Professional Documents
Culture Documents
March 2014
DYNAMIC MODEL TO FORECAST SLOPE STABILITY A CASE STUDY: GERADIELLA LANSLIDE SITE
Research Project Thesis submitted in partial fulfillment of the requirements for the degree of Bachelor of Science in Engineering
March 2014
Signature:
Date:
I had supervised and accepted this thesis for the submission of the degree.
... Date
ii
DEDICATION
iii
ACKNOWLEDGEMENT
It is a great pleasure to acknowledge to the many people who have supported us in the course of the project. We owe our deepest gratitude to all their probing and challenging suggestions and concerns on improving the outcome of the project. Foremost, we are extremely grateful to our supervisor Dr. Ranjith Premasiri for the immense support and guidance given throughout the project. One of the major reasons behind the success of this project is in direct consequence of his support, direction and flexibility provided since the initiation of the project. Foremost, we are grateful to Mr. M.I.D.H. Wijewickrama, the Director, Project Management Division of National Building Research Organization who provided us with the conceptual framework and the required data, generously without whom this project wouldnt have been a success. We would also like to express our great appreciation to all the other staff at National Building Research Organization including the Director General Dr. Asiri Karunawardhana who helped us in numerous ways in accomplishing most of the tasks of this project. We also like to thank Professor U G A Puswewala, the Dean, Faculty of Engineering, Dr. A M K B Abeysinghe, the Head, Department of Earth Resources Engineering, other academic staff in the University including all the lecturers in the Department and the key personnel of the institutes where we had our industrial placement for their generous dissemination of knowledge, wisdom and experience to develop our technical capability to such a level which was the driving force in the successful accomplishment of the project. Last but not least, a special thank goes to all the non-academic staff in the university who supported us in various ways during the whole course period, especially during the field visits.
iv
ABSTRACT
Forecasting the potential for disastrous events such as landslides, has become one of the major necessities in the current world. Most of all the landslides occurred in Sri Lanka are found to be triggered mostly by intense rainfall events. The study area is the landslide near Gerandiella waterfall which is located by the 41st kilometer post on Nuwara Eliya-Gampala main road in Kotmale Division in Sri Lanka. The landslide endangers the entire Kotmale town beneath the slope. Geographic Information System (GIS) platform is very much useful when it comes the need of emulating the real-world processes. The models are used in a wide array of applications ranging from simple evaluations to the levels of forecasting future events. This project investigates the possibility of developing a dynamic model to map the spatial distribution of the slope stability. The model incorporates several theoretical models including the infinite slope model, Green Ampt infiltration model and Perched ground water flow model. A series of rainfall values can be fed to the model as the main input to simulate the dynamics of slope stability. Hydrology model developed using GIS is used to quantify the perched water table height, which is one of the most critical parameters affecting the slope stability. Infinite slope stability model is used to quantify the degree of slope stability in terms of factor of safety. DEM was built with the use of digitized contour data obtained from the Survey department and total station survey results obtained from NBRO. Stratigraphy was modeled in Surfer using borehole data and resistivity images. Data available from rainfall gauges and piezometers were used in calibrating the model. During the calibration, the parameters were adjusted until a good fit between the simulated ground water levels and the piezometer readings was obtained. This model equipped with the predicted rainfall values can be used to forecast of the slope dynamics of the area of interest. Keywords: factor of safety, Geographic Information System, Hydrological model, slope stability
TABLE OF CONTENTS
DECLARATION PAGE OF THE CANDIDATE ..................................................................... i DECLARATION PAGE OF THE SUPERVISOR ...................................................................ii ACKNOWLEDGEMENT ........................................................................................................ iv ABSTRACT ............................................................................................................................... v TABLE OF CONTENTS .......................................................................................................... vi LIST OF FIGURES .................................................................................................................. ix LIST OF TABLES .................................................................................................................... xi LIST OF EQUATIONS ...........................................................................................................xii LIST OF ABBREVIATIONS ................................................................................................ xiii LIST OF APPENDICES ......................................................................................................... xiv 1 INTRODUCTION .............................................................................................................. 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 Chapter overview: ....................................................................................................... 1 Project background ...................................................................................................... 1 Knowledge gap identified ........................................................................................... 2 Scope ........................................................................................................................... 2 Contribution to knowledge .......................................................................................... 2 Statement of significance ............................................................................................ 3 Study area .................................................................................................................... 3 Objectives .................................................................................................................... 3 Project deliverables ..................................................................................................... 4
LITERATURE REVIEW ................................................................................................... 5 2.1 2.2 2.3 2.4 Landslides.................................................................................................................... 5 Types of Landslides .................................................................................................... 5 Landslide causative factors ......................................................................................... 9 Landslides in Sri Lanka ............................................................................................. 10 vi
Landslide triggering factors ...................................................................................... 11 Hydrology cycle ........................................................................................................ 12 Hydraulic conductivity in unsaturated zone .............................................................. 13
Brooks-Corey Model..................................................................................................... 15
2.7.1
2.8
2.8.1
2.9
2.10
2.10.1 2.10.2 2.10.3 2.10.4 2.10.5 2.10.6 2.10.7
METHODOLOGY ........................................................................................................... 23 3.1 3.2 3.3 Chapter overview ...................................................................................................... 23 Model conception ...................................................................................................... 23 Requirement analysis ................................................................................................ 24
Focus group ................................................................................................................... 24 Interviews ...................................................................................................................... 24 Functional requirements................................................................................................ 25
3.4 3.5
vii
3.6 3.7
3.7.2
3.8
3.8.1 3.8.2
3.9 4
RESULTS AND DISCUSSION ....................................................................................... 47 4.1 4.2 4.3 4.4 Results of Calibration for Groundwater levels .......................................................... 47 Results of Calibration for Factor of Safety ............................................................... 50 Results of Model Validation ..................................................................................... 51 Discussion ................................................................................................................. 53
CONCLUSION AND RECOMMENDATION ............................................................... 54 5.1 5.2 5.3 Chapter overview ...................................................................................................... 54 Conclusion................................................................................................................. 54 Recommendation ....................................................................................................... 55
viii
LIST OF FIGURES
Figure 1: Landslide nomenclature for an idealized slum-earth flow (Anon., 2004).................. 5 Figure 2: Different types of Landslides (Anon., 2004) .............................................................. 8 Figure 3: Descriptive representation of the hydrological cycle (Deodhar, 2008).................... 12 Figure 4: Illustration of infiltration capacity (Delleur, 1999) ................................................. 14 Figure 5: Constant-head permeability test (Das, 2010) ........................................................... 16 Figure 6: Relation among moisture retention parameters and soil texture class (Delleur, 1999) .................................................................................................................................................. 18 Figure 8: Model builder interface ArcGIS 10.1 ....................................................................... 19 Figure 9: Classification of elements in Model Builder (esri, n.d.) .......................................... 21 Figure 10: Rich picture of the model ....................................................................................... 23 Figure 11: Preparation of rainfall time series data to feed to the model .................................. 27 Figure 12: 1:10000 contour map of the AOI of 1 sq.km ......................................................... 28 Figure 13: Depression less DEM ............................................................................................. 28 Figure 14: Attributes of different geological units .................................................................. 29 Figure 15: Borehole lines visualized in Arc Scene .................................................................. 30 Figure 16: Upper most surface of soil layers visualized in Arc Scene .................................... 30 Figure 17: The total model process iterates for each rainfall value ......................................... 33 Figure 18: Basic hydrological cycle ........................................................................................ 34 Figure 19: Configuration of ArcGIS tools and nested sub-models to simulate the soil strata. 35
ix
Figure 20: Percolation from the soil layer immediately above the shear plane directly contributes to the increase in perched water table ................................................................... 36 Figure 21: Configuration of nested models and ArcGIS tools to stimulate the effect of water height on slope stability ........................................................................................................... 37 Figure 22: Three main components affecting the change in water table ................................. 37 Figure 23: Water height sub-model interface .......................................................................... 39 Figure 24: Factor of Safety calculation using Raster Calculator ............................................. 41 Figure 25: Factor of Safety sub-model interface ..................................................................... 42 Figure 26: Main model interface.............................................................................................. 43 Figure 27: Calibration Plots ..................................................................................................... 48 Figure 28: Ground water heights in the time step of Rainfall 80mm (2012-07-09) ................ 49 Figure 29: Ground water heights in the time step of Rainfall 24mm (2012-07-23) ................ 49 Figure 30: Raster images representing outputs of stability model ........................................... 50 Figure 31: Raster images representing the outputs of the stability model ............................... 50 Figure 33: Validation plots ...................................................................................................... 52
LIST OF TABLES
Table 1: Different types of landslides (Anon., 2004) ................................................................ 6 Table 2: Natural causes for landslides ....................................................................................... 9 Table 3: Case histories of Sri Lankan landslides ..................................................................... 10 Table 4: Estimated saturated hydraulic conductivities ............................................................ 17 Table 6: Representative Porosity Values (Laboratory, n.d.) .................................................... 18 Table 7: Details of sub models nested within the main model ................................................ 31 Table 8: Model Parameters for the Calibrated Model ............................................................. 45 Table 9: Goodness of Fit Parameters ....................................................................................... 47 Table 10: Goodness of Fit Parameters ..................................................................................... 50 Table 11: Goodness of Fit Parameters ..................................................................................... 51
xi
LIST OF EQUATIONS
Equation 1 ................................................................................................................................ 15 Equation 2 ................................................................................................................................ 15 Equation 3 ................................................................................................................................ 15 Equation 4 ................................................................................................................................ 16 Equation 5 ................................................................................................................................ 16 Equation 6 ................................................................................................................................ 16 Equation 7 ................................................................................................................................ 36 Equation 8 ................................................................................................................................ 38 Equation 9 ................................................................................................................................ 38 Equation 10 .............................................................................................................................. 40 Equation 11 .............................................................................................................................. 40 Equation 12 .............................................................................................................................. 45 Equation 13 .............................................................................................................................. 45 Equation 14 .............................................................................................................................. 46 Equation 15 .............................................................................................................................. 46 Equation 16 .............................................................................................................................. 46
xii
LIST OF ABBREVIATIONS GIS NBRO DEM ESRI USGS FoS Geographic Information System National Building Research Organization Digital Elevation Model Environmental Systems Research Institute United States Geological Survey Factor of Safety
xiii
LIST OF APPENDICES Appendix A Appendix B Appendix C Appendix D Appendix E Borehole data ( 5 boreholes) Daily rainfall data used in Model Calibration Daily rainfall data used in Model Validation Piezometer readings used in Model Calibration Gerandiella landslide site Survey data (CAD file)
xiv
INTRODUCTION
1.4 Scope
In our research the main focus is on investigating the possibility of developing a dynamic model to forecast the landslide susceptibility of the area of interest. Dynamic modeling of slope stability of an area of interest is not a common practice, which created the basic platform for this research. Landslides are triggered in most cases by a rise of pore pressure in the slope, which decreases the resistance of the soil material. Therefore temporal occurrences of landslides are related to periods with high rainfall amounts. A combined Hydrological Slope stability model will be used to investigate the spatial and temporal patterns of slope instability to be used as landslide forecasting model. The overarching research question was addressed by developing a GIS based dynamic model in ArcMap 10.1 which was validated using the existing rainfall data for the area of interest. Gerandiella landslide site was selected as the area of interest, and the existing elevation data, borehole data and the rainfall data were used as the inputs for the model for the validation purpose.
landslide susceptibility. Ground movement modeling activities can be exposed to the use of a novel technology which may have significant benefits compared to the existing methodologies in practice. The academic community will be benefited where they may gain access to new knowledge generated through this research. Based on the results of this research new studies can be developed and implemented for the benefit of the industry and the country as a whole.
1.8 Objectives
To investigate the potential of develop a dynamic model to forecast the landslide susceptibility of the area of interest To develop a dynamic model using ArcGIS (Arc Map 10.1) to calculate the Factor of Safety provided a series of rainfall data as the dynamic input 3
To develop the Digital Elevation Model using the existing elevation data To model the stratigraphy of the area of interest using the existing borehole data To validate the model using existing rainfall data.
LITERATURE REVIEW
2.1 Landslides
The term landslide describes a wide variety of processes that result in the downward and outward movement of slope-forming materials including rock, soil, artificial fill, or a combination of these. The materials may move by falling, toppling, sliding, spreading, or flowing (USGC Website, 2004).
Figure 1: Landslide nomenclature for an idealized slum-earth flow (Anon., 2004) Sudden movements of rock or soil mass are fairly common and are generally termed as landslides. A landslide can be a movement of either a sloping mass or the crest or the foot of a hill or even the cut surface of a slope (R.M.S.BANDARA, 2003).
Table 1: Different types of landslides (Anon., 2004) Landslide category Description Slide in which the surface of rupture is curved Rotational slides concavely upward and the slide movement is roughly rotational about an axis that is parallel to the ground surface and transverse across the slide (Fig.2A) The landslide mass moves along a roughly planar surface with little rotation or backward tilting (fig.2B) Translational slides A block slide is a translational slide in which the moving mass consists of a single unit or a few closely related units that move downslope as a relatively coherent mass (fig. 2C) It occurs in the form of a rapid mass movement in which a combination of loose soil, rock, organic matter, air and water mobilized as slurry that flow downslope (fig. 2F). Debris flows include <50% fines. Debris flows are commonly caused by intense surfaceDebris flow water flow, due to heavy precipitation or rapid snowmelt that erodes and mobilizes loose soil or rock on steep slopes. Debris-flow source areas are often associated with steep gullies, and debris-flow deposits Flows are indicated by the presence of debris fans at the mouths of gullies. Debris avalanche A variety of very rapid to extremely rapid debris flow (fig. 2G) Earthflows have a characteristic hourglass shape (fig. 2H). The slope material liquefies and runs out, Earth flow forming a bowl or depression at the head. The flow itself is elongate and usually occurs in fine-grained materials or clay bearing rocks on moderate slope and
Slides
under saturated conditions. However, dry flows of granular material are also possible. A mudflow is an earthflow consisting of wet material flowing rapidly and that contains at least 50 percent Mud flow sand, silt and clay-sized particles. In some instances, for example in many newspaper reports, mudflows and debris flows are commonly referred to as mudslides. Creep is the imperceptibly slow, steady, downward movement of slope-forming soil or rock. Movement is Creep caused by shear stress sufficient to produce permanent deformation, but too small to produce shear failure. (fig.2A) Toppling failures are distinguished by the forward rotation of a unit or units about some pivotal point, Topples below or low in the unit, under the actions of gravity and forces exerted by adjacent units or by fluids in cracks (fig. 2E). Falls are abrupt movements of masses of geologic Falls materials, such as rocks and boulders, that become detached from steep slopes or cliffs (fig. 2D). Lateral spreads are distinctive because they usually occur on very gentle slopes or flat terrain (fig. 2J). The dominant mode of movement is lateral extension accompanied by shear or tensile fractures. The failure is caused by liquefaction, the process whereby Lateral spread saturated, loose, cohesion less sediments (usually sands and silts) are transformed from a solid into a liquefied state. Failure is usually triggered by rapid ground motion, such as that experienced during an earthquake, but can also be artificially induced. When
coherent material, either bedrock or soil, rests on materials that liquefy, the upper units may undergo fracturing, extension, and may then subside, translate, rotate, disintegrate, or liquefy and flow.
Human-made causes of Landslides may include the following: Excavation of slope or its toe Loading of slope or its crest Drawdown (of reservoirs) Deforestation Irrigation Mining Artificial vibration Water leakage from utilities
Watawala railway, Nuwara- Damages to train Deforestation Liya and the railway 200mm rainfall within 12 hours
Hela-uda, Ratnapura
48 deaths
for
Koslanda, Badulla
Early NBRO
warning
by
10
Ela-patha, Ratnapura
87 deaths
Matara
(Kiripitiya, 24 deaths
Improper land-use
5 houses
Heavy rainfall
Nov, 2006
Peradeniya, Kandy
Heavy
rainfall
and
6 deaths 6 houses
Padiyapelella
Elamale 12 houses
junction, Nuwara-Eliya 11,12 Jan 2007 Diyanilla, Nuwara-Eliya 19 deaths Improper irrigation and vegetation
11
Figure 3: Descriptive representation of the hydrological cycle (Deodhar, 2008) The various processes in a hydrological cycle are as follows (Deodhar, 2008) 1. Evaporation: It is a process by which water from water bodies or land mass passes into vapour state and is diffused into the atmosphere. 2. Transpiration: It is the process by which water passes from the liquid state to the vapour state through plant metabolism. Evaporation and transpiration are sometimes combined, as often their separate assessment is difficult and are termed as evapotranspiration. 3. Precipitation: It is the general term for all the moisture emanating from the clouds and falling on the ground. Precipitation can occur in many different forms, including rain, hail, mist, snow, ice and so on. 4. Through fall: It is that part of precipitation, intercepted by vegetation, which then falls on the ground. 5. Interception: It is that part of precipitation, which is received and retained by vegetation and evaporates later.
12
6. Infiltration: It is that part of precipitation, which enters into the ground and then flows downwards. 7. Vapor diffusion: It is that part of water retained by the soil, which flows in the form of vapor towards the ground surface. 8. Surface detention: When a river flows, a significant volume of water is contained in the river channel. This temporary storage in the river channel is called surface detention. 9. Depression storage: When there is precipitation over a catchment, and before the river starts flowing, part of water is stored in ditches, small ponds, and so on, depending upon the topography of the catchment. This quantum of water thus stored is known as depression storage or surface retention. 10. Surface runoff: The precipitated water after meeting all the requirements flows in a stream. This is known as surface runoff or overland flow. 11. Interflow: When water infiltrates, it starts moving laterally towards a stream and appears on the surface. This is known as interflow. It is above the groundwater table. The velocity of flow is very low as compared to the surface flow. It is also known as sub-surface storm flow, sub-surface runoff, storm seepage and secondary base flow. 12. Groundwater flow: The infiltrated water may reach the saturated zone of water below ground and may get stored between its pores and voids between particles. This water stored eventually may flow towards a stream. This happens when the groundwater table is above the stream water level. However, when the stream water level is above the groundwater table, there may be a flow from stream to groundwater. This movement of water below the ground is known as groundwater
13
4. the horizontal hydraulic conductivity of the aquifer being recharged and its hydraulic gradient, which determine the rate at which recharged water will be carried out of the recharge area 5. Vertical hydraulic conductivity of the soil being recharged 6. Presence of man-made alterations to the subsurface, such as drainage tiles that carry water away to run off in surface streams (Delleur, 1999). Recharge is controlled by infiltration capacity. This phenomenon was studied by Horton and is the maximum rate at which a soil will permit the entry of water. It is generally a time dependent parameter because it depends upon rate of change of the saturation ratio and hence, the moisture content.
Figure 4: Illustration of infiltration capacity (Delleur, 1999) This relationship is illustrated with Figure 4. Initial infiltration into the soil is rapid as empty pores in dry soil begin filling. When infiltration continues, the infiltration capacity decrease to an eventual steady state rate where the infiltration water enters at same rate as that which it is transported downward by porous media flow. Once infiltration rate falls below the precipitation rate, the excess water will become runoff know as overland flow. Generally, this will happen after all surface storage requirements are met, e.g., leaf storage in trees, crops, and grasses, and depression storage in hollows in the ground. From these considerations, it is obvious that the most effective rain for watering plants and recharging groundwater supplies is the slow steady kind. And not the intense, high-rate storm, most of which runs off and does not infiltrate.
14
The modeling of water movement through soil typically requires the functional representation of soil hydraulic properties, such as soil moisture retention curve and the unsaturated hydraulic conductivity. There are two commonly used models such as Brooks and Corey (1964) and van Genuchten (1980). In general, the van Genuchten model matches experimental data more satisfactorily than the model of Brooks and Corey. However, the functional form of van Genuchten model is complicated and limits its usefulness for large number of analytical solutions available for infiltration and drainage problems. On the other hand, Brooks-Corey model yields conductivity and water retention functions that are easy to manipulate mathematically (Kumar & Mittal, n.d.). 2.7.1 Brooks-Corey Model The unsaturated hydraulic conductivity gets its maximum when the infiltration capacity reached. Water will not flow until the water content in the soil is sufficient and the pressure head becomes less negative, when the soil is initially dry. The specific discharge is then given as
Equation 1
= pressure gradient
Brooks and Corey (1964) give the formula for hydraulic conductivity as
Equation 2
() =( )
Where K () = hydraulic conductivity Ks = saturated hydraulic conductivity r = residual water content n = porosity and
Equation 3
= = () Where, Q = volume of water collected A = area of cross section of soil specimen T = duration of water collection
Equation 5
= Where specimen,
Equation 6
of
L=length
= ( )
2.9.2 Saturated volumetric moisture content Commonly used in the description of infiltration, is a volumetric expression for the amount of moisture in a soil. It is defined as the ratio between volume of water and total volume (Delleur, 1999). 2.9.3 Effective porosity Also known as drainable porosity, represents the total amount of water that can be stored in the soil. It is defined as the difference between the saturated and residual volumetric moisture content (Delleur, 1999). 17
2.9.4 Residual volumetric moisture content This represent the amount of water that remains in a soil mass even when a high vacuum pressure is applied to the soil. The relationships are shown in Figure7 and Table5.
Figure 6: Relation among moisture retention parameters and soil texture class (Delleur, 1999)
Table 5: Representative Porosity Values (Laboratory, n.d.) Material Type Clay Silt Sand Gravel Total Porosity 0.34-0.57 0.34-0.51 0.31-0.46 0.24-0.36 Effective Porosity 0.01-0.18 0.01-0.39 0.18-0.43 0.13-0.25
18
Model canvas: The model canvas is the white empty space in a model. Model diagram: The model diagram is the appearance and layout of the tools and variables connected together in a model.
19
2.10.1 Model elements There are three main types of model elements: tools, variables, and connectors. 1. Tools: Geo processing tools are the basic building blocks of workflows in a model. Tools perform various operations on geographic or tabular data. When tools are added to a model, they become model elements (Anon., 2012). 2. Variables: Variables are elements in a model that hold a value or a reference to data stored on disk. There are two types of variables: 3. Data: Data variables are model elements that contain descriptive information about data stored on disk. Properties of data that are described in a data variable include field information, spatial reference, and path. 4. Values: Value variables are values such as strings, numbers, Booleans (true/false values), spatial references, linear units, or extents. Value variables contain anything but references to data stored on disk. 5. Connectors: Connectors connect data and values to tools. The connector arrows show the direction of processing. There are four types of connectors: 1. Data: Data connectors connect data and value variables to tools. 2. Environment: Environment connectors connect a variable containing an environment setting (data or value) to a tool. When the tool is executed, it will use the environment setting. 3. Precondition: Precondition connectors connect a variable to a tool. The tool will execute only after the contents of the precondition variable are created. 4. Feedback: Feedback connectors connect the output of a tool back into the same tool as input.
20
A model process consists of a tool and all variables connected to it. Connector lines indicate the sequence of processing. Many processes can be chained together to create a larger. 2.10.3 Intermediate data When a model is run, output data is created for each process in the model. Some of this output data is only created as a middle step to connect to other processes that will create the final output. The data generated from these middle steps, called intermediate data, is often (but not always) of no use once the model has finished running. You can think of intermediate data as temporary scratch data that should be deleted after the model has run. However, when you run a model from the Model Builder window, intermediate data is not deletedit is up to you to delete it. 2.10.4 Model validation Model validation refers to the process of making sure all model variables (data or values) are valid. 2.10.5 Model parameter Model parameters are the parameters that appear on the model tool dialog box. Any variable in the model can be made a model parameter.
21
2.10.6 Running a model within Model Builder Selected processes in a model or the entire model can be run from within Model Builder. A model can be run from its tool dialog box. 2.10.7 Workspace environments There are four workspace environments that can be used in Model Builder to simplify model data management: 1. Current: Tools that honor the Current Workspace environment setting use the workspace specified as the default location for geo-processing tool inputs and outputs. 2. Scratch: Tools that honor the Scratch Workspace environment setting use the specified location as the default workspace for output datasets. The Scratch Workspace is intended for output data you do not wish to maintain. 3. Scratch folder: The Scratch Folder is the location of a folder you can use to write filebased data, such as shape files, text files, and layer files. It is a read-only environment that is managed by ArcGIS. 4. Scratch GDB: The scratch GDB is the location of a file geo-database you can use to write temporary data.
22
METHODOLOGY
Increase in Waterheight
Uppermost soil layer Increase in moisture content Percolation into the next soil layer
Percolation into the bottom-most soil layer Intermediate soil layers Increase in moisture content
Figure 9: Rich picture of the model
23
3.3
Requirement analysis
The main purpose of conducting a requirement analysis was to identify the main stakeholder groups and their requirements in order to define up to what level we can cater them. During this stage we could further narrow down our scope and modify the rich picture of the model. 3.3.1 Focus group We identified National Building Research Organization (NBRO) as the major focus group who operates under the Ministry of Disaster Management. It is the national focal point of Sri Lanka for all landslide related studies and services. NBRO also contributes to the development sector through its services in the field of Geotechnical Engineering, Research and Environmental studies etc. NBRO carries out applied research on landslides disaster management and constantly improves its expertise in landslide hazard identification, risk evaluation, appropriate construction and land use practices, cost effective mitigation measures and real time landslide early warning and forecasting. Landslide forecasting model has a high potential in helping them in their research and other activities. Our second target group is a broader one which comprises of Researchers and Students who are interested in the studies related to landslides and other ground movements. 3.3.2 Interviews Effective communication methods in the form of formal and informal interviews helped us in identifying the stakeholder requirements. After developing the rich picture of the model we interviewed several key personnel at NBRO to gather information on the methods they use currently to monitor landslide potential and the limitations and drawbacks of the methods being used. During the initial site visit to Gerandiella, we could interview several people who resides around the site in order to collect information as much as possible. From them we could obtain information related to landslide history, return period of rainfall, rainfall intensity etc. By conducting a literature review we could attain a theoretical knowledge. This knowledge could be enhanced and improved by conducting interviews where we could integrate practical knowledge and experience of the interviewees towards achieving a more reliable information and knowledge to be used in designing the model.
24
During the interviews we got insights to the technologies that are currently used in landslide monitoring and forecasting activities and also to the technologies that have the potential to be implemented in the near future. Besides we could realize that there are number of advanced technologies that have been implemented and in use, in other countries which are still not in use in Sri Lanka. Based on the knowledge and information we gathered we could identify the following categories of requirements. 1. Functional Requirements 2. Non-Functional Requirements 3. Performance Requirements 4. Design Requirements We defined the requirements specific to our model under each category as follows. 3.3.3 Functional requirements The functional requirements describe the core functionality of the application. We could define several data requirements and functional process requirements under functional requirements. 3.3.3.1 Data Requirements The main input data requirements for the model were identified as follows. 1) Rainfall time series 2) Evapotranspiration and interception time series 3) Terrain elevation data 4) Number and type of soil strata 5) Thickness of each soil stratum 6) Initial moisture content 7) Initial water height 8) Saturated hydraulic conductivity of each soil type 9) Saturated moisture content of each soil type 10) Residual moisture content of each soil type 11) Cohesion between bottom most soil layer 12) Angle of friction
25
3.3.3.2 Functional Process Requirements The process requirements of the model were identified as follows. 1) To extract rainfall value from the rainfall database 2) To calculate precipitation accounting for the loss caused by evapotranspiration 3) To calculate the increase in moisture content of each soil layer and update the moisture content by replace the existing one with the calculated figure. 4) To calculate the amount of percolation into the soil layer immediately below. 5) To calculate the changes to the water height caused by percolation from the bottommost soil layer and gravity flow. 6) To calculate and update the pore water pressure accounting for the increase in water height. 7) To calculate the factor of safety using the updated pore water pressure. 8) To perform all the calculations to each pixel simultaneously. 9) To iterate the total model process for each rainfall value in the time series data set. 3.3.3.3 Hardware and Software requirements Hardware Computer with a minimum 2GB RAM, minimum 2.2 GHz CPU speed Intel Pentium 4, Intel Core Duo, or Xeon Processors; SSE2 minimum Software ArcGIS 10.1 Surfer
Figure 10: Preparation of rainfall time series data to feed to the model Since the unavailability of evapotranspiration and interception time series data, 10% was written against the original rainfall values to calculate the precipitation values accounting for the respective losses caused by evapotranspiration and interception.
27
3.5.2 Digital Elevation Model Depression less DEM of the area of interest was created to be fed to the model as follows. DEM was required to calculate the flow direction raster and to calculate the slope angle of the terrain. The accuracy of DEM was considered to be critical in obtaining reasonable outputs from the model.
Figure 11: 1:10000 contour map of the AOI of 1 sq.km 3.5.2.1 Elevation raster Elevation raster was created using the 1:10,000 contour map of the AOI purchased from the Survey department of Sri Lanka. Topo to Raster tool in ArcGIS was used in creating the raster. 3.5.2.2 Locating sinks The elevation raster was fed to the Flow direction tool in ArcGIS and the resultant flow direction raster was fed back to the Sink tool in ArcGIS. The result is a raster that identifies any existing sinks in the data.
28
3.5.2.3 Filling the sinks By filling the Sinks using the Fill tool, a depression less DEM was created as shown in Figure14. 3.5.3 Geological unit information Using the borehole data the elevation values were interpolated using of the upper-most surface of each geological unit (soil layer) as shown in figure One separate raster for each geological unit (soil layer) was prepared, having its pixel values representing the average thickness of ach 50mX50m area on the ground. For each geological unit the following parameters were obtained through the literature survey and borehole data. Initial moisture content () Saturated moisture content ( ) Residual moisture content( ) Saturated hydraulic conductivity ( )
Figure 13: Attributes of different geological units 3.5.4 Other site specific parameters 3.5.4.1 Cohesion The cohesion is the term used in describing the shear strength soils. Higher the value of cohesion, lower the potential of a landslide. To analyse the worst case scenario, we can assume the cohesion to be Zero. However, in our design we defined the Cohesion as a model parameter, so that the user can adjust it according to the site specific conditions. 3.5.4.2 Friction angle Soil friction angle is also a shear strength parameter of soils. The model was designed giving the ability of the user to define the site specific angle of friction.
29
Figure 15: Upper most surface of soil layers visualized in Arc Scene
30
percolation into the next soil layer immediately below. Calculates the increase in
percolation into the next soil layer immediately below. Calculates the increase in
percolation into the next soil layer immediately below. Calculates the increase in
31
4) Soil Layer4
Percolation raster produced Calculates by Soil Layer3 Precipitation raster Initial moisture content Soil specific parameters: o Saturated hydraulic conductivity o Saturated moisture content o Residual moisture content Percolation raster produced Calculates by Soil Layer4 Precipitation raster Initial moisture content Soil specific parameters: o Saturated hydraulic conductivity o Saturated moisture content o Residual moisture content Ground water flow Initial moisture content Flow direction raster Saturated hydraulic conductivity Saturated water content Calculates
the
amount
of
percolation into the next soil layer immediately below. Calculates the increase in
5) Soil Layer5
the
amount
of
6) Water Height
conductivity, slope of the terrain, initial water height and width of a pixel Calculates the increase and
decrease in water height caused by gravity flow Calculates the increase in water height caused by percolation from the bottom-most soil layer into the shear plane Calculates and update the new water height accounting for the changes calculated above 32
7) Factor of Safety Recalculated water height Angle of friction Cohesion Bulk unit of soil and water Slope angle of the terrain
Calculates pore water pressure using the recalculated height Calculates the factor of safety using pore water pressure water
Figure 16: The total model process iterates for each rainfall value
33
Then the model reads the first precipitation value of the data set and calculates how much water (as a height) does the uppermost layer absorbs to increase its moisture content and how much does it allow to percolate into the next soil layer. This process is continued repetitively for all the soil layers. The basic hydrological cycle used in developing this model is presented in Figure19.
Figure 17: Basic hydrological cycle According to the stratigraphy of the study area, we could define four basic geological units (soil layers). The same process was simulated using ArcGIS model builder as shown in Figure20.
34
Figure 18: Configuration of ArcGIS tools and nested sub-models to simulate the soil strata 35
3.7.1.1 Percolation Unsaturated percolation to the ground water was assumed to take place by gravity according to the equation 1. For each soil layer the percolation flux (mm/day) was calculated using the equation. Equation 7 = ( ) Where, Pr Ks r s m = Percolating flux (mm /day) = Saturated hydraulic conductivity (mm/day) actual volumetric soil moisture content % = residual volumetric moisture content % = saturated volumetric moisture content % = constant (between 3 and 8)
Figure 19: Percolation from the soil layer immediately above the shear plane directly contributes to the increase in perched water table
36
Figure 20: Configuration of nested models and ArcGIS tools to stimulate the effect of water height on slope stability
3.7.1.2 Change in water height The model assumes the total impact on the water height is due to three main components as explained in figure 22.
Net inflow due to gravity flow 1
50m
50m
Increase in water height due to percolation from the bottom most soil layer 3
Figure 21: Three main components affecting the change in water table 37
By accounting for these three components we calculated the new water height using equation 2 given below.
Equation
8 = + 1 2 + 3
Each of these three components (1 , 2 , 3 ) were calculated as follows. Outflow due to groundwater flow (2 ) The routing of the ground water above the more or less impermeable rock occurs by the topographical surface. The flux Q is calculated according to the equation 4 as follows.
Equation
9 =
Where: Q Ks Zw B =Ground water flow (mm3/day) =Saturated hydraulic conductivity (mm/day) =Height of the ground water table (mm) =Width of flow (=width of pixel) (mm)
Sin =Sinus of the topographical slope Change in water table elevation due to this groundwater discharge out of a pixel was presented as a height using equation 5. 2 = Inflow due to gravity flow (1 ) For a particular pixel, there exist a number of neighbouring pixels around it of which the outflow may become an inflow to the pixel under consideration depending on the flow direction. ArcGIS software facilitates calculating the flow direction and flow accumulation using the digital elevation model and the outflow values. Using the Digital Elevation model flow direction raster was calculated. The flow direction raster was fed to the flow accumulation tool weighted by the calculated outflow from each 2 (4 + 0.01 4)
38
pixel(2 ). The resultant raster pixels had the values representing the amount of water height change caused by the groundwater inflow to each pixel. Inflow due to percolating flux(3 ) Effect to the water height caused by percolation from the soil layer which is immediately above the shear plane is calculated using equation 3 as follows: 3 = Where, 4 = Saturated moisture content of the soil layer immediately above the shear plane 4 = Moisture content of the soil layer immediately above the shear plane Perc4= percolating flux of the soil layer immediately above the shear plane (mm/day) 4 ( 4 + 0.01 4 )
3.7.1.3 Factor of Safety The stability of the slope is calculated in terms of a Safety Factor. The Safety Factor is the ratio between the driving and the resisting forces of the slope. The infinite slope model is used to calculate the stability. In this model it is assumed that the slip surface of the landslide is running parallel to the topographical slope. In that case the stability can be calculated for each pixel. 39
The safety factor was calculated using Equation 10 + tan (zcos 2 u) = sin cos Where: F = Factor of Safety c'=soil cohesion (kN/m2) =unit weight of soil (kN/m3) Z=depth of the soil (slip surface) (m) u=pore water pressure (kN/m2) =angle of the topographical slope
The pore water pressure is related to the height of the ground water above the slip plane in the slope as follows Equation 11 = 2 w Zw = Unit weight of water (kN/m3) = Height of the water table above the slip surface (m)
The height of the water table is related to the input of rain and the subsurface drainage of ground water along the slope. It is calculated with the hydrological model. Factor of safety is calculated using the new water height for each pixel. The basic assumption used in this model is that, all the other parameters required in calculating the factor of safety remain constant for the area of interest during the particular period of time. Thereby, it is considered that the only dynamic input for the model is rainfall and the other parameters are comparatively stagnant In this model we use rasters with a 50m resolution. For each pixel, the factor of safety is calculated at each iteration i.e. for each rainfall value. The model is iterated for each rainfall value in the time-series. At each iteration, impact on the existing water height is calculated and adjusted for. Also at each iteration a raster is created indicating how the slope stability changes according to the rainfall input. 40
Depending on the value of the Factor of Safety, the degree of stability is defined according to the criteria below: Factor of Safety > = 1 Stable Factor of Safety <1 Not Stable
. Figure 23: Factor of Safety calculation using Raster Calculator 3.7.2 Model outputs Final output of the model is a set of time-series rasters indicating the areas which have high or low susceptibility to landslides.
41
42
43
44
Hydrological model was calibrated by changing m value of Equation 7 in the range of 3-8. The extreme goodness of fit was obtained though applying m=4. The goodness of fit was calculated using following statistics. 1. Efficiency Coefficient The reliability of the hydrological model is evaluated based on the Efficiency Index (EI) as described by Nash and Sutcliffe (1970). The EI was developed to evaluate the percentage of accuracy or goodness of the simulated values with respect to their observed values. Following equations show the error parameters to be calculated during the calibration and validation of the model.
Equation 12
= 1
(, , )2 (, , )2
Where Qobs,i is observed discharge at time t, and Qsim,i is modeled discharge at time t. Qobs,mean is mean of observed discharge. An efficiency of 1 (E = 1) corresponds to a perfect match of modeled discharge to the observed data. 2. Standard Deviation This measures the variability dispersion of observed and simulated data. It is defines as following equation
Equation 13
= 45
( )2
Where, = standard deviation = mean of data set n = number of data in the data set
Equation 14
= (1 Where, = standard deviation of total data set = standard deviation of observed data = standard deviation of observed data 3. Root Mean Square Error (RMSE)
Equation 15
) 100
RMSE
i 1
( X obs,i X sim,i ) 2 n
Where Xobs is observed values and Xsim is modelled values at time i. 4. Mean Percentage Error (MPE) This is subject to the averaging of the positive and negative errors.
Equation 16
100] /
This method was used to increase the accuracy of the model. Visually compared the slide area and the factor of safety values in each cells. The model was calibrated to obtain where the factor of safety values smaller than 1cells representing the slide area. The parameters fed to calibrate the stability model is shown in Table 10.
46
47
48
Figure 27: Ground water heights in the time step of Rainfall 80mm (2012-07-09)
Figure 28: Ground water heights in the time step of Rainfall 24mm (2012-07-23)
49
Figure 30: Raster images representing the outputs of the stability model
50
Bore Hole
Efficiency Index
Root Mean Square Mean Percentage Error (RMSE) 44.04 99.95 Error (MPE) % 0.00 0.02
BH1 BH2
0.62 -1.42
51
52
4.4 Discussion
There were several important factors considered during model calibration and validation. There were assumptions, input parameter values and distributions, output values and conclusions. Initial calibration attempt concentrated on the output of the model. We did a real system measurements to analyze the model hence it is the most reliable and preferred method. In the case of calibration this entire model we were unable to approach the extreme goodness of fit. Finally we have chosen the most fitted model to do the validation. Therefore the validated results also representing some deviation with the observed data. There are many reasons for that the simulated model prediction does not show a proper agreement between the observed data. The main assumption of this model is that it does not encounter the entire catchment ground water flow. It represent only the ground water floor of the interested area. As this is a hilly area this will highly affect in ground water levels. The calibration should done in long term basis to determine invariance. Since the rainfall at Garandiella gauging station were available only for seven months, we had to select short time periods for model calibration and validation. At least there should be two year data (representing two water years) set to carry out calibration and validation for consecutive years. Otherwise it does not meet the storm and dry period in Sri Lankan climate condition. Generally the storm rainfall duration is not exceed more than 3 hours in Sri Lanka. Therefore rainfall runoff and groundwater flow can travel long distance and accumulate within a day. Therefore it is reasonable to use daily rainfall as a calibration parameter. Soil layer thickness is a major parameter for this model. It should be given for the entire area by layer by layer. It cannot represent when there are interconnecting soil layers exist. Sometimes they should be neglected. Therefore the properties of subsurface could be changed. We were able to obtain borehole logs from NBRO. The soil properties were procured from general data after doing a literature review. This intensifies the mismatch of real condition of subsurface model. Moreover we did not modeled geological features such as discontinuities, fractures etc.
53
5.2 Conclusion
Landslides are a type of the natural disasters that can create catastrophic effects endangering the lives, infrastructure and other environmental settings. These can be triggered by natural causative factors as well as man-made causative factors. This triggers the need of identifying and continuous monitoring of landslide prone areas where the model we developed has a high potential to be useful in such applications. The model was developed to investigate the ability of using a dynamic model to forecast landslide potential of a rainfall induced landslide site. The model was developed in ArcGIS platform using the model builder application in ArcMap10.1. The model was designed in such a way that a time series of rainfall values can be fed to the model to along with the site specific parameters and the model gives a time series rasters indicating the factor of safety. The model has certain set of constraints. The model assumes that the water height change is only triggered by rainfall events. The model will remain valid as long as all the other parameters except rainfall remain constant. The only dynamic input for the model is rainfall time series. Hence the model will be valid only to a specific locality. Besides it assumes the soil strata are uniform and in the same sequence within the area of interest. Most of the parameters of the model are site specific. Hence, several modifications need to be done, for it to be used for another site. The modifications to the model, and the execution of the model requires knowledge, experience and skills developed in GIS platforms. Hence may not be user friendly for the users with no GIS knowledge. The model requires large number of data to be fed in the form of geotechnical parameters which are not readily available and can be expensive to collect. 54
The model incorporates large number of theoretical equations which may not be realistic or applicable in the real world scenario.
5.3 Recommendation
The model gave reasonable results but can be improved further to get better results and make use of it in real-world applications. There are number of ways in which the functional requirements of the model can be improved into a better level. Model results and its validity can be improved with more frequent data collection followed by continuous updating of the model inputs. Hourly rainfall data will be more effective to be used in this kind of a model. Constraining the area further and increasing the resolution of the raster will improve the accuracy of the results. The model has a high level of potential to be improved into a landslide forecasting dynamic model if forecasted rainfall values are fed in as the input. The non-functional requirements also need to be identified and addressed when developing this model where they also act critical in determining whether this model is going to be useful for the users or not. One such non-functional requirement is the minimal lead time between data collection and modeling activities. A proper database need to be maintained in such a way that the user has access to relevant rainfall data in an effective and timely manner. The database can be designed with the ability to be updated with the forecasted hourly rainfall values as frequently as possible, which would result in a minimal lead time. Another important set of requirements that are non-functional are the communication requirements. The output of the model needs to communicate a warning signal to the relevant authorities and the community so that the overall purpose of developing such a model will become accomplished. Effective modes of communication and data exchange channels need to be integrated with the model into a system. It is recommended that this model can be developed into a software application that can be executed within a normal operating system so that the users will be happy that they are provided with a hassle free environment.
55
REFERENCES
Agriculture, U. S. D. o., n.d. Natural Resources Conservation Service. [Online] Available at: http://www.nrcs.usda.gov/wps/portal/nrcs/detail/soils/home/?cid=nrcs142p2_053573 [Accessed 24 09 2013]. Anon., 2004. [Online] Available at: http://pubs.usgs.gov/fs/2004/3072/fs-2004-3072.html Anon., 2012. ArcGIS Help 10.1. [Online] Available at: http://resources.arcgis.com/en/help/main/10.1/index.html#/What_is_ModelBuilder/002w0000 0001000000/ [Accessed 7 03 2014]. Bandara, R. M. S., 2003. VIDURAWA. [Online] Available at: http://thakshana.nsf.ac.lk/pdf/VIDURAWA/VIDUE-22(2)A/VIDUE-22(2)A9.pdf [Accessed 19 July 2013]. BAPTISTA, P., BASTOS, L., BERNARDES, C. & CUNHA, T. a. D. J., 2008. Monitoring Sandy Shores Morphologies by DGPS A practical tool to Generate Digital Elevation Models. Journal of Coastal Resarch, 24(6), p. 1516. Das, B. M., 2010. Principles of Geotechnical Engineering, Stamford: Cengage Learning. Delleur, J., ed., 1999. In: The Handbook of Groundwater Engineering. s.l.:CRC Press LLC. Deodhar, M. J., 2008. Elementary Engineering Hydrology. s.l.:Pearson Education. Kumar, C. P. & Mittal, S., n.d. DETERMINATION OF SOIL HYDRAULIC PROPERTIES IN A PART OF HINDON RIVER CATCHMENT USING SOILPROP SOFTWARE, Uttarakhand: Roorkee . Laboratory, A. N., n.d. Environmental Science Division (EVS). [Online] Available at: http://www.evs.anl.gov/ [Accessed 28 10 2013].
56
Little, J. D., Sandall, H., Walegur, M. T. & Nelson, F. E., 2003. Application of Differential Global Positioning Systems to Monitor Frost Heave and Thaw Settlement in Tundra Environments. PERMAFROST AND PERIGLACIAL PROCESSES, Issue 14, pp. 349-357. Schmidt, J. & Dikau, R., n.d. [Online] Available at: file:///C:/Users/User/Downloads/Schmidt-Dikau2005.pdf [Accessed 23 November 2013]. USGS, 2013. The Water Cycle - Water Science for Schools. [Online] Available at: http://ga.water.usgs.gov/edu/watercycle.html [Accessed 25 12 2013]. Young, E. J., 2012. Geomorphological Techniques. [Online] [Accessed 20 July 2013].
57
APPENDIX
58