fact that many saproliticslopes may fail during a rainy period of lesser intensitythan previous ones, which were withstood withoutfailure. The natural conclusion is that the soil elementhad been subjected to a higher pore pressure in thepast than that which caused failure in the present. Sincepore water pressures fluctuate seasonally, it could wellbe that the soil undergoes a weakening of the inter-granular bonds due to this cyclic behavior, which isgenerally below the critical level of pore pressures thatlead to failure. The basic assumption is that the cyclingof pore pressure below the one cycle failure porepressure value breaks the interlocking of the soilparticles, and it may fail with a lower failure envelope.


This phenomenon could be interpreted as a fatiguebehavior.

That is to say that the pore pressure cycling up to amaximum pore pressure that is less than that necessaryto fail the specimen in a triaxial stress path with mono-tonic increase of pore pressure. This cycling inducesaccumulated shear strains that help the breaking of thebonding or interlocking responsible for “true” (saturat-ed) effective cohesion, thus lowering the Mohr strengthenvelope, until the soil fails with a smaller pore pres-sure. The hypothesis that ϕ′ is constant during thisprocess is not unreasonable, and is adopted.

So, the soil may fail under an effective stress statebelow the failure envelope determined by shear strengthtests of intact soil specimens. This proposal was inde-pendently put forward by Eigenbrod et al. (1987) forcompacted soils.

Fig. 4 shows the typical situation in gneissic–graniticrocks, where the range of variation of piezometric levelin the lower aquifer of the saprolitic soil is shown. Fig. 5shows the failure by a monotonic increase of pore



DPR

OOF

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215

216

217

218

219

220

Fig. 4. Typical variation of piezometric levels in residual soils from gneissic–granitic rocks (Lacerda, 1989).


ARTICLE IN PRESS

REC

pressure, with the total stresses maintained constant;Fig. 6 defines pore pressures uo, umax and uf. Porepressure uo is that corresponding to an equilibriumfriction angle, such as the residual effective frictionangle. Pore pressure uf is that necessary to fail the soilby a monotonic increase in pore pressure, and umax isthe maximum value of the applied cyclic pore pres-sure in the cyclic triaxial test. The ratio Uc= (umax−uo) /(uf−uo) gives an idea of the degree of intensity of porepressure cyclic loading. Uc=1 means rupture in just onecycle.

UNCO

R

Fig. 5. Rupture by monotonic increase in pore water pressure (Brand,1981).


TEIt is difficult to obtain field data of pore pressuresprior to a failure, since the exact location of the ruptureis often unpredictable, but a simulation of this mech-anism can be made in the laboratory. Drained triaxialtests with cyclic pore pressure (CPP tests) were per-formed (Fig. 7). Tests were carried out on undisturbedsamples of a saprolitic soil from the Soberbo Road Slidein Rio de Janeiro. The tests were performed using thefollowing procedure: (i) saturation by back pressure;(ii) consolidation under isotropic condition; (iii) increaseof deviator stress with controlled stress (anisotropic

Fig. 6. Definition of pore pressures uo, umax and uf.



OF

221

222

223

224

225

226

227

228

229

230

231

232

233

234

235

236

237

238

239

240

241

242

243

244

245

246

247

248

249

Fig. 7. Cyclic pore pressure triaxial tests.

Fig. 9. Uc vs. number of cycles to failure.


ARTICLE IN PRESS

consolidation) until a desired value of shear stress; (iv)application of cyclic pore water pressure. During theCPP tests, axial and volumetric deformations weremeasured. Santos Jr. et al. (1997) describe the testingprocedures in detail.

Lacerda and Santos Jr. (2000) obtained yield surfacesfor this granitic saprolitic soil in a saturated conditionfrom drained isotropic compression tests, which con-firmed tests by Lacerda and Silveira (1992) on the samesoil, and from tests with failure induced by increasingthe pore pressure while maintaining the shear stressconstant (Fig. 8).

The same yield surface was defined by both tests. Thecurve of the yield or limit states (Fig. 8) has the ap-proximate shape of an ellipse in a p–q diagram, as also

UNCO

RREC 250

251

252

253

254

255

256

257

258

259

260

261

262

263

264

265

266

267

268

269

270Fig. 8. Yield surface of granitic residual soil.


TEDPR

Oshown by Sandroni and Maccarini (1981) and Leroueil(1997).

Fig. 9 shows these results (Santos et al., 1997). For acyclic pore pressure ratio Uc of 67% failure is reachedafter 400 cycles (forUc=1 the soil fails in the first cycle).This is a possible mechanism that helps to explain someslope failures under low accumulated rainfall.

It is important to note that this “fatigue” behaviorapplies only to stress states below the yield line. Whenthe soil is in a stress state above the yield region thecementation bonds have already been broken, and thesoil behaves essentially as a granular material. At shal-low depths the soil stress state is within the yield region.These results indicate that this hypothesis seems to betrue for a saprolitic granitic soil involved in the largehistoric landslide at Soberbo Road.

7. Loss of suction

Loss of suction is a major instabilization process, andthe resulting unstable mass has been the source materialfor some of the severest debris flows, when it occurs inclosed basins.

Vegetation increases the shear strength of the super-ficial soil, which will show a cohesion intercept due tothe root reinforcement. In slopes where the phreaticlevel is low, these slopes are essentially stable. Never-theless, infiltration of rain water brings this superficialsoil to a near saturated state.

If the saturation front reaches a depth situated belowthe root zone the loss of apparent cohesion may causethe slope to fail (Fig. 10). Wolle and Hachich (1989)have shown this phenomenon as the cause of thousandsof landslides near the summit of the coastal range nearCubatão, Sao Paulo, during periods of extremely heavyrainfall. The accumulated debris from the landslides inthe flatland near the sea was the cause of the mud



ROOF

271

272

273

274

275

276

277

278

279

280

281

282

283

284

285

286

287

288

289

290

291

292

293

294

295

296

297

298

299

300

301

302

303

304

305

306

307

308

309

310

311

Fig. 10. Shallow slide due to loss of suction.

Fig. 11. The Santa Genoveva Nursing Home accident: cross section(Barros et al., 1989).


ARTICLE IN PRESS

UNCO

RREC

covering the city of Caraguatatuba, near Santos in 1967(Jones, 1973).

8. Debris flows

When the rain intensity is higher than 70 mm/h, and ithappens after a period of extended rain (typically200 mm of accumulated rain in the last 7 days) there isthe danger of tens or thousands of shallow landslidesinitiating almost at the same time. Jones (1973), as wellas Barata (1969) and Costa Nunes (1969), describe theextreme events of 1966 and 1967 in this region. Locally,rainstorms can affect just a small region, and isolatedlandslides initiate debris flows. Sometimes the trigger ofa debris flow is a shock wave, produced by impactloading of rock or of a landslide that suddenly covers asaturated colluvium (Sassa, 1985; Barros et al., 1989).

9. Rupture induced by fall of rock slabs onto saturatedcolluvial soils

A debris flow of moderate proportions caused by theimpact loading of fragments of a large granite–gneissicrock slab is described by Barros et al. (1989), andLacerda (2004). They explain the triggering factor of thedebris flow as follows. A slab of appreciable dimensions(thickness from 2 to 4 m, length of more than 10 m)detached from a gneissic scarp of the Corcovado massifdue probably to cleft water pressures. The event tookplace in 1988, after 18 days of almost continuous rain(total of 840 mm), which were more than sufficient to


TEDP

completely saturate the 5 m thick colluvium, composedof more than 30% of rock blocks, some of which morethan 4 m in diameter, at a 20° inclination. Sometimesgeologists in Brazil use the term “talus/colluvium” todistinguish this particular type of colluvium. Thisamount of rain was divided in three distinct periods bySantos Barros and Brandão (1992); the accumulatedrainfall in the last 3 days was 441 mm, with a peakhourly rainfall of 56 mm/hr in February 22, the day ofthe debris flow. The slab fell from elevation 525 m andwas broken on its first impact on the rock scarp justbelow, at elevation 400 m; the resulting flying blocks(with an energy that cut some high trees in half)impacted the colluvium at elevation 280 m. Figs. 11



312

313

314

315

316

317

318

319

320

321

322

323

324

325

326

327

328

329

330

331

332

333

334

335

336

337

338

339

340

341

342

343

344

345

346

347

348

349

Fig. 12. The Santa Genoveva Nursing Home accident: plan view(Barros et al., 1989).


ARTICLE IN PRESS

Cand 12 show the section and the plan view of the site.The colluvium liquefied, and ran until a break in thedeclivity of the slope, in which there was a nursinghome, at elevation 180 m. Another debris slide occurredin an almost parallel paleo thalweg, also saturated, andwhich usually carried superficial water flows duringstorms. The two debris flows joined exactly at thenursing home, indicated by the black arrow, as can beseen in Fig. 12. About 53 people were killed in thisaccident alone. Figs. 13 and 14 show respectively the

UNCO

RRE

Fig. 13. Aspect of the nursing home 1 day aft


TEDPR

OOF

final position of the debris of the landslide, on a break ofthe slope, where the buildings were situated.

10. Rock slides along relief joints

The upper part of rock outcrops is sometimes steeplyinclined, and the relief joints are subjected to waterpressures due to rain infiltration. This can lead to thedetachment of individual blocks or slabs. When theseblocks fall on the scarp below, they may either initiate adebris flow, as discussed in Section 7, or, if the region isforested, and the soil below is not saturated, can accu-mulate on top of saprolitic or colluvial masses (Fig. 15),and become part of it. In tropical regions it is rare toobserve true talus deposits, which are more common indry, or temperate regions with little or no saprolitic soilformation.

11. Slide reactivation due to artesian aquifers

Saprolitic soils in the saturated state are often sepa-rated from the upper colluvial cover by a clayey soil,which can be an oldmature residual soil that was coveredby the colluvium, or a layer with a concentration ofelluviated, fine material from the more impermeablecolluvium. In this case, the aquifer of the saprolitic soilcan be confined, and semi-artesian or even artesian porepressures can develop in the lower stratum if the rate ofinfiltration from higher elevations is high. The exampleof a slide due to this cause is that of Urubu's Hill in thecity of Rio de Janeiro, which occurred on April 8, 1966.Several previous slides had been reported by the

er the landslide (Photo by Ary Maciel).



ROOF

350

351

352

353

354

355

356

357

Fig. 14. Aspect of the region washed through by the landslide. Gneissic rock appears where a 5 m thick colluvium/talus existed. At right and far left,note the intact colluvium (Photo by the Author in 1989).


ARTICLE IN PRESS

residents of the densely occupied area. These slideshappened 30 years before, and it is logical to assume thateven when the area was uninhabited they may have alsooccurred, but had been unnoticed. The slide involved the

UNCO

RREC

358

359

360

361

362

363

364

Fig. 15. Large fallen rock block, with more than 10 m in maximumdimension, in a forested slope (Tijuca forest, Rio de Janeiro). Photo bythe Author.


TEDPsubsidence of the upper part of the slope, forming a scarp

with 4.5 m, cutting a house in half. At the foot of theslope the ground suffered a heave of 1 m, and the soil andthe walls of the houses at this region were cracked. Allresidents had time to evacuate their houses, because theslide was not sudden. This accident is reported byMoreira (1974). Fig. 16 shows the plan of the site. Rotarydrill borings with Denison sampling, piezometers and 4inclinometers were installed. Sounding S-54, at the sideof inclinometer I-4, showed that the subsoil consisted ofa colluvium/talus up to 17 m thick resting on a clayey

Fig. 16. Urubu's Hill plan (adapted from Moreira, 1974).



OOF

365

366

367

368

369

370

371

372

373

374

375

376

377

378

379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

400

401

402

403

404

405

406

407

408

409

410

411

412

Fig. 17. Cross section A–A′ showing failure surface and piezometric lines (adapted from Moreira, 1974).

t1:1

t1:2

t1:3

t1:4

t1:5

t1:6t1:7

t1:8


ARTICLE IN PRESS

CORR

EClaterized residual soil 4 m thick. Below this mature soil,“young” residual soil, or saprolitic soil was encountereduntil a depth of 42m. From 42 to 47m the rock, a biotite–gneiss was very fractured, with schistosity inclined20 degrees to the horizontal; from 47 to 50 m, wheresound gneiss was finally reached. In the colluvium,weathered boulders of gneiss were encountered, onewith 4.2 m in diameter, another with 2.7 m, as shown incross section A–A′, Fig. 17. This profile is typical ofmany colluvium deposits in southeastern Brazil.

From the Denison “undisturbed” samples, classifica-tion, drained direct shear and isotropically consolidatedundrained triaxial tests were performed. These resultsare shown in Tables 1 and 2. It can be noticed that thesaprolitic soil effective friction angle is similar to thepeak friction angle, despite the 40% amount of clay.Lacerda and Fonseca (2003) have shown that this occursin lateritic soils, which behave as granular materials.The base rock is migmatite gneiss, and the existence of aquarry at the other side of the hill permitted the visualinspection of the sound rock.

Stability analyses carried out with the piezometriclines indicated in section A–A′ were performed. For aSafety Factor of 1.0 the contact between the colluviumand the clayey residual soil or the contact between the

UN

413Table 1Average from several classification tests in each layer (Moreira, 1974)

Depth (m) Soil Atterberglimits

Grain sizeanalysis

WL WP PI Clay Silt Sand

8–16 Colluvium matrix 42 22 21 33 17 501–15 Decomposed boulders – – NP 13 11 7217–18.4 Residual, clayey 58 34 24 40 25 3522–38 Saprolitic 48 32 16 14 11 75


TEDPRclayey residual soil and the saprolitic soil had to be in a

residual shear strength condition with c′ equal to zeroand ϕ′res equal to 30°. The inclinometric data showed amovement occurring along the thickness of the clayeysoil, with no movement in the saprolitic soil. The collu-vium moved as a block, accompanying the movement atthe contact with the intermediate, clayey layer.

Fig. 18 shows the 25-day accumulated rainfall for thesummer of 1966 (January to March). It can be seen thatthe rupture occurred in a period in which the accu-mulated rainfall was of the order of 350 mm, while amaximum of almost 600 mm was registered in January.It is postulated that during this time lag (the slide oc-curred in April 8) the pore pressure of the artesianaquifer reached a critical level. A similar situation wasalso reported by Jiao and Malone (2000).

12. “Injection” ofwater fromdeeply seated permeablerock veins

Water recharge by means of concealed springs con-nected to water bearing fractures in the underlying rockcan alter significantly the flow pattern in its neighbor-hood, and a suitably located piezometer would showartesianism. Local artesianism can initiate landslides inupper colluvial layers, as the Author has seen in Brazil.

Table 2 t2:1

Shear strength tests (Moreira, 1974) t2:2

t2:3Soil Test c′ (kPa) ϕ′ c′res (kPa) ϕ′res

t2:4Colluvium DST*, drained 24 29 – –t2:5Clayey residual soil CIU** 29 30 – –t2:6Saprolitic soil CIU** 19 32 5 30

*DST – direct shear test; submerged; **CIU – isotropicallyconsolidated triaxial test, saturated. t2:7



willy

Nota

I have sent separately a new version of this figure

414

415

416

417

418

419

420

421

422

423

424

425

426

427

428

429

430

431

432

433

434

435

436

437

438

439

Fig. 18. 25-day accumulated rainfall in the summer of 1966.


ARTICLE IN PRESS

In order to simulate this situation Borges and Lacerda(1986) made Finite Element analyses of a slope with aninitially low water table, and then applied a source witha piezometric pressure just above ground level, asFig. 19a) and b) from Lacerda (1999), shows. The arrow

UNCO

RREC

Fig. 19. Influence of a hidden spring on the f


PROO

F

indicates the direction of flow. The water level is signi-ficantly altered to a position close to the slope surface. Ifa cut were made in this slope in the dry season, it wouldeventually fail during the wet season, and, even withouta cut, the slope would certainly show signs of instability.The observation of some slides in natural slopes justafter a very heavy rainy period shows springs of waternear the crown of the slide.

13. Obstacles to ground water flow

The remnants of the of Soberbo Road debris flow inRio de Janeiro, 1967, have been extensively studied(Soares et al., 1988; Lacerda and Schilling, 1992). Thematerial deposited after the flow constitutes a colluvi-um, which moves continuously (in reality, intermittent-ly, according to the elevation of the ground water table).Fig. 20 shows a partial plan of the slide. Section C–C1along zones C, B and D is shown in Fig. 21. The twoupper branches of this complex slide present a colluviallayer on top of the residual soil. The thickness of thecolluvium lies between 6 and 10 m. The water table isclose to the surface, except at two zones, when it appears

TED

low pattern of a slope (Lacerda, 1999).



Fundação

Inserted Text

the city of

Fundação

Note

Accepted set by Fundação

Fundação

Inserted Text

in

Fundação

Note


willy

Nota

I have sent separately a new version of this figure

NCOR

RECT

EDPR

OOF

440

441

442

443

444

445

446

447

448

449

450

451

Fig. 20. Map of the Soberbo Road slide (Lacerda and Schilling, 1992).


ARTICLE IN PRESS

Uactually at the surface. At these points impermeablediabase dikes were found while perforating long (80 m)horizontal drains. The dykes are shown in Fig. 21,which also show the position of piezometers, inclinome-ters and superficial marks. The accumulated movementof the superficial marks can be seen to increase as the


diabase dykes are approached, as the arrows in Fig. 20show. Two piezometers and a water level indicator wereinstalled at most of the soundings, and they are shown aspositions A, B and C. The arrows in this figure indicatethe direction of movement of the flow lines. They areseen to bend upwards near the diabase dikes. The



DPR

OOF

452

453

454

455

456

457

458

459

460

461

Fig. 21. Section along the Soberbo Road Landslide showing two families of diabase dykes (Lacerda, 1999).


ARTICLE IN PRESS

C

position of the more impermeable dikes influences theflow lines. Thus, artesian pressures can be observed justbefore the dykes, as shown. In the Soberbo Road caseartesianism was indeed observed, the water level of thedeepest piezometer just at the upper contact with the

UNCO

RRE

Fig. 22. Displacements measured by inclinometers in the region


TEdyke rising more than 1 m above the ground elevation.Of course, the ground was very wet, with rivulets ofwater springing at the surface.

The local stability of the colluvium is decreased justabove the obstacle to flow which the diabase dykes

between the diabase dykes (Lacerda and Avelar, in press).



TEDPR

OOF

462

463

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

486

487

488

489

490

491

492

493

494

495

496

497

Fig. 24. Sequence of elevation of the water level inside the flume(Lacerda and Avelar, in press).


ARTICLE IN PRESS

REC

represent. Indeed, superficial horizontal movementswere larger at this location, as already discussed, and asuccession of cracks and the inclination and displace-ment of small trees and inclinometers showed the signsof this instability (Fig. 22).

In order to clarify the behavior of colluvial massesunder seepage obstructed by barriers, such as the dia-base dikes, a flume was constructed (Avelar, 2003), andtests performed with different inclinations of the box,filled with clean sand, under different inclinations. Thesand used in the tests was a washed, fine beach sandconsisting predominantly of quartz grains. The grainsize was between 0.149 and 0.420 mm (passing # 40sieve, retained in the # 100 sieve), and its specificgravity was 2647. The friction angles varied from 32 to35.5 degrees under a confining pressure of 25 kPa, insaturated–drained triaxial tests, with void ratios between0.75 and 0.85 (near the loose state). The permeabilityvaried between 2 and 3×10−2 cm/s for this range ofvoid ratios. The dry density was in the range between14.2 and 15.4 kN/m3.

The dry sand was loosely placed with the help ofspades, with the box resting horizontally. After thisloose placing the wand was wetted with a water hoseunder low pressure, and its surface was straightened outwith a steel ruler. The end result was a plane surface,with a depth of sand of 15 cm. The barrier consisted of apiece of wood with a height of 10 cm, firmly glued to thebottom and sides of the box.

Fig. 23 shows the Flume set up (Lacerda and Avelar,in press). Water reservoirs at the top and bottom of thebox are maintained with a constant water level, andobservations are made during the steady-state seepagepattern established. Electric and standpipe piezometersmonitored the advance of saturation. The displacementof the sand mass was measured at surface and with

UNCO

R

Fig. 23. Flume set up (Lacerda and Avelar, in press).Fig. 25. Slide of the soil mass above the barrier (dyke) (Lacerda andAvelar, in press).

Please cite this article as: Lacerda, W.A. Landslide initiation in saprolite and colluvium in southern Brazil: Field and laboratory observations.Geomorphology (2006), doi:10.1016/j.geomorph.2006.03.037.


498

499

500

501

502

503

504

505

506

507

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600


ARTICLE IN PRESS

UNCO

RREC

colored vertical dark-stained sand “piles”, simulatinginclinometers.

The procedure was to raise the box to the desired angleand then begin the process of water percolation, through ahose connected to the upper reservoir, maintaining aconstant water level. The phreatic line rose slowly, and thefirst readings of the piezometers were in those nearest thebarrier. Fig. 24 illustrates this sequence, together with theobserved displacement of the sand mass. Tests were runwith the inclination of the box varying from 10° to 30°, in5° intervals. For tests with box inclinations between 20°and 30° there was a clear formation of the shear zone justbefore the barrier. Fig. 24 shows the shear zone, seen indetail in the photograph of Fig. 25. The barrier forces theshear zone to curve upward, corroborating field evidenceof near circular slides at this location. The shear zoneexpands as the movement progresses.

14. Conclusions and remarks

1) Slides in shallow residual soils are generally sudden,due to their brittle stress–strain response, at low stresslevels. Deep seated slides in saturated saprolitic soilsmay not be sudden, because the stress strain behaviorat confining stresses larger than about 1000 kPa isgenerally strain hardening.

2) Saturated colluvium deposits are common in tropicalareas. Their brittle behavior is limited to confiningstresses lower than 50 kPa. They present a strainhardening stress–strain behavior, whichmeans that theyslide progressively, subject to pore pressure variation.

3) Many processes can destabilize colluvium:

• Elevation of ground water level in perched aquifers;• Cuts made at the toe of long colluvium slopes;• Embankments or loads applied to their middle or topsections;

• Impact of loads (rock falls or slides);• Effect of impermeable dykes or formations at somepoint along the slope.

4) Some slides may occur after the cessation of rainfall,due to the delay in the recharge of deep artesian aquifers.

15. Uncited reference

Leroueil et al., 1996

Acknowledgements

The PRONEX program (from the National ResearchCouncil - CNPq) has provided funds for the flume tests.


TEDPR

OOF

The author acknowledges Mr. Luiz de França, for thehelp in Figs. 1 and 5, 6, 7, 8, 9.

References

Ab'Saber, A., 2003. The Domains of Nature in Brazil – Landscapes.Atelie, S.Paulo, Brasil. (in Portuguese).

Avelar, A.S., 2003. Initiation and development of debris flows inslopes: triaxial impact tests and flume tests simulating the SoberboRoad landslide, Rio de Janeiro, PhD thesis, COPPE-UFRJ, Rio deJaneiro. (in Portuguese).

Barata, F.E., 1969. Landslides in the tropical region of Rio de Janeiro.Proc. 7th Int. Conf. on Soil Mechanics and Foundation Engineer-ing, vol. 2. ISSMFE, Mexico, pp. 507–516.

Barros, W.T., Amaral, C.P., Sobreira, F.G., D’Orsi, R.N., Maia, H.S.,Cunha, R.P., 1989. Catastrophic avalanche at the St. Genovevaslope. Solos e Rochas, ABMS (Brazilian Society for SoilMechanics and Geotechnical Engineering). São Paulo 11 (1),17–25 (in Portuguese).

Blight, G.E., 1997. Mechanics of Residual Soils. Balkema, Rotterdam.Borges, M.S.N., Lacerda, W.A., 1986. On the internal drainage of

cut and fill slopes. Proc. 8th Brazilian Conf. on Soil Mechanicsand Foundation Engineering, vol. 1. ABMS, Porto Alegre, Brazil,pp. 17–33 (in Portuguese).

Brand, E.W., 1981. Some thoughts on rain-induced slope failures.Proc. 10th Int. Conf. on Soil Mech. and Foundation Engineering,vol. 3. ISSMFE, Stockholm, Sweden, pp. 373–376.

Committee on tropical soils of the ISSMFE, 1985. Peculiarities of Geo-technical Behavior of Tropical Lateritic and Saprolitic Soil. ABMS,São Paulo.

Costa Nunes, A.J., 1969. Landslides in soil of decomposed rock due tointense rainstorms. Proc. 7th Int. Conf. on Soil Mechanics andFoundation Engineering, vol. 2. ISSMFE, Mexico, pp. 547–554.

Eigenbrod, K.D., Graham, J., Burak, J.P., 1987. Influence of cyclingpore-water pressures and principal stress ratios on drained defor-mations in clay. Canadian Geotechnical Journal 29, 326–333.

Fookes, P.G., 1997. Tropical residual soils. A Geological SocietyEngineering Group Working Party Revised Report. The Geolog-ical Society, London.

Fookes, P.G., Baynes, F.J.,Hutchinson, J.N., 2000. Total geological history:a model approach to the anticipation, observation and understanding ofsite conditions. Proc.Geo-Eng 2000, Intern. Conf. onGeotechnical andGeological Engineering, Melbourne, Australia. 91 pp.

Jiao, J.J., Malone, A.W., 2000. An hypothesis concerning a confinedgroundwater zone in slopes of weathered igneous rocks. Symp. onSlope Hazards and their Prevention. PRC, HongKong, pp. 165–170.

Jones, F.O., 1973. Landslides of Rio de Janeiro and the Serra das ArarasEscarpment, Brazil. Geological Survey Professional Paper, vol. 697.Department of the Interior, US Printing Office, Washington.

Lacerda, W.A., 1989. Fatigue of residual soils due to cyclic porepressure variation. Proc. 12th Int. Conf. on Soil Mech. and Found.Eng. ISSMFE, Rio de Janeiro, Brazil, pp. 3085–3087.

Lacerda,W.A., 1997. Stability of natural slopes along the tropical coast ofBrazil. Proc. Symp. on Recent Developments in Soil and PavementMechanics. COPPE-UFRJ, Rio de Janeiro, Brazil, pp. 17–40.

Lacerda, W.A., 1999. Local instability in saturated colluvial slopes inSouthern Brazil. Proc. Intern. Symp. on Slope Stability Engineer-ing - IS-SHIKOKU 99, vol. 1. Balkema, Rotterdam, pp. 199–204.

Lacerda, W.A., 2004. The behavior of colluvial slopes in a tropicalenvironment. Proc. Intern. Symp.on Landslides: evaluation andstabilization, Rio de Janeiro, Brazil, vol. 2, pp. 1315–1342.



Fundação

Note

Eliminate: "15. Uncited reference Leroueil et al, 1196"

Fundação

Note


601602603604605606607608609610611612613614615616617618619620621622623624625626627

628629630631632633634635636637638639640641642643644645646647648649650651652653

654

655


ARTICLE IN PRESS

Lacerda, W.A., Avelar, A.S., in press. Flume tests on sand subjected toseepage with the influence of hidden barriers. Proc. Intern.Workshop IW-FLOWS2003, Sorrento, Italy, 7 pp.

Lacerda, W.A., Fonseca, A.P., 2003. Residual strength of colluvial andresidual soils. Proc. 3rd Pan-American Conf. of Soil Mechanicsand Geotechnical Engineering, Boston.

Lacerda,W.A., Santos Jr., O.F., 2000.Water table fluctuation. A cause ofslide initiation in residual and colluvial soils in tropical regions. Proc.8th Intern. Symp. on Landslides, Cardiff, Wales, vol 2, pp. 877–880.

Lacerda, W.A., Schilling, G.H., 1992. Rain induced creep-rupture ofSoberbo Road Landslide. Proc. 6th Intern. Symp. on Landslides,Christchurch, New Zealand, vol. 1, pp. 145–152.

Lacerda, W.A., Silveira, G.C., 1992. Shear strength and compress-ibility characteristics of the Soberbo hill residual and colluvialsoils. Proc. 1st Brazilian Conf. on Slope Stability, Rio de Janeiro,Brazil, vol. 1, pp. 445–461 (in Portuguese).

Leroueil, S., 1997. Critical state soil mechanics and the behavior of realsoils. Proc. Int. Symp. onRecent Developments in Soil andPavementMechanics, Rio de Janeiro, Brazil. Balkema, Rotterdam, pp. 41–80.

Leroueil, S., Vaunat, J., Picarelli, L., Locat, J., Lee, H., Faure, R., 1996.Geotechnical characterization of slope movements. Proc. 7th Intern.Symp. on Landslides, Trondheim, Norway, vol. 1, pp. 53–74.

Moreira, J.E., 1974. Stability of slopes in residual soils of granite andgneiss. M.S. dissertation. COPPE-Universidade Federal do Rio deJaneiro. (in Portuguese).

Sandroni, S.S., Maccarini, M., 1981. Triaxial and direct shear tests on ayoung feldspathic gneissic residual soil. Proc. Brazilian Symp. on

UNCO

RREC


PROO

F

Engineering of Tropical Soils. COPPE-UFRJ, Rio de Janeiro,Brazil, pp. 334–339 (in Portuguese).

Santos Barros, L.H., Brandão, A.M. de P.M., 1992. Pluviosity – aconditioning factor of mass movements in the Serra Carioca, Rio deJaneiro. Proc. First Brazilian Conf. on Slope Stability, I COBRAE,ABMS, Rio de Janeiro, Brazil, vol. 2, pp. 697–708 (in Portuguese).

Santos Jr., O.F., Lacerda, W.A., Ehrlich, M., 1997. Triaxial tests withcyclic pore pressure variation for simulation of the seasonal variationof water level in slopes. Proc. Symp. onRecent Developments in Soiland Pavement Mechanics, Ed. M. Almeida, COPPE-UFRJ, Rio deJaneiro. Balkema, Rotterdam, pp. 279–282.

Sassa, K., 1985. The mechanism of debris flow. Proc 11th Intern.Conf. on Soil Mechanics and Foundation Engineering, SanFrancisco, U.S.A., vol. 3, pp. 1173–1176.

Soares, M.M., Pedrosa, M.G.A., Lacerda, W.A., 1988. Mechanism ofmovements in colluvial slopes in Rio de Janeiro. Proc. 5th Intern.Symp. on Landslides, Lausanne, Switzerland, vol. 2, pp. 1211–1216.

Thomas, W.F., 1994. Geomorphology in the Tropics, A Study ofWeathering and Denudation in Low Latitudes. Wiley, Chichester,England.

Wolle, C.M., 1985. Slope Stability in Tropical Lateritic and SaproliticSoils, ISSMFE Committee on Tropical Soils. Progress Report1982–1985. ABMS, São Paulo, pp. 164–221.

Wolle, C.M., Hachich, W., 1989. Rain-induced landslides in Southeast-ern Brazil. Proc. 12th Int. Conf. on Soil Mechanics and FoundationEngineering, Rio de Janeiro, Brazil, vol. 3, pp. 1639–1642.

TED



Fundação

Note

Eliminate this reference - "Leroueil et al, 1996..."

Fundação

Cross-Out

Fundação

Cross-Out

willy

Caixa de texto

1. LACERDA, W.A. e AVELAR, A.S. (2003) Flume tests on sand subjected to seepage with the influence of hidden barriers; IW-FLOWS 2003-International Workshop on the Ocurrence and Mechanisms of flow-like landslides in natural slopes and earthfills, Sorrento, Ed. Luciano Picarelli, Pàtron Editore, Bologna, pp.155-162

willy

Nota

1.Lacerda, W.A. & Avelar, A.S. (2003) Flume tests on sand subjected to seepage with the influence of hidden barriers; IW-FLOWS 2003-International Workshop on the Ocurrence and Mechanisms of flow-like landslides in natural slopes and earthfills, Sorrento, Ed. Luciano Picarelli, Pàtron Editore, Bologna, pp.155-162

Landslide initiation in saprolite and colluvium in southern Brazil: 3 Field and laboratory observations

Documents