**G. L. Sivakumar Babu and M. D. Mukesh**

Assistant Professor, Senior Research Fellow, Department of Civil Engineering

Indina Institute of Science, Bangalore.

[email protected]

**Abstract**

Geographic Information Systems (GIS) have become the promising tool for an effective analysis associated with the study of geologic hazards. GIS is an ideal tool for landslide modeling owing to its versatility in handling a large set of data, providing an efficient environment for analysis and display of results with its powerful set of tools for collecting, storing, retrieving, transforming and displaying spatial data from the real world. This paper demonstrates the ability of the GIS to incorporate the spatially varying data of ground elevation, soil properties, etc. in the engineering analysis of the slope stability. The key factor in landslide hazard mapping is the assessment of risk associated with the failures. Though in general, models are available, better interpretation and understanding of the failures could be derived from them in a GIS environment. A typical landslide area at Mangalore is under investigation for its hazard potential mapping. This paper discusses the analysis of the earthquake and rainfall induced slope failures in GIS and formulates the possible framework.

**Introduction**

Landslides constitute one of the major hazards that cause losses in lives and property. Landslides are one of the complex analyses, involving multitude of factors and need to be studied systematically in order to evaluate the hazard. The increasing computer-based tools are found to be useful in the hazard mapping of landslides. One of such significant tools for hazard mapping of landslides is Geographic Information Systems (GIS). A GIS is defined as a powerful set of tools for collecting, storing, retrieving at will, displaying, and transforming spatial data (Burrough and McDonnel, 1998). One of the main advantages of the use of this technology is the possibility of improving hazard occurrence models, by evaluating their results and adjusting the input variables. An important aspect of landslide investigations is the possibilities to store, treat, and analyze spatiotemporal data that are available.

**GIS and Landslide analysis**

The unique capability of GIS to work (to capture, store and manage the data) with data referenced by vast spatial or geographic coordinates and its ability to incorporate appropriate engineering models, have caused its proliferating application across the wide sections of engineering, especially in civil engineering where, management of spatial-data is pivotal for the analysis. As the typical landslide analysis demands, collection of numerous data, storage of them and using them in the analysis could be handled well in the GIS environment. Any spatially-distributed data with a geo-reference to real world could be stored as points, lines and polygons (vector model) or as continuous fields (raster data model). Beyond GIS being used as a spatial database, it assists in modelling applications through handling a special form of data that would otherwise be compromised in conventional analysis (Miles et al. 1999). Also, GIS does not only serve as a database for parameter data, rather qualitative and quantitative data can be integrated through spatial relationships rather than through relationships between attributes that may not exist (Frost et al. 1997). The other facilities such as: Query languages and user interfaces permitting rapid modification of parameter values; Convenient and quick updating of model parameters; Overlay function, where multiple maps are either visually or topologically combined and the its potential in visualization of data using the graphic features, that assists the engineer in verifying data and information pertaining to the model and its application; and developing elevation maps and subsequent slope, aspect and hillshade themes (which are useful in the landslide analysis) are worth mentioning, as they are common to many applications of engineering models (Miles et al. 1999). In particular, the ability of GIS to present the data and analysis results in map forms plays a key role in identifying the critical areas (where more rigorous analysis and improved solution is required) by its interactive visualization in a spatially optimized mode.

**Fig. 1:** *Landslide area at Mangalore, Karnataka*

**Hazard mapping of landslides**

Landslide hazard maps have been constructed by different methods, such as, from inventories; by consideration of site conditions including geology, hydrology, topography, and/or geomorphology; by statistical correlation of landslide frequency with geologic and geomorphic factors; using safety factors from stability analysis. Mechanics based models have also been used to estimate failure probability based on uncertainties about infiltration of rainfall, or pore pressure, and soil strength (Wu et al. 2000). An a priori study of causal factors of landslides indicate, the ground conditions, geomorphological processes, physical processes and man-made processes as the significant contributors. The pertaining data needed for landslide analysis for a particular slope are:

- Geologic and Geomorphologic features/setting
- Types and quantification of soil and/or rock properties
- Details of vegetation cover, folds, faults, etc.
- Past records of rainfall and earthquake incidences
- Appropriate hydrologic and stability models

**Formulation of typical landslide analysis**

As part of the research project, a landslide at Mangalore, Karnataka is in investigation. A slope stability analysis combined with risk analyses is proposed to carry out. The area under investigation is shown in Figure 1, marked with 100 point elevations and contours of 2m interval. It could observed that the hill is sloping downward towards southeast. The slide has occurred at the upper regions and hence short-term remediation in the form of relief wells and reinforcing piles were employed at appropriate locations. Various man-made features are also present, like roads and drainage structures, a stream falling in the downslope. However, the theoretical framework for the hazard mapping of the slide area presented has been done in detail. The steps involved in the landslide hazard mapping are given in Figure 2 as flow chart. The detailed formulation of the work is described below. Further, an illustration with a hypothetical slope for simulated soil and earthquake conditions is demonstrated at the end. The methodology proposed would be able to develop the hazard maps, incorporating the triggering causal factors, viz., both rainfall and earthquakes.

**Site conditions**

The ground surface elevations could be taken from the topographic maps or digital elevation models (DEM). The other input parameters for the predictive model, involves the detailed characterization of the soil conditions, viz., the description of soil strength properties (cohesion, angle of internal friction, unit weight), permeability characteristics, depth of soil cover, vegetation pattern. The soil cover depth is important in case of shallow soil sliding, which occurs prevalently. The presence of vegetation covers influences the stability calculations significantly as the forested area has increased cohesion value of the soil than the clear-cut areas. Pertinent literature shows that forested areas has upto 20% increased cohesion value (Wu et al. 2000). The values of strength properties would eventually vary spatially throughout the slope. If considerable effort is paid in detailed site investigation, in determining the strength values at more number of points, then a separate statistical analysis (Geostatistics or random field analysis) could be performed, to evaluate the correlation distance. With the information of the correlation distance of each property varying in both vertical and horizontal directions, the interpolation of strength values at unsampled points could be done in Arc View through the Kriging operation. The Arc View facilitates the user to write an avenue coding through which this operation could be performed.

**Fig. 2:** *Formulation of landslide hazard mapping*

The geostatistical interpolation is more accurate than the available IDW method or spline methods of interpolation, especially for the geological properties. **Infiltration and drainage model**

The rainfall-induced failures are the common occurrences in soil slopes. Depending upon the intensity and duration of rainfall and permeability characteristics of the slope, infiltration causes development of pore pressures. Rise in piezometric level in hillside slopes, decreases the effective stress, eventually leading to failure of soil mass. Hence prediction of piezometric levels constitutes an important element in the evaluation of landslide hazards. Suitable hydrologic model predicting the level of pore pressures induced for a particular intensity of rainfall and for the given soil, hydraulic and slope conditions should be coupled with the stability model. The lumped parameter model given by Wu et al (1996), gives the recharge of the groundwater as a function of rainfall characteristics and soil properties. Then a finite-difference solution should be used to calculate the pressure head at different points in a slope.

**Stability model**

Slope failures triggered by any causal factors, involves movement of mass or slide. The exact configuration of the movement and its mass is difficult to predict. Though, depending upon the material characteristics, slope conditions, suitable theoretical models could be used for analysis. Limit equilibrium solutions are available for stability analysis, depending on the assumption of the failure plane, whether it is plane failure of finite/ infinite slope or circular failures. However, in actual slope failures, the assumption of single slope failure is not realistic. It may include a slide of mass with circular failure surface and also with plane surface failure, as multiple slope failures are possible. Hence, the stability model should capture both the types of failures according to the conditions of its occurrences. Appropriate criterion could be established for identifying which type of failure would occur for the given conditions.

As Figure 3(a) shows a slope of height H, inclined at angle y, with a back slope a, composed of rock/soil with cohesion c, friction angle j, and unit weight g (Christian et al. 1997).

For a failure surface (Figure 3(b)) inclined at an angle q, the static factor of safety F is

Ignoring the effect of the vertical acceleration on the calculated static factor of safety, the above equation becomes F*, while the ground acceleration is ah and the amplification factor in the slope is A,

The infinite slope model (Figure) is also used for plane slopes, where g, gs, gw are the unit weights of unsaturated and saturated soils and water, respectively; hs = hp/cos2b is the water level in the slope; b is the slope angle; cr and c’ are the cohesion due to roots and effective cohesion; j’ is the effective angle of internal friction; and z is the depth.

The stability analysis of a plane slope does not account for the possibility of failure where the local slope angles are smaller than the general slope angle. Hence, from the method of slices could be used to calculate the safety factor Fc, for failure along the circular arc.

**Hazard mapping**

The analysis using above hydrologic and stability models could be readily incorporated in the Arc View either using map calculator or avenue scripts. The results give the factors of safety in the form of maps or contours. The task of engineer becomes easier to identify, visually which areas are critical that deserve further attention. Hazard rating could be assigned based on suitable criterion. Natural hazard is defined as the probability of occurrence of potentially damaging phenomena within a specified period of time and within a given area. Zonation also refers to the division of the land in homogenous areas or domains according to the degree of actual or potential hazard (Varnes, 1984). Following gives such a rating criterion, based on a quantitative estimation of landslide occurrence over a given region, without mentioning the time period.

**Table 1. ***Quantitative Hazard Rating of Landslides (Miles et al.,1999)*

Stability Criterion | Hazard rating |

F > 1.5 | Stable slope |

1.25 > F < 1.5 | Low Hazard |

1.00 > F < 1.25 | Medium Hazard |

0 > F < 1.00 | High Hazard |

F is the factor of safety

**An Illustrative example**

The preliminary analysis of the slope stability is done in Arc View GIS, with hypothetical slope and simulated earthquake conditions, to validate and explore the capabilities of GIS. The area of analysis is 80X80m, with the crest of 85m and mean slope as 53°. The slope, soil properties (cohesion, friction angle, and unitweight) are given as input in attribute table and developed as themes through event theme menu. Then interpolation for the elevation and soil properties for whole slope area is done through spline interpolation. The following table gives the statistics of the given inputs.

**Table 2:** *Input Parameteres for the hazard predition model*

Statistics | c (kPa) | f degree) | g(KN/m) | Slope (°) |

Mean | 38.4 | 46.2 | 18.12 | 53.4 |

Standard Deviation | 1.9 | 1.467 | 0.86 | 16.57 |

Coefficient ofVariation (%) | 5 | 3 | 4.7 | 31 |

Minimum value | 28 | 41 | 16 | – |

Maximum value | 42 | 49 | 21 | – |

In stability calculations, it is assumed the failure is planar surface and hence equations 2 and 3 are used for computing safety factors. As this stability model is capable of including the earthquake effects, a pseudo-dynamic analysis is also performed. The results of this analysis are given in table 3. The factors of safety maps/contours with hazard rating are shown in figures. Assuming that the probability distribution of factors of safety follows normal distribution, the reliability index and the corresponding probabilities of failure are computed, for both static and pseudo-static results.

**Table 3.** *Results of static and pseudo-static analyses*

Static FS | Dynamic Factors of Safety (Aah) | |||||

0.1 | 0.2 | 0.3 | 0.4 | 0.5 | ||

Mean of FS | 2.6 | 1.95 | 1.52 | 1.21 | 0.98 | 0.8 |

s of FS | 0.23 | 0.19 | 0.19 | 0.19 | 0.19 | 0.21 |

C.O.V (%) | 8.8 | 37 | 12.5 | 15.8 | 20.3 | 26 |

b | 6.89 | 5 | 2.73 | 1.11 | -0.01 | -0.97 |

P[f] | 0 | 0 | 0.02 | 0.11 | 1.5 | 1.54 |

The safety factor maps gives the visualization of the weaker and stable zones, spatially. Accordingly, the distribution of hazard areas could be spotted for further in-depth analysis or for suitable remediation program. Relevant maps and contours are given in Figures 4 a,b,c and d.

*Fig. 4(a): Elevation map with contours of the slope Fig. 4(b): Slope map of the analysis area*

**Concluding remarks**

Geographic Information Systems (GIS) is being exploited widely in many engineering problems, which involves spatial data management. Landslide hazard mapping, one of the important task in disaster/hazard mitigation projects is a typical problem involving huge database. A framework for the landslide hazard mapping in GIS is discussed with an illustrative example. The posterior analysis of GIS results gives the engineer a better understanding and visualization of the problem and results.

*Fig.4(c): Static factors of safety of the slope Fig. (d): Dynamic factors of safety for the slop with acceleration coefficient of earthquake, 0.4. *

**Acknowledgements**

This work is carried out as part of the CSIR sponsored research project, “Risk Assessment in landslides”, at Indian Institute of Science, Bangalore. Their financial assistance is gratefully acknowledged.

**References**

- Burrough, P.A and McDonnell, R. 1998. Principles of geographical information systems, Oxford University Press, London.
- Christian, J.T. and Urzua, A. 1998. Probabilistic evaluation of earthquake-induced slope failure, Journal of Geotechnical and Geoenvironmental Engineering, Vol. 124, No.11, 1140-1143.
- Frost, J.D., Caroll, D.P., and Rochaway, T.D. 1997. Spatial liquefaction analysis, Spatial analysis in soil dynamics and earthquake engineering, Geotech. Spec. Publ. No. 67, ASCE, Reston, Va., 70-86.
- Lee, W.A., Lee, T.S., Sharma, S., and Boyce, G.M. 1996. Slope stability and stabilization methods, John Wiley & Sons, 363 – 389.
- Miles, S.B. and Ho. C.L.1999. Applications and issues of GIS as tool for civil engineering modeling, Journal of computing in civil engineering, ASCE, Vol.13, No.3, 144-161.
- Varnes, D.J. 1984. Landslides hazard zonation: A review of principles and practice, Natural Hazards No.3, Commission on Landslides of the Int. Assoc. of Engrg. Geology, UNESCO, Paris, 63.
- Wu, T.H, and Abdel-Latif, M.A. 2000. Prediction and mapping of landslide hazard, Canadian Geotechnical Journal, 37: 781-795.
- Wu, W., Abdel-Latiff,M.A., and Wieczorek, G.F, 1996. Landslide hazard prediction. In Proc. Of the 7th Int. Symp. On landslides, Trondheim, Vol 1, pp. 423-428.