Application of Remote Sensing and GIS for Natural Hazards - Disaster Assessment and Management Landslides
Application of Remote Sensing and GIS for Natural Hazards - Disaster Assessment and Management Landslides

Introduction

Landslides are a sudden, short-lived geomorphic event that involves a rapid-to-slow descent of soil or rock in sloping terrains. They occur worldwide, often in conjunction with natural hazards like earthquakes, floods, or volcanic eruptions. Landslides can also be caused by excessive precipitation or human activities, such as deforestation or development that disturb natural slope stability.

Landslides in the United States alone cause $1 to $2 billion in property damage and over 25 fatalities per year. Posing threats to settlements and structures, landslides often result in catastrophic damage to highways, railways, waterways, and pipelines. According to the U.S. Transportation Research Board, annual costs for the repair of minor slope failures by state departments of transportation exceed $100 million.

To determine where protective measures are necessary, scientists and technicians produce landslide inventory and risk assessment maps for many areas around the world Landslides unfortunately, do not display a clear relationship between magnitude and frequency as do earthquakes and floods. Landslide studies are challenging to scientists, due to the difficulty to represent landslide hazards in quantitative terms over large areas.

Satellite Images - Landslide Lake, Tibet

Landslide Lake, Tibet (ASTER - DEM)

Credit: NASA/GSFC/METI/ERSDAC/JAROS,and U.S./Japan ASTER Science Team

Remote Sensing for Landslide Location and Causes

Remote sensing techniques greatly aid in the investigations of landslides, on both a local and regional scale. Remote sensing offers an additional tool from which we can extract information about landslide causes and occurrences. Most importantly, they greatly aid in the prediction of future landslide occurrences, which is very important to those who reside in areas surrounded by unstable slopes. Satellite Imaging Corporation (SIC) offers satellite imagery from Stereo IKONOS, SPOT-5, ASTER, LiDAR, and SAR, depending on your project needs, location, and terrain condition.

Landslide studies can be organized into three phases:

1. Detection and classification

2. Monitoring activity of existing landslides

3. Analysis and prediction of slope failures in spatial distribution and temporal distribution

Remote sensing techniques can be and are often used in all three stages of a landslide investigation and monitoring.

ASTER Satellite Image - Landslide, Dominican Republic

Landslide in Dominican Republic (ASTER)

Credit: NASA/GSFC/METI/ERSDAC/JAROS,and U.S./Japan ASTER Science Team

Detection and Classification of Landslides

To detect and classify the landslide, you need to be able to view the size and contrast of the landslide features and the morphological expression of the topography within and around the landslide. Interests in determining parameters are the type of movement that has occurred, the degree of present activity of the landslide, and the depth to which movement has occurred. The most common remote sensing tools used for the detection and classification of landslides are satellite imagery and aerial photography.

Monitoring Landslide Movement

Monitoring landslide movement involves the comparison of landslide conditions over time, including the aerial extent of the landslide, the speed of movement, and the change in the surface topography. Satellite imagery and aerial photography are commonly used in this stage of a landslide investigation.

Landslide Hazard Analysis Mapping

Landslide hazard maps typically aim to predict where failures are likely to occur without any clear indication of when they are likely to occur. They are useful for providing landslide hazard information needed for planning and protection purposes.

Satellite Image - Landslides, Pakistan

Landslides in Pakistan (ASTER)

Credit: NASA/GSFC/METI/ERSDAC/JAROS,and U.S./Japan ASTER Science Team

Analysis and Prediction of Landslides in GIS (Geographic Information Systems)

A large database is necessary for the analysis and prediction of slope failures. It needs to be able to store, manipulate, and apply the data collected in first two stages which are recognition and monitoring. A Geographical Information System is ideal for this stage in a landslide investigation because it is capable of handling large amounts of past, present and future data and integrating this data with predictions. It is capable of data storage, visualization and manipulation of the environment within the application it can also has regional databases perform both local and regional modeling.

Geographic Information System - GIS Overview

Most landslide potential models determine terrain instability by combining slope maps with soils data, then selecting from the resultant soil/slope categories the combinations that are rated for severe erosion potential. Vegetative cover considerations extend the model. When available, maps showing historic landslide sites are added for both establishing and testing landslide potential.

Available and Desired Base Data (Listed in Order of Importance)

  • Terrain Data (DEM derived from elevation contours)
    Landform data to include derived factors of slope and flow accumulation are required for the landslide potential model; aspect and roughness
  • Additional Environmental Data
    Edaphic data to include soil type, depth to bedrock, and parent material
  • Orthorectified Mosaic Stereo IKONOS Satellite Image or Aerial Photo
    Vegetation data to include derived factors of vegetation type and vegetation density
    Disturbance data to include the locations of recent terrain and cover modifications
  • Other Information
    Hydrography data to include streams would greatly strengthen the model
    Climate data to include rainfall, snowfall, and mean temperature
  • Culture
    Historical landslide data to include location and frequency of occurrence would greatly strengthen the model as a means for evaluating model performance

GIS Modeling Approach/Logic

SHALSTAB model which is used frequently by most GIS specialists and Scientists combines slope steepness with flow accumulation to classify risk of slope failure. Standard soil parameters are employed to redefine the model as the mechanical properties of soils can profoundly affect slope stability. However, much available soil data is at very small scale (1:240,000) and only useful for regional analysis and if available at larger scale requires expert interpretation for regional variants.

An extended SHALSTAB model is proposed depending on availability of base and derived data. Promising extensions include by data availability:

Terrain Factors

  • Aspect is computed from DEM data (derived as first derivative of elevation at minimal effort) and depending on regional climate can marginally effect slow stability with south facing slopes susceptible
  • Terrain Roughness is easily computed from DEM data (derived as coefficient of variations of surrounding slope conditions at minimal effort) and depending on regional climate can marginally effect slope stability with "smoother" sloped terrain more susceptible.

Vegetation Factors

  • Vegetation Type by broad categories of forest, non-forest, bare, water land cover can be derived (multispectral classification at moderate effort) with bare vegetation being more susceptible
  • Vegetation Density by broad categories of dense, moderate and sparse land cover can be derived (multispectral classification at minimal additional effort beyond vegetation type classification) with sparse being more susceptible

Disturbance Factors

  • Terrain/Cover Modifications such as forest harvesting, wildfire, mining and other activities can be derived (multispectral classification at minimal additional effort beyond vegetation type classification) with disturbed areas more susceptible

Hydrological Factors

  • Headwaters of streams (if available as mapped data or could be derived through manual image interpretation at considerable effort) in steep terrain can be derived with uphill locations surrounding headwaters more susceptible

Climate Factors

  • Precipitation by broad categories of wet, moderate and dry conditions (if available as mapped data or could be derived from weather station records at moderate effort) with wet conditions more susceptible
  • Temperature by broad categories of hot temperature and cool conditions (if available as mapped data or could be derived from weather station records at moderate effort) with temperate conditions more susceptible

Historical Factors

  • Location/Frequency of previous landslides in the project area (if available as mapped data or could be derived through manual image interpretation at considerable effort) for use evaluating model performance
  • SHALSTAB landslide potential model works best with detailed elevation data. A Digital Elevation Model (DEM) spatial resolution of 10 meters is desired. Raster-to-Vector procedures are used to convert base maps to project grid and raster-to-vector procedures are used to output a model.

Landslide Diagram

Natural Resources Canada

The basic SHALSTAB model primarily involves point-by-point processing with its results dependent on the direct coincidence at each map location. Several of the proposed extensions such as terrain roughness and headwater proximity, address "contextual relationships" of surrounding conditions.

A nine level landslide potential index is proposed:

  • 0 = not susceptible (lakes)
  • 1 = low potential
  • 2-3 = minimal
  • 4-5 = moderate
  • 6-7 = high
  • 8-9 = extreme

Salt and pepper smoothing is proposed to eliminate individual outliers of varying classification. Locations of 6 or higher will be isolated to generate a map of high landslide potential and can be converted to a vector map.

Edge match procedures will be used to insure continuous coverage of the final map and might require tiling into manageable processing units depending on the project size and shape.

Source: http://www.satimagingcorp.com