TERRAIN STUDIES FOR SLOPE INSTABILITY IN PARTS OF TEHRI DISTRICT, GARHWAL HIMALAYA, U.P., INDIA

Abstract

Himalaya is the youngest mountain range in the world. They owe their origin to the collision of Indian Plate with the Eurasian plate. With the continual subduction of Indian Plate, the Himalaya is still rising, thus, form a very unstable belt on the northern part of India. On the basis of lithotectonic characters, this mountain range has been classified in to six longitudinal zones - as sub-Himalaya in the south, followed northerly by lesser Himalaya, Higher (or greater or central) Himalaya, Tibetan (or Tethys) Himalaya, Indus-Tsangpo suture zone and Trans-Himalaya (Schawn, 1981). During rainy seasons Himalaya witnesses inumerable landslides which cause major hazards. In general, it has been observed that the Lesser Himalayan region is more affected by landslide activity as compared to sub-Himalayan region (Joshi, 1987). In the Garhwal region of the Lesser Himalaya, a study conducted by Central Building Research Institute, Roorkee indicates annual land degradation of 120 sq.m per kilometer of roads. Although these areas are highly landslide prone, it appears that land failure phenomena are not observed every where. They are found to occur and recur in the areas characterised mainly by unique combination of topographic, structural, lithologic, hydrologic, climatic and vegetative features. These features to gether with mass wasting play important role in changing the geomorphic regime of an area, both on long and short term basis. Hence, the present study is an endeavour towards identification, (v) evaluation and forecasting such hazardous zones of instability due to landslide mainly through geomorphological signatures in the Garhwal region of the Lesser Himalaya of a part of Tehri district, U.P., India. Geologically, the area is made up of two major lithostratigraphic succession - the Bijni Member of the Lansdwone Formation and the Younger succession of Blaini, Krol, Tal and the Subathu Formations, consisting mainly of quartzite, limestone/ dolomite, siltstone, slate and shale. The Bijni quartzite of Lansdowne Formation has angular unconformable contact with the overlying younger stratigraphic units. Also, the contacts between Phulchatti Quartzite and Manikot Shell Limestone of Upper Tal and the Subathu Formation with underlying rocks is marked by an unconformity. In the south-western part of the area, Bijni Member overlies the Subathu (Eocene) Formation and this abnormal position is due to a major fault known as Singtali Fault or Garhwal Thrust, which trends, in general, WNW-ESE with steep southerly dip varying between 70° to 80°. These rocks in general, have four sets of joints dipping NE, SSW, ESE and NW direction. The terrain under study is characterised by rugged topography and typical ridge and valley features. The slopes are found to be highly variable and in general steep. Approximately 90% of the hillface is occupied by natural vegetation cover. Landslide affected areas in the terrain are found to occur at any elevation which ranges from about 500 m to more than 2200 m. In (vi) general, with increase in elevation, the landslide affected area though tends to increase, yet is not uniformly distributed. Such areas are prominently found to occur mostly at elevations above 1800 m, 1500 - 1600 m and 1100 - 1300 m, close to the upper part of the ridges, and also at elevation of 500 - 600 m which is generally found along the hillslope near the road running almost parallel to the river Ganga. Slopewise also, the percent landslide affected area is not uniformly distributed in different slope facets. It is maximum in facets with slope angle between 30 to 35°. Landslide distribution has also been found to have intimate relation with vegetation cover. It is observed that maximum landslide affected areas fall within barren to sparsely vegetated land, of which barren lands have been affected the most, and found associated with steeper slope areas of upper part of ridges. The effect of lithology on the landslide area distribution is not conspicuous as most of the area (about 80%) is occupied by a single rock type i.e. quartzite. However, a careful observation indicates that the argillites and carbonates of the Lower Krol Formation, the quartzite of the Bijni Member of Lansdowne Formation and Blaini Formation have been affected the most. The steep sided valleys have led to increase in the susceptibility of the rocks to landslide activities. The slope movement is also governed by occurrence, orientation and trend of local structures, such as bedding plane and joints. Various workers like Horton (1932, 1945), Strahler (1952, 1971) Schuum (1956), Morissawa (1962,1963,1968,1985) have studied (vii) landforms and proposed a number of morphometric parameters which are used for terrain evaluation. Despite their wide use for morphometric studies, little attempt has been made to assess them as instability indicators. The basins of the third order which formed good working units, were the targets of study on morphometry vis-a-vis instability of the area. The correlation matrix prepared from twenty variables indicates that Fraction landslide area (Ls) in a basin has statistically significant correlation coefficient of about 0.86, 0.84, 0.68 and -0.55 (99% confidence level) with Drainage texture (DT), Stream frequency(SF), Drainage density (DD), and Basin circularity (BC) respectively. Drainage density is sensitive to changes in lithology, vegetation cover, slope, structure, and hydrologic aspect of stream system (Carlston, 1963; Strahler, 1964). Stream frequency is largely dependent on rock type. Therefore drainage texture is one single morphometric .parameter in a basin that has in it, the influence of many morphometric parameters which in turn, are reflection of the sum effect of elevation, slope, lithology, structural features, vegetation and hydrological condition. Based on regression analysis the following relationship between Fraction landslide area and Drainage texture (DT) has been worked out. Ls = 0.0279 + 0.0052 DT This relationship when tested to estimate Fraction landslide area in four randomly selected drainage basins in the area was found

to be useful within the error limit of 25 percent. Despite the usefulness of Drainage texture as an indicator of instability of a terrain, this parameter cannot be used to study instability condition at local level of a particular hillslope section. Also, it does not indicate the nature of failure and processes of instability involved on the hillslope. Nevertheless, it indicates that morphological features can be very useful in investigation of stability of hillslopes in a watershed. Morphology of slopes can dictate type of movement and extent of instability (Crozier, 1973). He showed that significant morphometric indices can be obtained from a landslip as indicators of slope movement processes. Crozier investigated slope morphology on the basis of sixty six landslips as observed in a terrain made up largely of concavo-convex slopes with very few bedrock outcrops and free faces associated with a thick soil and regolith mantle cover in Newzealand. In the present investigation, a similar approach has been made on the basis of sixty transverse slope profiles in the twelve randomly selected drainage basins. However, critical values of various indices namely classification index, dilation index, flowage index, displacement index and tenuity index given by Crozier (1973) could not be applied as such in this area. Hence, these critical values as predictive limits for likely process of slope movement were modified. The modified mean values of five morphometric indices namely classification, dilations, flowage, displacement and tenuity indices were calculated to be 7.68, 0.35, 75.92,

48.04 and 1.20 respectively for rotational slide (RS). Likewise the mean values for five above mentioned indices for planar slide were calculated to be 7.22, 0.36, 44.0, 73.31 and 0.62 respectively. Similarly, for slide flow (SF) the mean values of classification, dilation, flowage, displacement and tenuity indices were calculated to be 4.39, 0.70, 73.94, 54.25 and 1.62 respectively. The calculated mean values for fluid flow (FF) were found to be 1.88, 1.78, 215.95, 53.22 and 2.37 for classification, dilation, flowage displacement and tenuity indices respectively. The efficacy of the modified parameters as predictive limits for the slope movement processes involved on a hillslope was established when this approach was applied successfully to three different areas of similar topography located at 122 km NNW (Mussoorie ByPass); 80 km NNE (Kaliasaur) and 57 km SSW (Chilla) of the area of study. Displacement index (DPI) appears to indicate the potential of failure of a section. A section is likely to become unstable and fail by rotational slide, planar slide, slide flow and fluid flow if the displacement index is less than 48.04, 73.31, 54.25 and 53.22 respectively. With a view to work out a simple quantitative statistically significant criteria to differentiate between the various slope movement processes based on simultaneous use of all the five morphometric indices, multivariate discriminate function analysis was used. The multivariate criteria have been developed and tested successfully to discriminate slide flow from fluid flow, (x) fluid flow from planar slide, fluid flow from rotational slide, planar slide from slide flow, rotational slide from slide flow and planar slide from rotational slide. Discriminant analysis when used as search technique, indicated that of all the morphometric indices, classification, dilation and tenuity indices contribute significantly in discriminating various slope movement processes. Based on these indices, bivariate and unvariate plots have been prepared. These plots can also be used as tools to screen and find out the slope movement process at a given site. The developmental activity has affected the hillslope adjoining the road sections at a number of places by various types of slope movement, particularly planar and wedge failure. In order to evaluate the hillslope failure along the road, twenty one landslides were identified between Byasi to Devaprayag in the study area. Six sites, designated herein as L-^, L5, L6, L7, Lg and L14 were selected for detailed study. Slope orientation, joint orientation, weathering of joint wall, joint continuity, spacing, filling and separation of joints were measured. Stability analyses were performed using Slope Mass Rating (SMR) and limit equilibrium technique (Hoek and Bray, 1981) . To use SMR it was necessary to determine joint wall (JCS) and uniaxial compressive strength of rock material. The L-type Schmidt hammer (rebound hammer) is in use by many research workers world over, for determination of JCS. The correlation coefficient between joint wall uniaxial compressive strength (JCS) measured

in the laboratory on rock samples (on cubes of 25 mm length) and the rebuond hammer test data on natural discontinuity surfaces (using Miller equation, 1965) was found to be significantly low i.e. close to zero. A close scrutiny of samples and data revealed that hammer can give acceptable strength (JCS) of joint wall, if the surface is dry and smooth, and not relatively uneven, which occures in most cases in field. Slope Mass Rating of six selected sites indicates that the rating for sites hlf L5 and L14 are 22, 28 and 42 respectively. These conclusions proved correct on field observations. However, rating for sites L6, L7 and Lg was calculated to be 18, 24 and 6 respectively and do not represent the stability condition as observed in the field. It therefore, indicates that SMR method of stability analysis is not a foolproof approach of indicating stability condition. A critical examination of these sites indicates that when the continuity along the dip of a discontinuity surface is less than five percent of the total height of affected slope, it appears that the SMR technique does not depict the true picture of stability condition and extent of failure. Slope stability analysis based on short solution of Hoek and Bray (1981) suggests that all these slopes are unstable in wet condition and are stable under dry dynamic (earthquake) condition. The conclusion draws its support from the fact, that year after year, these slopes fail only during the monsoon

season, due to dual action of pore water which exerts pressure to destabilise the slope on one hand and decreases the strength of wet rock mass along discontinuity surfaces on the other hand. It is also important to note that these sites remained unaffected by an earthquake of 6.5 on Richter scale which devastated the area in and around Uttarkashi (situated 35 km to the north of study area). Computer slope analysis recommends provision of adequate drainage in form of drill holes with folded geofabrics to stabilise rock slopes. The study, in short indicates that i) Slope movement in the area is the out-come of the action and interaction of factors like, topographic elevation, slope, structural, lithological, hydrological conditons, vegetation cover. The sum effect of these factors is also reflected in the morphometric signature of a terrain. ii) L-type rebound hammer should be used on smooth natural surfaces only iii) It appears when continuity along dip of a discontinuity surface is less than five percent of total height of affected slope, SMR may not give adequate picture of stability condition, (iv) Slope movement along the road is due to excess pore water pressure build ups and decrease in the strength of rock mass during rainy season. The slopes remain stable under dry dynamic (earthquake) condition.