Hazard Modelling Resources Workshop D: Introduction to Geographic Information Systems (GIS): Hazard, vulnerability and risk analysis

Introduction

Introduction to GIS

Introduction to GIS: Ilwis - The Basics

Introduction to Geographic Information Systems (GIS): Hazard, vulnerability and risk analysis

Practical Exercise

This case study is intended to illustrate the meaning of hazard, vulnerability and risk, using a very simple data set on the national-scale of Colombia (South America). The occurrence of a disaster depends on two factors:

- Hazard: the potential for adverse consequences of some primary event, sequence of events or combinations of circumstances

- Vulnerability: the degree of loss resulting from the occurrence of the phenomenon.

You will first generate a qualitative hazard map by combining several factor maps. Then a vulnerability map is made. The hazard and the vulnerability map are combined into a risk map.

Getting started

Download and UnZip this data file Ilwis Hazard Vulnerability Risk to an appropriate directory on your n:\ drive.

  Load the ILWIS program

Change the working drive and the working directory until you are in the correct data directory.

Visualization of the data

Before you can start with the actual analysis, it is important to have an idea of the input data. The following raster maps are available:

Colombia                  The departments of Colombia.

Landslid                   Regions with severe landslide problems.

Volcanic                   Regions with volcanic hazards.

Seismic                       Regions with seismic hazard.

Tsunami                     Regions with tsunami hazard (seismically induced waves).

Inundat                      Regions with flood hazards.

Rivers                         Regions with torrential river activity.

Beach                          Regions with beach erosion and/or accumulation.

Industry                    Main industrial regions.

Infraseg                    Main infrastructure.

Concentr                  Concentration of economic activities.

Topograp                  Topographic region of Colombia..

There is only one polygon map present:

Colombia                    The departmental boundaries of Colombia.

You will first have a look at the maps before you start the analysis. 

Double-click the polygon map Colombia.

Now a Display Options dialog box is shown in which you can modify how the map will be shown.

Select the check box Boundaries Only.

Accept all default settings and press OK. The map is shown in a map window.

Drag the map window to the side of the screen.

Drag-and-drop the raster map Colombia into the map window.

The Display Options dialog box is opened. Click OK.

The raster map Colombia is now displayed, together with the boundaries of the polygon map.

Click on several units in the map to find out their meaning.

Now you will look at the maps that contain information on the different types of hazard in Colombia.

Drag-and-drop the raster map Seismic into the map window.

Accept the defaults in the Display Options dialog box and click OK.

The raster map Colombia is now replaced by the raster map Seismic.

Check the meaning of the mapping units.

Evaluate the other raster maps

( Landslid, Volcanic, Tsunami, Inundat, Rivers, Beach and Topograp)

with the same procedure.

Open the Pixel Information and drag-and-drop the raster maps Colombia, Seismic, Landslid, Volcanic, Tsunami, Inundat, Rivers, Beach and Topograp to it.

Check the combined information of the maps while you move the mouse pointer through the map window.

Also have a look at the maps used for the vulnerability analysis.

Drag-and-drop the map Seismic to the map window.

Accept the defaults in the Display Options dialog box and click OK.

Raster map Colombia is now replaced by the raster map Seismic.

Check the meaning of the mapping units.

Evaluate the other maps ( Industry, Infraseg and Concentr) by dragging and dropping them in the map window.

Check their contents.

Add these maps also to the Pixel Information window.

Check the combined information of the maps while you move the mouse pointer through the map window.

Close the map window and the pixel information window

When you have a sufficient idea of the input data you can continue with the next topic.

Creating the hazard map

In this exercise you will look at the first aspect: hazard.

In later exercises you will evaluate the vulnerability and finally the risk.

The degree of hazard of a certain area is determined by a combination of factors.

The different factors, which influence the degree of hazard, can be observed separately, although they do influence each other (e.g. in an area with high seismic hazard there will be more landslides).

The factors are provided in the form of parameter maps, each of them describing a potentially damaging phenomenon. The following phenomena are taken into account:

- Landslides

- Seismic hazard

- Tsunami hazard: earthquake induced flood waves

- Volcanic hazard

- Flood hazard

- Erosional hazard by torrential rivers

- Beach erosion and accretion

- Topographic regions

In the previous section you have seen the spatial distribution of the various hazards in Colombia.

The next step is to combine this information into one map. You can follow two approaches:

- Simply sum up all the maps, with equal weight, or

- Assign different weights to different hazard types.

Here you will follow the second approach.

The impact of the different hazards to human activity is not equal for all types of hazards.

Strong earthquakes will have a much more devastating effect on an area than landslides.

Therefore you have to assign weights to each of the classes within the individual hazard maps, taking into account their importance in producing damage.

The amount of weight that will be given to a certain factor and the way this factor will be classified is highly subjective.

This method is also referred to in the literature as blind weighting, as we do not dispose of quantitative data to decide whether the relation between high seismic hazard and high landslide hazard should be 1 or 10 or 100.

The expert’s opinion is used in defining these weights.

This has the consequence that nearly every scientist may assign different values.

Later on the weighting values will be used instead of the map-values.

In this way, the separated parameter maps of the factors involved in the hazard-analysis become weighting maps.

This exercise consists of several steps:

- Step 1: Assigning weight values to the classes of the parameter maps.

The weighting values will be assigned in tables connected to the raster maps.

You will create a table for each map and then create a column weight, in which you will edit the weight values for the different classes.

- Step 2: Renumbering the parameter maps to weight maps.

The combination of each parameter map with the weight values derived from the table created in the previous step is called renumbering.

This way you will change the maps with classes into value maps, with weight values.

- Step 3: Combining the weight maps into one single hazard map.

The weight maps will be combined in this exercise by simply summing them up.

It is also possible to define weights to the individual maps.

For example, the weights for seismic hazard can be two times as important as those for landslide hazard.

- Step 4: Classifying the combined weight map into a final hazard map.

The combined weight map, which has many classes, will be simplified by classifying the values into four classes.

Step 1: Assigning weight values using attribute tables

The weight values that should be assigned to the individual classes are given in the tables below:

Seismic 

Volcanic 

Landslid 

Tsunami 

Beach 

All classes in the map Inundat will all receive a weight of 5. Class "No inundation hazard" will receive a weight of 0.

All classes in the map River will receive a weight of 4. Class "No river" will receive a weight of 0.

All classes in the map Topograp will receive a weight of 2, and only class "Altitude less than 1000 m" a weight of 0.

Here only the procedure for the first map (Seismic) is explained. The other maps can be treated with the same procedure.

Select the menu options: File, Create, Create Table.

Enter the following filename: Seismic.

Select the domain Seismic and click on OK.

A domain in ILWIS determines the content of a map, or a table.

If we have classes, the domain is a list of the possible names of the classes.

A table which will be linked to a map should therefore have the same domain.

The table is now shown.

You see the domain items in the grey column on the left hand side.

Select: Columns, Add Column.

Enter the column name: Weight, and select the domain: Value.

Enter the values 0 and 10 for the value range, and 1.0 for the precision. Click OK.

Fill in the weights of the three classes, as indicated in the table  detailed above

Close the table.

Now you have made a table with weight values for the map Seismic.

Repeat the same procedure for the maps Landslid, Volcanic, Tsunami and Beach.

Create a table for each map using its own domain.

Before you can go on to the next step you still have to indicate that the tables should be linked to the maps.

This is done by changing the properties of a map.

Click with the right mouse button on the map Seismic and select from the context-sensitive menu the option Properties.

Now a dialog box appears in which the properties of the map are shown, such as the domain, the number of pixels, etc.

Click on the option Attribute Table and select the table Seismic.

Press OK.

Now the table Seismic is linked to the map Seismic.

Follow the same procedure for the other maps ( Landslid, Volcanic, Tsunami and Beach).

The creation of weights for the other three maps ( Inundat, River, and Topograp) will be done differently.

Since practically all units will receive the same weight it is better to use another method of reclassification: a Map calculation formula.

Step 2: Renumbering the parameter maps to weight maps

Now that you have made a table with weight values for the parameter maps, we can now renumber the maps to weight maps.

For this operation we will use the operation Attribras.

From the menu of the Main window, select the menu items: Operations, Raster operations, Attribute Map.

Select Seismic as the raster map; select Seismic as the table and Weight as the attribute.

Type Wseismic as output map.

Type: Renumbered map Seismic with weights as Description.

Click Show and OK.

In the Display Options dialog box, select the representation Pseudo.

Click OK. The weight map is displayed.

If you click on a unit in the map, you will no longer see a description, but a number, which is the weight assigned earlier.

To check this with the original map Seismic use the PixelInfo

You can do this by selecting File, Open Pixel Information.

In the pixel information window, select File, Add Map and select the map Seismic

When you move the mouse over the map, you will see the name of the original unit in map Seismic and the weight value.

Close the map window and the pixel information window.

Now you have renumbered the map Seismic into a weight map Wseismic

Repeat the same procedure for the maps:

  Landslid

Volcanic

Tsunami and

Beach

Creating weight maps called: Wlandsli, Wvolcani, Wtsunami and Wbeach.

Now another method will be used to reclassify the maps Inundat, River and Topograp.

You will directly create the weight map, using a formula, without first storing the weights in a table.

Type the following formula on the command line of the main ILWIS window:

Winundat=iff(Inundat=”No inundation hazard”,0,5)

NOTE CASE AND FILE NAME SENSITIVITY

In the Raster Map Definition dialog box, select the domain Value with values range between 0 and 5 and a precision of 1.0.

Click OK.

No calculation takes place yet.

Only the definition is stored and a raster map icon for the map Winundat is made.

The actual calculation takes place only after opening the map.

Double-click raster map Winundat.

The map is calculated, before the Display Options dialog box is opened.

Select the representation Pseudo, and click OK.

Design the formula for creating the map Wriver yourself.

All classes of the input map River will get the value of 4, except for the class “No river” which gets 0.

Also make the map Wtopogra, in which all classes of the input map   Topograp will get the value of 2, except for the class

“Altitude less than 1000 m”, which gets 0.

Make sure not to make a typing mistake in the long name, otherwise the result will be undefined.

Now you have all the weight maps ready.

Step 3: Combining the weight maps

The next step in the creation of the hazard map is the combination of the individual weight maps.

Type the following formula on the command line:

Hazard=Wseismic+Wlandsli+Wvolcani+Wbeach+Wtsunami+Wriver+Winundat+Wtopogra

Click OK in the Raster Map Definition dialog box.

Display the map Hazard. Use the Pseudo representation.

Use the pixel information window to read the values of the input maps, together with those of the map Hazard.

Check if the computer can add up values correctly.

Close the map window and the pixel information window.

Step 4: Classifying the hazard map

The hazard map which was made in the previous section has many different values. The range of values can be evaluated by calculating a histogram.

In the Main window, select the following menu items: Operations, Statistics, Histogram.

Select the map Hazard and click OK.

The histogram is calculated and the result shown in a table.

In the table window, select the menu items: Options, Show Graph.

The Graph dialog box is opened.

Select Value for the X-axis and Npix for the Y-axis and click OK.

The Edit Graph dialog box is opened.

Select: line and click OK.

The histogram will be shown on the screen.

Evaluate the values in the histogram.

On the basis of these values we can make a subdivision in hazard classes.

Close the graph and the histogram table.

From the range of values you can see that the best way to classify the map hazard is to divide it into classes of 5 units.

In the Main window, select File, Create, Create Domain.

Create a new domain Hazclas.

Select the options Class and Group. Click OK.

The Domain Editor is opened.

Select Edit, Add Item and fill in the following classes:

Close the Domain Editor.

In the Main window, select: Operations, Image Processing, Slicing.

Select the raster map Hazard and give Hazclas as the output raster map.

Use the domain Hazclas. Click OK.

Open the representation of Hazclas, and edit the colours so that “Very low” is green and “Very high” is red, with intermediate changes.

Display the map Hazclas.

Use the Pixelinfo to check the result.

Create annotation for the map Hazclass (grid, scale bar, title, legend), and save the result as a map view with the name Hazclas.

Close the map window and the pixel information window.

This ends the first exercise in the creation of a hazard map. In the next exercise you will create a vulnerability map.

Creating a vulnerability map

In this exercise you will look at the input data for a vulnerability analysis.

The final aim of the exercises at the national scale is to make a qualitative risk map, displaying the areas where there is a high probability that a disaster may occur.

For this we also need to know the vulnerability: the degree of loss to a given set of elements at risk resulting from the occurrence of the phenomenon.

Elements at risk are the population, properties, economic activities, etc. at risk within a given area A vulnerability classification can be carried out in the same way as the method used in the creation of the hazard map: assigning weights to each of the factors, and summing up the weights.

The following maps are used for this:

Infraseg                      Major infrastructures

Industry                      Main industrial centers.

Concentr                    Concentration of economic activities.

Popdens                     Population density.

The last map Popdens does not exist yet. You will first have to create it. This will be done in the first step of the analysis.

Creating a population density map

For this calculation you will need the following input data:

- The map Colombia from which you can calculate the area of each department.

- A table (below) with the number of inhabitants per department.

The map is using the following steps:

- First you will calculate the histogram of the map Colombia , and find out the area (in square meters) for each department.

- Secondly you will make a table Colombia and you join it with the histogram table to obtain the area of each department.

- Thirdly you create a new column Population in the table Colombia, in which you enter the population numbers per department.

- Then you divide the column Population by the column Area to find out the population density.

- Finally, reclassify the map Colombia with the column Popdens and create the population density map.

Calculate the histogram for the raster map Colombia.

Create a new table, Colombia, with the domain Colombia.

In the table window, read in the areas of the department by executing the following formula:

Area=Colombia.his.Area

Accept the default values.

Create a column Population (type= Value, with a range 0 to 6000000, and precision 1).

Fill in the values from the table below.

Table: Number of inhabitants per department.

After entering the values, calculate a column Popdens (population density = number of persons per km² ).

Note that the values in the column Area in m² .

Close the table window.

  Creating the vulnerability map

In the same way as the preparation of the weight maps, used for the hazard map, a series of weight maps should be created to produce a vulnerability map.

The weight values, which will be used in this exercise, are based upon the relative importance of each of the elements at risk with respect to the damage caused by a disaster.

In reality these vulnerability values will be different for different hazard types.

For example: the vulnerability of roads is less for inundations than it is for landslides or earthquakes, as during a flood the roads cannot be used temporarily, whereas during landslides or earthquakes they may be (partly) destroyed.

We have not taken this aspect into account here. The following weight values are used:

Concentr

Industry

Infraseg

Popdens

Since the calculation of the vulnerability map is very similar to the one of the hazard map, not all individual steps are explained in detail.

All steps are combined below.

Create tables for the maps Industry, Concentr, and Infraseg, and fill in the weights.

Create weight maps using the maps Industry, Concentr, and Infraseg, and the tables you have just made.

The resulting maps are called Windustr, Wconcent, and Winfras.

Create weights for population density in the table Colombia using the values in the table  previously detailed on the  page.

Create the weight map Wpopdens by reclassifying the map Colombia with the reclassified population densities.

Combine the four weight maps Windustr, Wconcent, Winfrase, and Wpopdens by adding them up.

The resulting map is called Vulnerab.

Create a class/group domain Vulclas with 4 classes (Very low vulnerability, Low vulnerability, Moderate vulnerability, High vulnerability).

Classify the map Vulnerab into 4 classes using the operation slicing

The result is called Vulclas.

Display the result and make a representation, and annotation.

At this point you have a hazard map and a vulnerability map, and now you can combine them into the final risk map.

Creating the risk map

The final stage in a risk analysis is the creation of a risk map.

According to Varnes (1984), risk can be defined as “the expected degree of loss due to particular natural phenomena”.

In a real project this would be the multiplication of costs*vulnerability*recurrence interval of natural damaging phenomenon.

We do not dispose of cost data, and neither of recurrence intervals in this simple example.

Therefore you will simplify the procedure by calculating the risk as the combination of natural hazard and the vulnerability.

You will make a qualitative risk map, giving the general relation between hazard and vulnerability.

The combination will be done using a two-dimensional table.

In the Main window, select File, Create, Create 2 Dimensional Table.

Enter the Table Name Risk.

Select Hazclas as the Primary domain, and Vulclas as the Secondary domain.

Press the Create button next to the Domain list box.

The Create Domain dialog box is opened.

Create a class domain: Risk, with three classes (Low risk, Moderate risk, and High risk).

Close the domain editor, and click OK in the Create 2 Dimensional Table dialog box.

The 2 dimensional table is opened. It will contain undefined values for all combinations.

Below we have already filled in some of them, for example:

When the hazard is very low, it doesn’t matter whether the vulnerability is low or high; the risk will be low in all cases.

When the vulnerability is very low (meaning that the area doesn’t contain any important elements at risk), the risk is always low.

Vulnerability

Fill in the missing classes in the two dimensional table above.

Give the reason for assigning the risk classes high, medium or low risk.

Click the upper left field in the 2-dimensional table.

Edit the fields using the left arrow.

Press the down arrow to go to the next field.

Fill in the entire table. It is faster to work with the arrow key, than to click each field, select the right name from the list box, and click another field.

You only have to click the upper field of the next column when you are finished with the last field of the previous column.

When you are finished, close the table window.

The two-dimensional table is now complete. We can use it in a calculation on the command line. The calculation formula should have the following syntax:

Output map = Two-dim table [Map1, Map2]

Type the following formula on the command line of the main window:

Risk=Risk[Hazclas,Vulclas]

Press Enter. The raster Map Definition dialog box is opened. Click OK.

Open the map Risk and check its contents.

Close the map window.

Create annotation for the map  Risk.

The creation of a risk map is not the final step in a disaster management.

It is merely a tool which has to be used by decision makers in the planning process.

One of the politically important aspects is the percentage of each department, with high, medium or low risk.

These percentages will be calculated in this section.

Cross the maps Colombia and Risk and create a cross table.

Open the cross table and type the following formula on the command line:

High:=iff(Risk=“High”,Npix,0)

Also calculate the columns Medium and Low in the same way.

Close the cross table and open the table Colombia.

Use join operation to read in the maximum value of the columns High, Medium and Low of the cross table.

Use table join again to obtain the sum of the column Npix from the cross table.

Calculate the percentages of each department with high, medium and low risk.

Find the five most endangered departments of Colombia (the ones with highest percentage of their area classified as high risk).

 Feedback

www.itc.nl/ilwis/applications/application01.asp

References

Romero, J.A., Florez, A. and Sanchez, H.A. (1989). Inventario inicial de riesgos Naturales en Colombia. Análisis Geográficos No.16. Instituto Geográfico Augustín Codazzi, Bogotá, Colombia.56 pp.

Varnes, D.J. (1984). Landslide Hazard Zonation: a review of principles and practice. Commission on landslides of the IAEG, UNESCO. Natural Hazards No. 3, 61 pp.

Van Westen, C.J., Van Duren, I, Kruse, H.M.G. and Terlien, M.T.J. (1993). GISSIZ: training package for Geographic Information Systems in Slope Instability Zonation. ITC-Publication Number 15, ITC, Enschede, The Netherlands.Volume 1: Theory, 245 pp. Volume 2: Exercises, 359 pp. 10 diskettes.

Velasquez, A. and Meyer, H.J. (1990). Un ensayo de evaluación de las amenazas, de los riesgos y de los desastres en Colombia. Proceedings 1er Seminario Andino de Geología Ambiental, 1er Conferencia Colombiana de Geología Ambiental, 3er Conferencia de Riesgos Geológicos del Valle de Aburra. AGID Report No.13, paper 42, pp 547-580.

  If you wish to download this software for use at home then go to

http://52north.org/downloads/ilwis

DPG Version 3 December 2011