Accessibility Modelling in ArcView 3

An extension for computing travel time and

market catchment information

By

Andrew Farrow and Andy Nelson

2001
Accessibility Modelling in ArcView 3 CIAT

Contents Acknowledgments
Acknowledgements
• The data used in the examples were taken from the CIAT
Installation Honduras database. More details can be found here.
http://www.ciat.cgiar.org/scale/data

Introduction
• The Technical Notes section has been adapted from the
Description documentation provided with ArcView 3.2 .

Example • Several parts of the Accessibility Analyst have been based on

scripts found at the ESRI ‘Arcscripts’ website
http://www.esri.com/arcscripts
References

ArcView Technical Notes Script Author

Cost Distance Grid Tools ESRI
Graphics.Graphics2Shapes [email protected]
Oregon Department of Forestry
Make Extension Tools ESRI
Clipping Grids Tom Van Niel

• CIAT donors generously provided the funding for the

development of this software. In particular we would like to
mention

The Ecoregional Trust Fund for Methodological Development

The World Bank
UNEP

• The following people at CIAT have contributed to the ideas that

lead to this software and manual.

-1-
Accessibility Modelling in ArcView 3 CIAT

Andy Farrow, Andy Nelson, Manuel Winograd, Jeremy Eade,

Gregoire Leclerc, Ron Knapp and William Bell.

Installation under Windows Installation under UNIX

This extension requires This extension requires

Arcview 3 and the Spatial Analyst extension Arcview 3 and the Spatial Analyst extension

1 Unzip the file CIAT_access.zip. 1 Uncompress CIAT_access.Z.

2 Place ciat_access.avx in the extension directory. $> uncompress CIAT_access.Z

2 Place ciat_access.avx in the extension directory.

Assuming that your system has the environment variable
$AVHOME.

$> cp access.avx $AVHOME/ext

You may need to be superuser to do this.

Alternatively, if your system has the environment variable

$USER_EXT, you can place ciat_access.avx in that
directory.
You may need administrator privileges to do this.
3 Place the ciat_access directory anywhere you like.
3 Place the ciat_access directory anywhere you like. It contains the manual and the example datasets.
It contains the manual and the example datasets.
4 Start ArcView and the extension will be available in the
4 Start ArcView and the extension will be available in the Extensions dialogue as Access Analyst (Ver. 2.0)
Extensions dialogue as Access Analyst (Ver. 2.0)

-2-
Accessibility Modelling in ArcView 3 CIAT

the farmer largely inhabits a walking world. In his introduction to

Transport and the Village, Cleaver states that:

Introduction It is clear that the extremely poor state of off the road transport
system in Africa acts as a powerful brake on agricultural
Accessibility productivity and growth. Improved accessibility will reduce the
economic costs of moving goods from local markets and ease
Transportation is a critical function for an economy as it affects the the barriers to social facilities. This will contribute to economic
movement of people, goods and services, and development. In most growth and enhance social well-being.
countries, a complex transportation network permeates the
countryside with paths for walking, paths for horses and donkeys, But before national governments can provide environments
single lane seasonal roads, one and two lane all-weather gravel conducive to the development of local infrastructure, there is a need
roads, local surfaced roads, regional and national surfaced roads, for a model that can incorporate the local environmental factors that
and motorways or freeways with dividers separating the define rural inaccessibility. Also any model should be flexible
carriageways. Each of these has typical speed of travel associated enough not only to quantify time and cost of travel but also
with it, which can be adjusted to take into account slope, the type of infrastructure improvement (or degradation) scenarios to be created
vehicle traversing it, rainfall over a 2, 3 or 7 day period, or an and evaluated.
adjustment for congestion. In the end it is possible to work out how
long it takes to get from anywhere on the map to the nearest market Clearly there are different levels of infrastructure availability,
town or hospital or school using a particular vehicle class. The time transport availability, and many different environmental effects, and
taken to reach a desired location can be generally referred to as its economic considerations. For example it would be impossible to
‘accessibility’. apply the same model of rural accessibility to Sub Saharan West
Africa and Central American Hillsides, but any model should be
Accessibility can be defined as the ability for interaction or contact capable of incorporating all of these factors in a way that is not only
with sites of economic or social opportunity, (Deichmann, 1997), sensible and geographically sensitive but also educational and
however there are a multitude of ways in which this intuitive concept explanatory.
has been expressed in the literature. Goodall (1987) defines
accessibility as the ease with which a location may be reached from
other locations, and Geertman (1995) states that the concept of
Background
accessibility can be used in rural development policy as an indicator Spatial accessibility is determined by geographic location in relation
or rural deprivation and as a variable on location analysis to target locations, and by the transportation facilities that are
available to reach those destinations. Accessibility is also influenced
by social factors such as knowledge and information and by
Over the last ten years there has been a growing awareness that
rural accessibility concerns more than just roads, (Barwell, 1996). economic factors since the use of transport and communication
Reports and surveys carried out in developing regions paint a picture facilities is usually associated with some monetary cost. Although it
of rural isolation and unproductive use of limited resources, in which is difficult to incorporate these economic factors explicitly in spatial
models, the selective weighting of one market can imply the social
and economic dimensions over another.

-3-
Accessibility Modelling in ArcView 3 CIAT

In public planning, an important objective is the equal provision of

services for all people and in all parts of a country. For example, the
population of a country or region should have similar access to public Description
services, independent from their location of residence. Poor
infrastructure and consequently limited accessibility implies fewer To begin with, it is necessary to clarify some of the terminology that
opportunities for improving economic status, health or social will be used in this model. Accessibility is calculated on a friction
standing. Ravallion (1996), for example, speaks of a “spatial poverty surface. A friction surface consists of a regular two-dimensional grid
trap” that may prevent the poor from breaking out of local level where each cell in the grid represents either a transport route such
constraints. as roads, railway lines, tracks, or navigable rivers, or relatively
inaccessible land and water bodies.
Operational quantitative measures of accessibility are therefore
useful in policy analysis where the focus is on service provision, for Different types of transport infrastructure have different
example, in the health sector. Other areas in which accessibility is characteristics. A surfaced road, for example, allows faster travel
important are economic applications, and, in fact, many of the speed than a dirt road. In practice, it is therefore not enough to
operational measures of accessibility have first been developed for measure the distance of a road connection between two points.
labour market research and retail analysis. Instead, a measure of travel cost is preferable. This cost can be
measured in monetary terms or as travel time. As will be discussed
Geographical information systems (GIS) lend themselves naturally to later in this section the cost to travel across different types of land
the computation of accessibility indicators. GIS can represent can be easily estimated.
networks, villages or facilities and provide functions to compute
distances and to define relationships among spatial objects. Targets (places of interest) such as villages, hospitals, or national
Consequently, certain accessibility measures can be computed using parks are usually located on the transport network and can therefore
packages such as IDRISI, GRASS and ArcView. be represented in another grid as cells that have a certain
characteristic. This characteristic could be the village’s population,
However, there is a distinct lack of any step-by-step, or easy to use the hospital’s number of doctors or beds, or the number of teachers
models to develop transport maps, accessibility surfaces and the at a school. If we are simply interested in is the presence or
economic catchment areas of towns and markets. This is a strange absence of a place of interest, then all cells can have the same
omission from the planner's toolbox, when accessibility is considered value.
such a fundamental indicator of economic potential and an
undeniable pressure on ecological systems. One possible All that is needed to create an accessibility map are:
explanation is that accessibility is so poorly defined; many people
talk about accessibility and there are many differing opinions as to • A coverage of markets or places of interest
what it actually is. The following section describes a simple but
flexible GIS method and tool for deriving verifiable accessibility • A grid where each cells value represents the cost of
models. traversing that particular cell.

-4-
Accessibility Modelling in ArcView 3 CIAT

These two datasets are input to a COST-DISTANCE model, which is

discussed in more detail in the Technical Notes section.

So far the problem seems trivial, the cost-distance function is simple The ArcView extension addresses all of these issues, with a simple
in concept and the data requirements are minimal. In fact these are definition or concept of accessibility.
two of the reasons for selecting this type of analysis. However there
are several points to note. What is the cost of getting from any location
to the nearest place of interest?
1. You need geographic data that is in an Equal Area
Projection (for example Lambert Azimuthal), to preserve This is a fairly general question, where cost can be:
areal qualities.

This is a key point. The ArcView GIS and ArcInfo technical • Real cost (cost of transport)
notes, state that any data that are to be used as inputs for
the cost-distance algorithm should be in an equal area • Perceived cost (cost of transport with other factors)
projection. For very small regions or areas near the
Equator, this may not be so critical. • Time (simply time taken to get there)

2. You are unlikely to have a friction surface ready to hand. • Perceived time (travel time plus factors introduced by the
nature of the transport)
3. The friction surface is totally dependent on the user and
purpose, and will possibly change for differing scenarios; • Any other concept of cost that the user is able to express
i.e. foot based/vehicle based, wet season/dry season.
A place of interest can be:
4. The pre-processing required to create the equal area
projected grids. • A market where goods can be bought or sold (large towns)

With that in mind, an interface was developed at CIAT to perform this • A service industry (so called enterprise zones)
type of spatial analysis quickly and intuitively.
• A hospital or school
ArcView GIS was chosen as the working environment for this
activity. Previous work in CIAT had used ArcInfo and its associated • A national park or any other area under threat by human
macro language AML to produce accessibility surfaces, but by development
version 3, ArcView had become mature enough to allow
development of the interface using its Avenue scripting language. The location can be anywhere in the area of study.

-5-
Accessibility Modelling in ArcView 3 CIAT

The following section is a step by step introduction to the ArcView Open a new view
extension.

Example
Load the Access Analyst Extension

The Accessibility menu will be added to the View menu bar. There
are two options in the menu, English or Español. Select your
preferred language to start the analysis. The Accessibility Dialogue
will appear, asking for a directory in which to place all temporary
grids and coverages. You will not be able to move on to the next
stage until you have entered a valid directory. Once a directory has
In the extension dialogue, select Access Analyst from the list. This been entered, click Next.
will load both the Access Analyst and Spatial Analyst extensions.

-6-
Accessibility Modelling in ArcView 3 CIAT

Add friction themes to the view

There are five options available,

A new View called Friction Surface Themes will be created. At least • Rivers
two datasets are needed for creating the friction theme; a transport • Slope
or road cover, and a cover that defines the limit of the analysis. • Land Cover
These two must be loaded before any optional friction components • Urban Areas
can be included. Clicking the buttons will open the Add Theme pop- • Barriers
up window and you will be prompted to add the themes.
Depending on the analysis, you may want to include one or more of
The text alongside each button will confirm that the theme has been these themes as a contributing factor to the friction surface. In this
loaded. You can always change the themes by clicking the example we will add Rivers, Slope and Land Cover.
respective theme button to delete the current choice and selecting a
new one. Once these two themes have been added, you can either When a theme is loaded, a new x button appears alongside, which if
click Next to proceed or you can add optional themes by clicking the clicked will delete the respective theme allowing you to replace it with
check box. In this example, we will add more friction themes. another choice.

-7-
Accessibility Modelling in ArcView 3 CIAT

Once all the optional fiction themes have been added, click the Add There are three options for making sub selections on the point
Optional Themes check box to hide the optional themes, and then dataset.
click Next. Use the Back button if you want to go back to a previous
stage at any time in the analysis • Select from Table
• Select from Query
Add a source or target theme to the view • Select from the Map
The final necessary theme is the source or target point theme, i.e. a
dataset that contains all the points to calculate accessibility to. As Here we will make a sub selection based on Table of the dataset.
before, click the Add Target Theme button to load a source/target Clicking, Select From Table brings up the table, record can be
dataset. Once loaded, the Make a Selection check box becomes selected which will be highlighted (usually in yellow) in the view. In
visible. If you do not wish to alter the points that you have loaded this case we select two points out of the sixteen.
you can just click Next to proceed, but if you wish to make a
selection, click the check box. Once a selection has been made, click the check box once again to
hide the options and click Next.

-8-
Accessibility Modelling in ArcView 3 CIAT

Clip all themes to a common spatial extent All the clipped themes will be transferred to a new View called
By now there are at least three themes in the view and perhaps as Clipped Themes. Now click Next to proceed to the next stage.
many as eight. In the event that these themes do not share a
common spatial extent, it is necessary to clip them.

A pop-up window will explain that a red box will appear in the view
Convert all themes to grids
The cost-distance algorithm will only work with grid datasets, so this
that can be interactively resized should you wish to clip the themes.
stage converts all the datasets to grids. A pop-up window will ask for
Click OK and the red box will appear. You can resize the box until
the analysis properties to be set, the most important property being
you have focused on your area of interest. When ready, click Clip
the cell-size. In this case we can set the cell-size to be equal to the
Themes. If some of your themes are complex or big, this can take
slope grid (100m), which is also an easy number to use for the
some time, this is a good time for a coffee break.
calculations that have to be made in the latter stages. When the cell-
size has been set, click OK in the pop-up to start the grid conversion
If you do not wish to clip the themes, leave the box in its initial
process. All the grids will appear in a new View called Gridded
position and simply click Clip Themes.
Themes.

-9-
Accessibility Modelling in ArcView 3 CIAT

For each grid, you will be asked to select the item from the table to Reclass the grids to reflect their friction value
use for the grid values. Select whichever value is most appropriate, In this stage all the grids must be reclassed so that the value of each
and select No, when asked to join attributes to the grid. cell represents the cost or time required to traverse the cell. For
example with a grid consisting of roads, we might estimate that a car
When all the themes have been converted, click Next. can travel at 60km/hr. so all road cells in a 100m grid would be given
the value 6

60 km/ hr = 60,000 m/hr = 1,000 m/min so 100m takes 6 seconds.

Or for land cover, average walking speed is estimated at 4km/hr, but

is reduced to 3km/hr for forest, giving values of 90 and 120
respectively.

4 km/ hr = 4,000 m/hr = 67 m/min so 100m takes 90 seconds.

3 km/ hr = 3,000 m/hr = 50 m/min so 100m takes 120 seconds.

-10-
Accessibility Modelling in ArcView 3 CIAT

In this case these values have to be calculated for each

• road type
• river type
• land cover type (background value)
• boundary (this value is not important)

In this case we have small cell sizes so the cost will be calculated in
seconds, but it could also be minutes or hours. So the general
conversion factor for turning km/hr into seconds is.

1
Time = Cell size x
1000
Speed(km/hr) x
3600

Slope is a special case where we need to calculate a slope factor,

i.e. how much does slope affect the speed of travel over the other
surfaces. For example slopes between 0 and 5 degrees may not
have any influence so the factor would be 1, but slopes between 5 Combine the friction grids into one
and 10 degrees would slow travel by half so the factor would be 2, The cost-distance algorithm requires only two inputs, a friction theme
etc. There is no accounting for slope direction, it is assumed that and a source theme. In this stage, all the friction components
travelling both up-slope and down-slope incurs a reduction in travel (roads, rivers, boundary, etc.) are merged to create one grid. Click
speed. the Combine button to do this. When the Friction grid appears, click
Next to continue
Click on the Reclass button to start and use the pop-up windows to
reclass each grid. All the grids will appear in a new View called The grids are merged using the following logic, cell-by-cell.
Reclassed Themes. If you wish to change the classification of any of
the grids, click Reclass again to repeat the process. When all the
themes have been reclassed, click Next.

-11-
Accessibility Modelling in ArcView 3 CIAT

for each in_cell Run the Cost-Distance function

{ Clicking the cost-distance will start the algorithm. Three grids will be
if in_cell = barrier created.
out_cell = barrier x slope
else if in_cell = road • allocation grid - Indicating the catchment area of each target
out_cell = road x slope
• direction grid - Indicating the direction of travel through each
else if in_cell = river
cell
out_cell = river
else if in_cell = urban area • time to market grid - Indicating the cost of travel from each cell
out_cell = urban area x slope to the nearest target (target defined by the allocation grid)
else if in_cell = land cover
out_cell = land cover x slope
}

Here we are showing the time to market grid that was calculated in
seconds but has been reclassed into minutes for clarity. We have
also loaded the two targets into the view. Click Next to continue.

-12-
Accessibility Modelling in ArcView 3 CIAT

Convert outputs to Shapefiles

The final stage allows the user to convert the three outputs to
shapefiles. If the study area is large or the time to market surface is
complex, this may take a long time. Optionally you might want to
reclass the grids (seconds to minutes or minutes to hours for
example) and then convert them to shapefiles.

A composite map of time to markets, the transport network and all

populated places in the vicinity of Yorito and Sulaco in Honduras.

-13-
Accessibility Modelling in ArcView 3 CIAT

Bibliography Ritsema van Eck J.R. and de Jong T. (1999), Accessibility analysis and spatial
competition effects in the context of GIS-supported service location planning. In:
Computers, Environment and Urban Systems, volume 23, pp. 75-89.
Archondo-Callao R. and Faiz A. (1994), Estimating vehicle-operating costs. World
Bank technical paper No. 234. pp 82. Teeffelen van P. Barkhof M. and de Jong T. (2000), Applying a GIS for integrated
accessibility planning in Laos using a PC based GIS -FLOWMAP- as a tool to plan
Barwell I. (1996), Transport and the village. Findings from Africa-level travel and facilities. Fifth seminar on GIS and developing countries GISDECO, International Rice
transport surveys and related studies. World Bank Discussion Paper No. 344. pp 66. Research Institute, Laguna, Philippines.

Chesher A. and Harrison R. (1987), Vehicle Operating Costs. Evidence from Winograd, M. Farrow, A. and Eade J. (1997), Environmental and sustainability
developing countries. The Highway design and maintenance standards series. World indicators outlook for Latin America and the Caribbean. CD-ROM with manual.
Bank Publication. pp 374. UNEP-CIAT, Cali, Colombia. pp 27

Deichmann U. (1997a), Accessibility and Spatial Equity in the Analysis of Service Winograd M. Fernandez N. and Farrow, A. (1998), Herramientas para la toma de
Provision, Workshop on Geographical Targeting for Poverty Reduction and Rural decisiones en AmericaLatina y el Caribe: indicadores ambientales y sistemas de
Development, The World Bank, Washington, D.C., November 11, 1997. informacion geograficos/ Tools for Making decisions in Latin America and the
Caribbean: environmental indicators and Geographical Information Systems. CIAT
Deichmann U. (1997b), Accessibility Indicators in GIS, United Nations Statistics and the United Nations Programa de Naciones Unidas para el Medio Ambiente
Division, Department for Economic and Policy Analysis, New York. (PNUMA).

Deichmann U. and Bigman D. (2000), Spatial indicators of access and fairness for Wolff de T. Kruska E Ouma E. Thornton P. and Thorpe W. (2000), Improving GIS
the location of public facilities. In Bigman, David and Hippolyte Fofack (Editors) derived measures of farm market access: An application to mil markets in the East
Geographic targeting for poverty alleviation, methodology and application. World African highlands. Fifth seminar on GIS and developing countries GISDECO,
Bank Regional and Sectoral Studies, Washington DC. International Rice Research Institute, Laguna, Philippines.

Eade J. (1997), The accessibility surface. Electronic Mail from

[email protected], to [email protected] (Accessed Feb 21, 1997).
CIAT, Cali, Colombia.

Geertman S.C.M. and Reitsema van Eck J.R. (1995), GIS and models of
accessibility potential: An application in planning. International Journal of
Geographical Information Systems 9, 67.

Dixon-Fyle K. (1998), Accessibility planning and Local Development. International

Labour Organisation. Technical report.

Kammeier H.D. (1999), New tools for spatial analysis and planning as components of
an incremental planning-support system. In: Environment and Planning B, vol 26, pp.
365-380.

Leinbach T. (1995), Transportation and Third World Development: Review, Issues,

and Prescription. Transportation Research. Vol.29A, No.5, pp. 337-344.

Ravallion M. (1997), Good and Bad Growth: The Human Development Reports
World Development, Vol. 25, No. 5, pp. 631-638.

-14-
Accessibility Modelling in ArcView 3 CIAT

ArcView Technical Notes Cost Distance Modelling Discussion

These functions are similar to Euclidean distance functions, but
aGrid.CostDistance(costGrid,direction,allocation,maxDistance) instead of calculating the actual distance from one point to another,
they determine the shortest cost distance (or accumulated travel
A Spatial Analyst function that calculates for each cell the least- cost) from each cell to the nearest cell in the set of source cells. A
accumulative cost-distance over a cost surface, costGrid, to cells second exception is that cost distance functions apply distance not in
with values other than No Data in aGrid, optionally producing geographic units but in cost units.
direction and allocation grid data sets (direction and allocation).
All cost distance functions require a source Grid and a cost Grid. A
direction can be a FileName or Nil, and defines the FileName for the source Grid can contain single or multiple zones, which may or may
optional cost direction grid data set. allocation can be a FileName or not be connected. All cells that have a value (including 0) are
Nil, and defines the FileName for the optional cost allocation grid processed as source cells. All non-source cells need to be assigned
data set. If either direction or allocation is set to Nil the optional No Data on the source Grid.
output grid data set for the argument that is set to Nil is not
calculated. A cost Grid assigns impedance in some uniform-unit measurement
system that depicts the cost involved in moving through any
maxDistance can be a Number or Nil, and defines the threshold that particular cell. The value of each cell in the cost Grid is assumed to
the accumulative cost values cannot exceed. If a cell exceeds the represent the cost-per-unit distance of passing through the cell,
value it is given the value of No Data. If maxDistance is set to Nil, where a unit distance corresponds to the cell width. These costs may
then the value of infinity is used. be travel time, dollars, preference and so forth.

This request will trigger the evaluation of aGrid, if aGrid has not yet
been evaluated. For more information on the evaluation of Grid
objects, see the Discussion topic for the Grid class.

This request is only available with the Spatial Analyst.

-15-
Accessibility Modelling in ArcView 3 CIAT

The Cost Calculations

The aGrid.CostDistance request creates an output Grid in which
each cell is assigned the accumulative cost to the closest source
cell. The algorithm utilizes the node/link cell representation. In the
node/link representation, each centre of a cell is considered a node
and each node is connected by links to its adjacent nodes.

Every link has impedance associated with it. The impedance is

derived from the costs associated with the cells at each end of the
link (from the cost surface) and from the direction of movement. If
moving from a cell to one of its four directly connected neighbours,
the cost to move across the links to the neighbouring node is 1 times If the movement is diagonal, the cost to travel over the link is
the cost of cell 1 plus the cost of cell 2 divided by 2. 1.414216 (or the square root of 2), times the cost of cell 1 plus the
cost of cell 2 divided by 2.
a1 = cost1 + cost2 / 2
a1 = 1.414216(cost1 + cost2) / 2
where cost1 is the cost of cell 1, cost2 is the cost of cell 2 and a1 is
the length of the link from cell 1 to cell 2. But when determining the accumulative cost for diagonal movement
the following formula must be used.

accum_cost = a1 + 1.414216(cost2 + cost3) / 2

The accumulative cost is determined by the following formula.

accum_cost = a1 + (cost2 + cost3) / 2

where cost2 is the cost of cell 2, cost3 is the cost of cell 3 and
accum_cost is the accumulative cost to move into cell 3 and cell 1.

-16-
Accessibility Modelling in ArcView 3 CIAT

The Algorithm The accumulative cost values are arranged in a list from the lowest
Creating an accumulative cost-distance Grid using graph theory can accumulative cost to the highest.
be viewed as an attempt to identify the lowest cost cell and adding it
to an output list. It is an iterative process that begins with the source
cells. The goal of each cell is to be assigned quickly to the output
cost-distance Grid.

The lowest cost cell is chosen from the active cost list and the value
for that cell location is assigned to the output cost-distance Grid. The
list of active cells is now expanded to include the neighbours of the
chosen cell, because those cells now have a way to reach a source
Only those cells that can possibly reach a source can be active in the
list. The cost to move into these cells is calculated using the
accumulative cost formulas.

In the first iteration, the source cells are identified and assigned to
zero since there is no accumulative cost to return to themselves.
Next, all the source cell’s neighbours are activated and a cost is
assigned to the links between the source cells nodes and the
neighbourhood cell’s nodes using the above accumulative cost
formulas. Each of these neighbourhood cells can now reach a
source, consequently, they can be chosen or assigned to the output
accumulative cost Grid. To be assigned to the output Grid, a cell
must have the next least-cost path to a source.

-17-
Accessibility Modelling in ArcView 3 CIAT

Again, the active cell on the list with the lowest cost is chosen, the This updating can occur with the advent of new paths for cells on the
neighbourhood is expanded, the new costs are calculated and these active list as more cells are allocated to the output Grid. When the
new cost cells are added to the active list cell with the lowest value on the active accumulative cost list is
allocated to the output Grid, all the accumulative costs are
Source cells do not have to be connected. All disconnected sources calculated. These costs are also calculated for the neighbouring cells
contribute equally to the active list. Only the cell with the lowest of the newly assigned output cell, even if the neighbouring cells are
accumulative cost is chosen and expanded, regardless of the source on the active list through another cell. If the new accumulative cost
to which it will be allocated. for the locations on the active list is greater than the one that the
cells currently has, the value is ignored. If the accumulative cost is
less, then the old accumulative cost for the location is replaced on
the active list with the new value. That cell, which has discovered a
cheaper and more desirable path to a source, then moves up on the
active chosen list. In the example below, the cell location at row 3,
column 1 (highlighted by the box) had an accumulative cost of 11.0
when it was put on the active list to reach the source at the top of the
Grid. But because the lower source expanded to this location, the
cell had access to a cheaper accumulative cost path to reach a
source. The value for the location was updated on the active list and
allocated to the output earlier, because of this lower accumulative
cost.

Furthermore, cells on the active list are updated if a new, cheaper

route is created by the addition of new cell locations to the output
Grid.

If there are multiple zones or disconnected sets of source cells on

the input source Grid, the growing process continues and allocates
the cheapest cost cell from the active list, regardless of which source

-18-
Accessibility Modelling in ArcView 3 CIAT

it is from. When the growth fronts meet, the least cost path back to
the source proceeds until all eligible cells have received a cost value.

When all cells have been chosen from the active list, the result is the
accumulative-cost or weighted-distance Grid. The procedure used
ensures that the lowest accumulative cost is guaranteed for each
cell.

t is conceivable that when the growing patterns meet, cells from one
growth pattern will discover that they can reach a source cell in
another set or growth pattern more cheaply; if so, they will be
reassigned to the new source. This behaviour was witnessed by the
cell at row 3, column 1 earlier, but is also exemplified below by the
cell located at row 3, column 6.

-19-
Accessibility Modelling in ArcView 3 CIAT

Cost Back Link Cost Allocation

The cost-distance Grid identifies the accumulative cost for each cell Cost allocation produces a Grid very similar to the Euclidean
to return to the closest cell in the set of source cells. It does not show allocation function; like the Euclidean function, it returns a Grid
which source cell to return to or how to get there. The cost back link identifying which cells will be allocated to which source, but unlike
returns a Grid with a value range from 0 to 8 that can be used to the Euclidean function, on the basis of the lowest accumulative cost
reconstruct the route to the source. Each value (0 through 8) to reach a source.
identifies which neighbouring cell to move into to get back to the
source. 0 is a source, 1 indicates move to the immediate right, 2 to
the lower right diagonal, 3 to the cell immediately south, etc..

If the cell is assigned 5 as part of the least cost path to a source, the
path should move to the left neighbouring cell. If that cell has 7, the Cost Path
path should move due north. Once the accumulative cost and back-link Grids are created, least-
cost path routes can be derived from any designated destination cell
or zone(s). The cost-path request retraces the destination cells
through the back-link Grid to a source. If there are multiple cells or
zones as input destinations, the least cost path can be calculated
from each cell (resulting in multiple paths, one path for each cell), or
each zone (one path from each zone).

-20-
Accessibility Modelling in ArcView 3 CIAT

locations with No Data in the cost Grid act as barriers in the cost
surface functions. Any cell location that is assigned No Data on the
input cost surface will receive No Data on all three output Grids of
the aGrid.CostDistance request. If a mask has been set in the
analysis environment, any masked cells will be treated as No Data
values.

When two or more cost paths from different zones converge on the
way to a source and continue the remaining distance together, the
joint path to the cell is assigned 2, since the owner of the segment
cannot be determined. The value assigned to each path is assigned
by the order in which the zone cells are encountered in the scanning
process.

The set of source cells consists of all cells in the source Grid that
have valid values. Cells that have No Data values are not included in
the source set. The value 0 is considered a legitimate source. Cell

-21-