Processing

technical

Data availability and requirements

for flood hazard mapping
by Zelda Els, Aurecon, and Adriaan van Niekerk, University of Stellenbosch

This study investigated whether the existing data in South Africa is adequate for flood hazard mapping in rural areas.
The most appropriate data sources were identified and, through the combination of GIS and hydraulic modelling,
evaluated for determining the likely extent, frequency and depth of predicted flood events. The conclusion is that
one-dimensional flood modelling can be carried out in some areas, but that more investment is needed for the
collection of data appropriate for flood hazard mapping.

F
lood hazard maps play an integral Flood parameter (bold
part in flood risk assessments Requirement Scale/level indicates minimal
as they communicate the extent parameters to be mapped)
and probable recurrence of predicted National/regional 1:100 000 - 1 :1 000 000 Extent
floods events [1, 2, 3, 4]. However, Disaster
management Extent, depth and
very few flood hazard maps exist for
planning Local 1:5000 - 1:50 000 other parameters where
rural areas, particularly in developing appropriate
countries such as South Africa, mainly
National/regional 1:100 000 - 1:500 000 Extent
because adequate data is often
Town planning Extent for different
unavailable [5, 6, 7]. The aim of this
(land use) Local 1:5000 - 1:25 000 probabilities, depth,
article is to determine whether floods
velocity and duration
can be modeled using the limited data
sources available for rural areas in National/regional 1:100 000 - 1:500 000 Extent

South Africa. The article first introduces Emergency Extent and depth for
planning and different return periods,
the concept of flood hazard mapping Local 1:5000 - 1:25 000
management and other parameters
and then provides an overview of the
where appropriate
existing methodologies, software and
data. The suitability of available data Public Extent for different
Local 1:10 000 - 1:25 000
awareness probabilities and depth
is evaluated using HEC-RAS, a popular
flood modelling package. The paper Extent for different
concludes with a synthesis of the Insurance Local 1:10 000 - 1:25 000 probabilities, depth, and
findings and identifies critical datasets velocity (if significant)
that need to be routinely collected for
Table 1: Flood parameters and scale according to requirements. (Adapted from: Martini and
mapping flood hazard in rural areas. Loat (2007: 13-16).

Flood hazard mapping

in flood hazard mapping as their data for mapping of the extent and
Flood hazard maps are the results
requirements (see Table 1) determine depth flood parameters at a scale of
of hazard analyses and indicate the
the type, scale and level of flood 1:50 000 or larger.
locations of flood events for certain
parameters that should be mapped Various software packages are
return periods [8].The purpose of flood
[10, 11, 12]. available for the modelling of flood
hazard mapping is to determine the
following flood parameters [9, 10, 11]: Each requirement has different levels/ hazards. Differentiation between these
scales at which the specific flood software packages can be made based
• Frequency, i.e. the probability of an
parameters can be mapped. For on modelling type, dimension (i.e.
occurrence in a specific time period;
disaster management, the mapping one, two or three dimensions), data
• Extent, i.e. the geographical area of requirements and the applications
of the flood extent at national and
impact; for which it was originally developed.
regional levels are required at a scale
• Severity (intensity), i.e. the water between 1:100 000 and Pender & Néelz [2] compiled a table
level or water depths; and 1:1 000 000,while at local level, the (see Table 2) listing the various flood
• Flow velocity and/or direction extent and depth can be mapped at a modelling software that are available
of flow. scale between 1:5000 and 1:50 000 as well as their respective potential
Stakeholders (e.g. disaster managers, [11]. This article focusses on local applications.
emergency workers, insurance disaster management planning and will Two- and three-dimensional models
companies) play an important role consequently evaluate the available require very high resolution data,

PositionIT – January/February 2013 1

 Processing technical

as well as data pertaining to flood Distinguishing

wave characteristics, to determine Dimension Available software Potential application
features
the duration and peaks [12, 13].
No physical Broad scale assessment
One-dimensional software is normally ArcGIS, Delta
0D laws included in of flood extents and flood
used to perform basic hydraulic Mapper, etc.
simulation depths
analysis of, for example, water level
Solution of one- Design scale modelling
and extent. The water level is then
dimensional Infoworks RS (ISIS), which can be of the order of
combined with a digital terrain model 1D
St. Venant Mike 11, HEC-RAS 10-100s of km, depending
(DTM) to determine the water depth equations on catchment size
or inundation zones [12, 13, 14].
Design scale modelling
Infoworks RS, Mike 11 and HEC-RAS 1D plus flood which can be of the order of
are examples of software packages that storage, cell 10-100s of km, depending
carry out one-dimensional modelling. Infoworks RS (ISIS),
1D+ approach to the on catchment size, also
Mike 11, HEC-RAS
Both HEC-RAS and Mike 11 are simulation of has the potential for broad
accepted by the Federal Emergency floodplain flow scale application if used with
Management Agency (FEMA) for sparse cross-section data
hydraulic modelling of one-dimensional 2D minus the law
Broad scale modelling of
flow [14]. of conservation of
2D- LISFLOOD-FP urban inundation depending
momentum for the
on cell dimensions
Data requirements analysis floodplain flow
The availability of data determines the Design scale modelling of
methodology and the flood parameters Solution of the the order of 10s of km. May
that can be modelled in flood hazard two-dimensional TUFLOW, Mike 21, have the potential for use
2D
mapping [7].The following data inputs shallow wave TELEMAC in broad scale modelling if
equation applied with very course
are required for modelling the most
grids
basic flood parameters (e.g. extent and
depth): Predominately coastal
modelling applications
• Topographical data describing the 2D plus a solution
where 3D velocity profiles
topography of the study area for vertical
2D +
TELEMAC 3D are important. Has also
[7, 12, 13, 15, 16, 17] velocities using
been applied to reach scale
continuity only
• Land cover (land use) data to river modelling problems in
calculate Manning’s roughness research projects.
coefficient [7, 13, 17, 18], and Solution of the
Local prediction of three-
• Historical data for the calibration three-dimensional
CFX, FLUENT, dimensional velocity fields
and validation of the flood 3D Reynolds averaged
PHOENIX in main channels and
Navior Stokes
modelling [7, 11, 13, 17, 18] floodplains
equations
Topographical data
Table 2: Flood modelling software and applications. (Source: Pender and Néelz (2007: 108).
The quality of the topographical
data source is an important factor
in the accuracy and reliability of the for stream channel descriptions in the earth [24]. Often, these two terms
generated flood hazard maps hydraulic analyses. Although a TIN can are used interchangeably, but in this
[11, 12, 13, 16], thus it is important to be created from a DEM, the stream study land cover will be the preferred
select the most suitable topographical channel detail is still not sufficient for term as it has a more direct bearing
data source. The topographical data on the data required for hydraulic
hydraulic analysis [19, 20, 21, 22].
sources available in South Africa are modelling.
Generally, TINs are recommended as
summarised in Table 3. Land cover data is required to
they can potentially provide a better
It is evident from Table 3 that very representation of the river channel and determine the Manning’s roughness
limited “wall-to-wall” topographical data surrounding area [21, 22, 23]. The coefficient needed in hydraulic
is available for South Africa. According available DEMs have resolutions that modelling [25, 26]. Manning’s
to Martini and Loat [11], a DEM with a vary between 30 m and 1 km, and do roughness coefficient refers to the
resolution of between 10 m and not adhere to the required guidelines. energy loss due to the friction caused
5 m is required for flood modelling. In The best topographical data that is by the surface roughness of the
addition, hydraulic modelling requires available is the 5 m contours (scale channel wall, the vegetation, channel
an accurate representation of the 1:10 000). irregularities (e.g. depressions, ridges,
stream channel and surrounding area. sand bars and holes) and the curvature
Low resolution (>10 m) DEMs are Land cover (land use) data of the channel [27].The available
insufficient for hydraulic modelling of Land use refers to the human activity land cover data sets (see Table 4) for
river channels as it cannot represent (e.g. grazing, conservation) associated South Africa include both national and
sharp changes in the topography of with a specific land unit, while land international data sources.
the terrain. DEMs are sufficient for cover refers to all natural features (e.g. For local flood modelling, a land cover
extracting watershed characteristics in vegetation (natural or planted), water, data set at a scale of 1:50 000 or larger
hydrological analyses while triangulated ice, bare rock) and man-made features is needed. International land cover data
irregular networks (TINs) are suitable (e.g. buildings, roads) on the surface of sources (e.g. Globcover and GLC 2000)

2 PositionIT – January/February 2013

 Processing technical

are insufficient for local applications due Resolution/

to their coarse resolutions, 300 m and Type Dataset Coverage Source
interval
1 km respectively. The NLC 2000 is
CD:NGI South Africa (SA)
consequently the most suitable of the 5-20 m* CD:NGI (2011a)
1:10 000 contours (partial coverage)
remaining four data sets as it was Contour
CD:NGI
carried out at a scale of 1:50 000 and 20 m* SA CD:NGI (2011a)
1:50 000 contours
contains 45 land cover classes. The
CD:NGI SA (partial
NLC 2009 was also considered, but 25 m CD:NGI (2011a)
25 m DEM coverage)
it differentiates between significantly
fewer land covers (8) compared to the ASTER GDEM 30 m Global ERSDAC (2009)
45 of the NLC 2000. Using the latter CD:NGI SA (partial
DEM 50 m CD:NGI (2011a)
dataset will consequently improve the 50 m DEM coverage)
accuracy of the Manning’s roughness SRTM
90 m Global ERSDAC (2009)
coefficient. 90 m DEM
GTOPO 30 1 km Global ERSDAC (2009)
Historical data
Table 3: Topographical data sources in South Africa. (* Vertical interval)
Historical data is required for
hydrological data input and for the
calibration and validation of the flood Type Name Resolution/scale Coverage Source
modelling [7, 11, 13]. Sources of SA, Lesotho &
historical data can include [11, 28]: Vector NLC 1994 1:250 000 Fairbanks et al. (2000)
Swaziland
• Dated flood maps Vector NLC 2000 1:50 000 SA Schoeman et al. (2010)
• Water level records of rivers Vector ENPAT 1:250 000 SA ENPAT (2001)
• Gauge station records (for velocity) Raster NLC 2009 30 m SA SANBI (2010)
• Flood marks Raster Globcover 300 m Global Bicheron et al. (2008)
• Pictures, paintings and drawings Raster GLC 2000 1 km Global Mayaux et al. (2006)
• Newspaper articles about past flood
Table 4: Land cover data sources for South Africa.
events, with dated photographs
• Historical reports or books about
flood events HEC-RAS 4.1 was selected for
sets are needed to perform calibration
• Aerial and satellite photos. and validation [33]. The flood extent evaluating the data in Table 6 as it
The capturing of data on hazard image should preferably be captured is the most widely used computer
events, both spatially and non-spatially, within 8 to 24 hours after the flood program for floodplain analysis [35]
is very limited in South Africa [29]. event has occurred [34]. The selection and has been approved by FEMA [14].
Details of historical events are mainly of imagery for the verification and ArcGIS 9.3 and the HEC-GeoRAS
obtained from newspaper articles and calibration of the flood model depends 4.1 extension were chosen for data
South African Weather Services (SAWS) on the selected study area and the preparation and visualisation of the
reports [30]. availability of imagery for the period flood modelling results as indicated
during which the historical flood events in Fig. 1.
Water level and velocity data
occurred.
The Hydrological Services at There are three main phases in flood
Images listed in Table 5 with modelling analysis using HEC-RAS,
Department of Water Affairs (DWA)
resolutions of 30 m or better (e.g. namely data preparation, hydraulic
are responsible for providing
GeoEye, SPOT, Landsat, CBERS CCD, analysis and post-processing. Data
hydrological data and information. Their
RESOURCESAT-1, ERS AMI SAR 2
Hydrological Information System (HIS) preparation and post-processing
(RADAR)) can be considered for
consists of various databases which are performed in ArcMap with the
calibration and validation purposes.
include data on river stations, river flow HEC-GeoRAS extension, and all
Higher resolution imagery (e.g.
and other related information. More hydraulic analysis is performed within
GeoEye, WorldView) enables the
than 800 gauging stations for river flow HEC-RAS. HEC-GeoRAS provides the
identification of flooding remnants
exist and each station has one or more capability to capture the geometric
(e.g. erosion terraces, debris,
monitoring points [31]. data, prepared in ArcMap, according to
alluvial deposits).
the HEC-RAS format required for the
Aerial and satellite imagery
Flood modelling methodology hydraulic modelling. Thus, HEC-GeoRAS
Imagery is used to capture required
is mainly responsible for data conversion
data layers for hydraulic analysis [22] It should be clear from the previous
between the GIS software package,
and for the calibration and validation sections that the level of detail of
ArcMap, and the hydraulic software,
of the flood model [7, 11, 13]. Table available topographical and flow data is
HEC-RAS [25, 36, 37].
5 lists the relevant imagery sources insufficient for complex flood modelling.
available in South Africa. Only one-dimensional, steady flow HEC-RAS was applied in the Drieheuvels
hydraulic modelling can potentially be area (Fig. 2) located along the Great
Images captured before and during
carried out using the available data Berg River between Piketberg and
the disaster are required to allow for
(see Table 6) to determine primary Riebeeck West.
proper identification of previously
existing water bodies for calibration and flood parameters (e.g. extent and The hydraulic results provide two
validation purposes [32]. Separate data depth) at a scale of 1:50 000. parameters for flood hazard mapping,

PositionIT – January/February 2013 3

 Processing technical

namely extent and depth for 20-, 50-

Resolution/
and 100-year (Fig. 3) flood events. After Type Name Coverage Source
scale
performing a visual inspection along
1:20 000 SA CD:SM (2006)
the boundary of the flood extent, it
1:30 000 SA CD:NGI (2011b)
was found that none of the homesteads
located along the river will be affected 1:40 000 SA CD:SM (2006)
CD:NGI
by a 100-year event. Although the Panchromatic (Pan) 1:50 000 –
SA CD:SM (2006)
clearance is estimated at 50 m for some 1:60 000
Aerial
homesteads, it should raise concern photos 1:80 000 SA CD:SM (2006)
for a 500- or 1000-year event. Several 1:150 000 SA CD:NGI (2011b)
small areas of cultivated agricultural 1:10 000 SA CD:NGI (2011b)
land will be flooded by all three selected
CD:NGI Colour 1:20 000 SA CD:SM (2006)
occurrence interval events.
1:30 000 SA CD:SM (2006)
Data limitations and Orthophotos CD:NGI Colour 1:10 000 SA CD:NGI (2011b)
recommendations GeoEye-1 Pan 0,41 m Global GeoEye (2011)

Topographical data WorldView-1 Pan 0,5 m Global DigitalGlobe (2011b)

The availability of suitable topographical WorldView-1 MS 0,55 m Global DigitalGlobe (2011b)

data to accurately represent the river WorldView-2 Pan 0,46 m Global DigitalGlobe (2011c)
channel and surrounding terrain is WorldView-2 MS 0,52 m Global DigitalGlobe (2011c)
one of the main impediments for Quickbird Pan 0,65 m Global DigitalGlobe (2011a)
flood modelling in rural South Africa.
IKONOS Pan 0,82 m Global GeoEye (2010)
Topographical data is fundamental to
GeoEye-1 MS 1,65 m Global GeoEye (2011)
flood modelling and is used throughout
the process. A larger scale data source SPOT 5 Pan 2,5 m Global SPOT Image (2008)
(i.e. larger than 1:10 000) is ideally Quickbird MS 2,62 m Global DigitalGlobe (2011a)
needed to produce better results. IKONOS MS 3,2 m Global GeoEye (2010)
Better topographical data would also RapidEye (Level 1B) 5m Global RapidEye (2011)
enable the creation of larger scale
RESOURCESAT-1
locality maps for more detailed disaster 5,8 m Global ISRO (2011)
(IRS-P6) Liss 4 MS
management planning. For such
SA (partial Koekemoer, Govender
detailed analyses, DEMs of 5 m to Sumbandilasat MSS 6,25m
coverage) & Smit (2011)
Satellite
10 m resolution or 1 m vertical interval
imagery SPOT 1, 2, 3 Pan 10 m Global SPOT Image (2010)
contours are required [11]. Although
(optical)
SPOT 4 Pan 10 m Global SPOT Image (2010)
the recently-released 5 m-resolution
Stellenbosch University DEM (SUDEM) SPOT 5 MSS 10 m Global SPOT Image (2008)
of South Africa will likely be of great Landsat 7 Pan 15 m Global GLCF (2009)
value for flood modelling [38], higher SPOT 1, 2, 3 MS 20 m Global SPOT Image (2010)
resolution DTMs are required for more
SPOT 4 MS 20 m Global SPOT Image (2010)
detailed modelling applications.
CBERS CCD 20 m Global CBERS (2011)
High resolution DEMs can be developed
RESOURCESAT-1
using photogrammetry and stereo 23,5 m Global ISRO (2011)
(IRS-P6) Liss 3 MS
imagery [39]. Accuracies of 10 cm
Landsat 4, 5 MSS 30 m SA GLCF (2009)
(vertical) have been achieved using
Landsat 7 MSS 30 m SA GLCF (2009)
very high resolution (0,5 m) satellite
imagery (e.g. WorldView-2 and Landsat 1-4 MSS 60 m SA GLCF (2009)
GeoEye), although such imagery is Modis MS(band 1-2) 250 m Global NASA (2011)
still relatively expensive. The South Modis MS(band 3-7) 500 m Global NASA (2011)
African National Space Agency (SANSA)
Modis MS(band
is investigating the inclusion of the 1 km Global NASA (2011)
8-36)
Indian P5 (Cartosat-1) satellite data
NOAA AVHRR 1 km Global NOAA (2011)
to its current portfolio. This will enable
TerraSAR-X High
the development of 5 m DEMs that can 1m Global Infoterra (2009)
Resolution Spotlight
be applied in hydraulic modelling [40].
TerraSAR-X Spotlight 2m Global Infoterra (2009)
The Centre for Geographical Analysis
Satellite Radarsat-2 3m Global CSA (2011)
at Stellenbosch University is currently
imagery
carrying out research to automatically TerraSAR-X StripMap 3m Global Infoterra (2009)
(radar)
extract 2 m resolution DEMs from Radarsat-1 8m Global CSA (2011)
0,5 m stereo aerial photography
ERS AMI SAR 2 30 m Global Eurimage (2011)
over large areas. Such data will
TerraSAR-X ScanSAR 100 km Global Infoterra (2009)
significantly improve the accuracy of
hydrologic modelling. Table 5: Aerial and satellite imagery available for flood model calibration and validation in South Africa.

4 PositionIT – January/February 2013

 Processing technical

Light detection and ranging (lidar) is

currently the most accurate method
for generating high resolution (up
to 0,1 m) DEMs, but this technology
is still very expensive [41, 42],
especially for large areas [39]. Other
technologies, such as Interferometric
Synthetic Aperture Radar (IFSAR) are
also available to develop DEMs with
up to 0,5 m vertical accuracy [39]. A
number of satellite systems (e.g. ERS,
Fig. 1: The software used in the flood modeling methodology.
RADARSAT, TerraSAR-X) provide this
data [43].

Land cover data

Land cover data is needed for
the hydraulic model to determine
Manning’s coefficient values along
each cross-sectional cut line. Most of
the current land cover data sets in
South Africa are very dated and at a
small scale. The CD:NGI appointed
the Satellite Application Centre (SAC),
now SANSA, to facilitate workshops,
during January and February 2008,
with all users of land cover data to
determine a methodology for creating
a new land cover data set. Over 180
land cover classes were defined during
this process and it was decided to
condense this classification into eight
super classes and 32 sub-classes.
These classes will be derived from a
combination of satellite imagery that
includes Landsat 7 ETM+, SPOT 5, SAR
and lidar technologies [44]. CD:NGI
compiled bid specifications to test the
methodology at five different district
municipalities across South Africa
using an automated classification [45].
Such land cover data will significantly
enhance hydraulic modelling as more
detailed allocations of Manning’s
coefficient values would be possible.
The larger scale of the data will allow
more detailed modelling.

Historical data
• Water level and velocity data
Peak discharge values are required Fig. 2: Drieheuvels demonstration area.
to calculate the flow for different
occurrence intervals. Although gauge
stations exist for both major and There is a need for the data recorded
Data input Name
minor rivers, flow data is often absent, at gauge stations to be more complete
inconsistent or incomplete for many by also including water depth and CD:NGI 1:10 000
Topographic
stations. In such cases, a hydrological discharge values. Firstly, attention (5 m) contours
analysis needs to be carried out for the should be given to updating existing Land cover NLC 2000
specific river catchment before gauge stations to record comprehensive
hydrological databases. Thereafter, DWA’s Hydrological
the hydraulic modelling can commence Water level
Information System
[46]. In addition, very few of these flood-prone areas along the river
gauge stations contain verified should be identified where additional
Table 6: Data sources suitable for
data sets. gauge stations could be positioned flood hazard mapping.

PositionIT – January/February 2013 5

 Processing technical

[47]. In addition, these gauge stations

can be equipped to act as early
warning systems for flood events, thus
serving a dual purpose.
Alternatively, various government
organisations, from national to
municipality level, can have their own
water monitoring programmes that
capture detailed data with regard to
water level, velocity and pollution.
Possible planning and coordination
between these organisations can allow
stakeholders access to these water
monitoring databases to resolve the
gauge station data shortage.
• Aerial photography and
satellite imagery
Although a wide selection of satellite
systems are available, very few of
these have the appropriate temporal
resolution to collect imagery frequently
(i.e. daily). Even if such sensors were
available, cloud cover will likely limit
the use of optical sensors for studying
flood events. High-resolution active
sensors (i.e. radar) with the capability
to penetrate clouds are consequently
required to accurately calibrate flood
models. Such data is currently still
prohibitively expensive. Fortunately the
SPIDER programme was established
in 2006 by the United Nations Office
for Outer Space Affairs (UNOOSA) with
the objective to provide access and
develop the capacity of all countries
and organisations (international and
regional) in the use of all space-based
information during the whole disaster
management cycle. One of the Fig. 3: Flood depth map for 100-year flood level.
frameworks used by the UN-SPIDER,
is the use of SpaceAid that facilitates
fast and efficient access to satellite South Africa urgently needs to improve facilitate communication in flood risk
imagery (including radar) during a the existing data sources required management.Environmental Hazards 7:
disaster on a 24 hours per day/7 days for hydraulic modelling as it will 106-114, (2007).
a week basis [32]. significantly strengthen the capacity [3] V Schneider J Verdin and J Bales
to produce flood hazard maps. The Flood hazard inundation mapping:
Concluding remarks Paper presented at the Global Facility
current situation of limited availability
This study has shown that flood for Disaster Reduction and Recovery
of flood hazard maps in rural areas can
Seminar Series, Washington, (2007).
modelling using available data is have a disastrous impact, especially on
[4] MF Viljoen and HJ Booysen: Planning
possible in some rural areas in South already vulnerable communities.
and management of flood damage
Africa. HEC-RAS was identified as a control: The South African experience.
suitable software package as it Acknowledgements
Irrigation and Drainage 55: S83-S91,
requires minimal data input to perform The authors would like to thank (2006).
hydraulic analysis. Prof. Dušan M. Sakulski from Free State [5] RP Osti Flood mapping: A point of
University for his guidance and support departure for flood risk management.
The identification of a suitable area
during this study. Paper delivered at the World
for demonstrating how flood modelling Meteorological Organisation’s Expert
can be performed was challenging due References Group Meeting on Flood Mapping,
to the limited coverage of 1:10 000 Geneva, (2008).
[1] K Hardmeyer and MA Spencer: Using
scale contour maps and river gauge [6] H Shamaoma, N Kerle and D Alkema:
risk-based analysis and geographic
stations. Furthermore, the model result Extraction of flood-modelling related
information systems to assess flooding
base-data from multi-source remote
could not be calibrated and validated problems in an urban watershed
sensing imagery. Paper delivered
due to limited imagery and thus should in Rhode Island.Environmental
at ISPRS Commission VII Mid-term
consequently only be considered as a Management 39: 563-574, (2007).
Symposium “Remote Sensing: From
prediction of the flood parameters for [2] G Pender and S Néelz: Use of Pixel to Processes” conference,
large scale purposes. computer models of flood inundation to Enschede, (2006).

6 PositionIT – January/February 2013

 Processing technical

[7] M Zimmerman: Data mining for flood [20] PB Andrysiak and D Maidment: Visual [34] GMES (Global Monitoring for
mapping. Paper delivered at the World floodplain modelling with geographic Environment and Security): Floods
Meteorological Organisation’s Expert information systems (GIS). Publication [online].Frascati: European Space
Group Meeting on Flood Mapping, No 00-4. Austin: Bureau of Engineering Agency, (2005). Available from:
Geneva, (2008). Research, University of Texas, (2000). www.emergencyresponse.eu/gmes/
[8] M Bründl, HE Romang, N Bischof and [21] D Maidment and D Djokic: Hydrologic docs_wsw/RUB_119/ssp_gmes_
CM Rheinberger: The risk concept and and hydraulic modelling support with datasheet_01_floods_vi.pdf [Accessed
its application in natural hazard risk Geographic Information Systems. 9 June 2011].
management in Switzerland. Natural Redlands: ESRI Press, (2000). [35] G Dyhouse, J Hatchett and J Benn:
Science and Earth System Sciences 9: [22] F Rusinga: Interview about HEC-RAS Floodplain modelling using HEC-RAS.
801-803, (2009). flood modelling and topographic data. Michigan: Haestad Press, (2003).
[9] ISDR (International Strategy for Cape Town, Aurecon South Africa (Pty)
[36] CT Ackerman, TA Evans and
Disaster Reduction): Living with risk: Ltd (28 October 2009).
GW Brunner: Hec-GeoRAS: Linking GIS
A global review of disaster reduction [23] A Casas, G Benito, VR Thorndycraft and to hydraulic analysis using ARC/INFO
initiatives. Geneva: ISDR, (2004). M Rico: The topographic data source of and HEC-RAS. Paper delivered at the
[10] H Kleeberg, K Röttcher and H digital terrain models as a key element 19th Annual ESRI User Conference,
Spanknebel (eds): Flood hazard map in the accuracy of hydraulic flood San Diego, (1999).
guidelines [online]. Heft: German modelling.Earth Surface Processes and
Landforms 31: 444-456, (2006). [37] V Merwade: 2009.Tutorial on using
Working Group of the Federal States on Hec-GeoRAS with ArcGIS 9.3 [online].
Water Issues, (2006). Available from [24] M Thompson: A standard
West Lafayette: University of Purdue.
www.lawa.de/documents/Flood-Hazard- land-cover classification scheme for
Available from: http://web.ics.
Map-Guidelines_ea8.pdf [Accessed 19 remote-sensing applications in South
purdue.edu/ ~vmerwade/education/
July 2010]. Africa.South African Journal of Science
georastutorial.pdf [Accessed 3 May
92: 34-42, (1996).
[11] F Martini and R Loat (eds): Handbook 2010].
on good practices for flood mapping [25] CT Ackerman: HEC-GeoRAS, GIS tools
[38] A Van Niekerk: Stellenbosch University
in Europe.Emmeloord: DrukkerijFeiko for support of HEC-RAS using ArcGIS.
Davis: United States Army Corps of digital elevation model (SUDEM):
Stevens, (2007).
Engineers, (2005). 2011 edition. Stellenbosch: Centre for
[12] H De Moel, J Van Alphen and Geographical Analysis, (2012).
JCJH Aerts: Flood maps in Europe — [26] T Hernandes and PE Zhang: Floodplain
analysis using computational tools. [39] R Sanders, F Shaw, H MacKay, H Galy
methods, availability and use. Natural
Paper presented at the World and M Foote: National flood modelling
Hazards and Earth System Sciences 9:
Environmental and Water Resources for insurance purposes: Using IFSAR
289-301, (2009).
Congress of the Environmental and for flood risk estimation in Europe.
[13] B Büchele, H Kreibich, A Kron, Hydrology and Earth System Sciences
Water Resources Institute, Tampa,
A Thieken, J Ihringer, P Oberle,
(2007). 9: 449-456, (2005).
B Merz and F Nestmann: Flood-risk
mapping: contributions towards an [27] WS Janna: Introduction to fluid [40] N Nieckau: Providing access to cutting
enhanced assessment of extreme mechanics. 4th ed. Boca Raton: CRC edge satellite imagery [online].
events and associated risks. Natural Press, (2009). Johannesburg: EE Publishers (Pty) Ltd.
Science and Earth System Sciences 6: [28] FEI: Flood mapping [online]. Helsinki: Available from: www.eepublishers.
485-503, (2006). Finland’s Environmental Administration. co.za/article/providing-access-to-
Available from: www.ymparisto.fi cutting-edge-satellite-imagery.html
[14] FEMA (Federal Emergency Management
[Accessed 2 November 2007]. [Accessed 1 June 2011], (2011).
Agency): Numerical models meeting
the minimum requirement of NFIP [29] A Halloway, G Fortune and [41] A Combrink: Introduction to lidar-based
[online]. Washington: FEMA, 2009. V Chasi: Risk and development annual aerial surveys (Part 2). PositionIT.
Available from www.fema.gov/plan/ review: Western Cape 2010. Cape June 2011.
prevent/fhm/en_hydra.shtm [Accessed Town: Peri-Peri Publications, Disaster [42] K Joyce: Remote sensing and
10 August 2009]. Mitigation for Sustainable Livelihoods
geological hazard assessment. Paper
Programme, (2010).
[15] D Maidment: ArcHydro: GIS for water delivered at the Earth Observation
resources. Redlands: ESRI Press, 2002. [30] D Sakulski: Spatial and temporal and Geological Hazard Assessment
frequency and distribution of hazardous Workshop, Pretoria (2011).
[16] M Sane and M Huokuna: Procedure
events over South African territory for
for identifying automatically possible [43] Astrium 2012: Radar technology
the period between years 1800 and
flood risk areas. Paper presented at can see through clouds and smoke
1995 based on ‘CAELUM’ document,
the Thematic Workshop for Flood to deliver striking imagery [online].
Pretoria: Cooperation Unit for Southern
Mapping of the World Meteorological SuresnesCedex: ASTRIUM. Available
Africa, United Nations University,
Organisation, Dublin, (2009). from: http://www.astrium.eads.net/
(2007).
[17] JM Wright: Floodplain management: en/space-benefits-for-earth/radar-
[31] DWA (Department of Water Affairs):
Principles and current practices. technology-can-see-through-clouds-
Hydrological Services [online]. Pretoria:
Knoxville: University of Tennessee, and-smoke-to-deliver-striking-imagery.
DWA. Available from: www.dwaf.gov.
(2008). html [Accessed 17 October 2012].
za/Hydrology/about.htm [Accessed
[18] M Mavrova-Guirginova, D Pencheva and 18 October 2009]. [44] W Lück, P Mhangara, L Kleyn, and
M Belyashka: Maps in the context of H Remas 2010. Land cover field guide.
[32] O Altan, R Backhaus, P Boccardo and
the European Flood Risk Directive. Pretoria: CSIR Satellite Application
S Zlatanova (eds): Geoinformation
Paper presented at 3rd International Centre (SAC). Available from: www.ngi.
for disaster and risk management:
Conference on Cartography and GIS, Examples and best practices. gov.za/images/stories/technical_info/
Nessebar, (2010). Copenhagen: Joint Board of Geospatial publication/CDNGI_Land_Cover_
[19] DJ Anderson: GIS-based hydrologic Information Societies, (2010). Methodology_and_Field_Guide_v2.pdf
and hydraulic modelling for floodplain [Accessed 3 February 2011].
[33] PA Longley, MF Goodchild, DJ Maguire
delineation at highway river crossings. and DW Rhind (eds). Geographic [45] Parker A: (aparker@ruraldevelopment.
Master’s thesis. Austin: University of Information Systems and Science. 2nd gov.za) 2011. Re: NGI Land Cover
Texas, Faculty of Graduate School, ed. New York: John Wiley & Sons Ltd, Initiatives. E-mail to A van Niekerk
(2000). (2005). ([email protected]) (18 July 2011).

PositionIT – January/February 2013 7

 Processing technical

[46] F Rusinga: Interview about for South Africa [online]. Pretoria: Brandenburg an der Havel: RapidEye.
calibration and validation of flood Department of Environmental Affairs Available from: www.rapideye.de/
modelling results. Cape Town, Aurecon (DEA). Available from: upload/ RE_Product_Specifications_
South Africa (Pty) Ltd (12 February www.environment.gov.za/enviro-info/ ENG.pdf [Accessed 2 August 2011].
2010). enpat.htm [Accessed 24 September [72] SANBI (South African National
[47] B Clarke & G Ratcliffe (eds) 2007. 2009]. Biodiversity Institute) 2010. National
Berg River Baseline Monitoring [58] ERSDAC (Earth Remote Sensing Data land cover 2009 [online]. Pretoria:
Programme: Final Report–Volume 5: Analysis Center) 2009.Aster [online]. SANBI. Available from: http://bgis.
Synthesis. Publication No P WMA 19/ Tokyo: ERSDAC. Available from sanbi.org/ landcover /project.asp
G10/00/2107. Pretoria: Department of www.ersdac.or.jp/GDEM/E/2.html [Accessed 6 October 2010].
Water Affairs. [Accessed 25 September 2009]. [73] F Schoeman, TS Newby, MW Thompson
[48] P Bicheron, P Defourny, C Brockmann, [59] Eurimage 2011.ERS [online]. Rome: and EC Van Den Berg: South African
L Schouten, C Vancutsem, M Huc, Eurimage. Available from: National Land-cover Change Map.
S Bontemps, M Leroy, F Archard, www.eurimage.com/products/ers.html Publication No GW/A/2010/47.
M Herold, F Ramera and O Arino 2008. [Accessed 2 June 2011]. Pretoria: Institute for Soil, Climate
GLOBCOVER: Products description [60] DHK Fairbanks, MW Thompson, and Water, Agriculture Research
and validation report [online]. DE Vink, TS Newby, HM van den Berg Council, (2010).
Toulouse: Medias France. Available and DA Everard 2000. The South [74] SPOT Image 2008. Technical
from: http://globcover.s3.amazonaws. African Land-cover Characteristics information [online]. Toulouse: SPOT
com/LandCover_V2.2/ GLOBCOVER_ Database: A synopsis of the landscape. Image. Available from:
Products_Description_Validation_ South African Journal of Science 96: www.spotimage.com/automne_
Report_I2.1.pdf [Accessed 18 May 69-82. modules_files/standard/public/ p234_0
2011]. 916b39e58c710b3bb5788fdcc025f80niv
[61] GeoEye 2010.GeoEye constellation.
[49] CD:NGI (National Geo-spatial Virginia: GeoEye. eau_anglais_2008.pdf [Accessed
Information) 2011a. Maps and 09 October 2010].
[62] GeoEye 2011.GeoEye-1 fact sheet.
geo-spatial information [online]. [75] SPOT Image 2010. SPOT satellite
Herndon: GeoEye.
Cape Town: CD:NGI. Available technical data [online]. Toulouse:
from: www.ngi.gov.za/ index.php/ [63] GLCF (Global Land Cover Facility) 2009.
SPOT Image. Available from: www.
Image-tabs-home/maps-and-geo- Landsat [online]. Maryland: GLFC.
spotimage.com/automne_modules_
spatial-information.html [Accessed Available from www.landcover.org/data/
files/standard/ public/p229_48f99c03cb
15 May 2011]. landsat/ [Accessed 20 October 2009].
2bc7f6beb7acc41f29fffSpotSatelliteTech
[50] CD:NGI (National Geo-spatial [64] Infoterra 2009.TerraSAR-X basic image nicalData_EN_Sept2010.pdf [Accessed
Information) 2011b. National aerial products [online]. Friedrichshafen: 5 June 2011].
photography and imagery programme Infoterra GmbH. Available from:
Contact Zelda Els, Aurecon,
[online]. Cape Town: CD: NGI. www.infoterra.de/basic-imagery
Tel 021 526-6038,
Available from: www.ngi.gov.za/index. [Accessed 2 August 2011].
[email protected] 
php/Image-tabs-home/ national-aerial- [65] ISRO (Indian Space Research
photography-and-imagery-programme. Organisation) 2011.IRS-P6 [online].
html [Accessed 21 May 2011]. Andhra Pradesh: ISRO. Available from:
[51] CD:SM (Chief Directorate: Surveys and www.nrsc.gov.in/irs-p6.html [Accessed
Mapping) 2006. Mapping South Africa. 2 August 2011].
Cape Town: CD: SM. [66] Infoterra 2009.TerraSAR-X basic image
[52] CBERS (China-Brazil Earth Resources products [online]. Friedrichshafen:
Satellite) 2011. CBERS 1, 2 and Infoterra GmbH. Available from:
2B cameras [online]. São José dos www.infoterra.de/basic-imagery
Campos: National Institute of Space [Accessed 2 August 2011].
Research. Available from: www.cbers. [67] J Koekemoer, N Govender and I Smit:
inpe.br/?hl=en&content=cameras SumbandilaSat provides valuable
1e2e2b [Accessed 4 June 2011]. imagery. PositionIT. April/May, (2011).
[53] CSA (Canadian Space Agency) [68] P Mayaux, H Eva, J Gallego,
2011. RADARSAT systems: Satellite AH Strahler, M Herold, S Agrawal,
characteristics [online]. Quebec: CSA. S Naumov, EE De Miranda, CM Di
Available from: www.asc-csa.gc.ca/ Bella, C Ordoyne, Y Kopin and PS Roy:
eng/satellites/radarsat/radarsat- Validation of the global land cover
tableau.asp [Accessed 2 August 2011]. 2000 map. Geoscience and Remote
[54] DigitalGlobe 2011a.Quickbird [online]. Sensing 44: 1728-1739, (2006).
Longmont: DigitalGlobe. Available [69] NASA (National Aeronautics and Space
from: www.digitalglobe.com/index. Administration) 2011.MODIS [online].
php/85/QuickBird [Accessed 2 August Washington: NASA. Available from:
2011]. http://modis.gsfc.nasa.gov/about/
[55] DigitalGlobe 2011b.WorldView-1 specifications.php [Accessed
[online]. Longmont: DigitalGlobe. 2 August 2011].
Available from: www.digitalglobe.com/ [70] NOAA (National Oceanic and
index.php/86/WorldView-1 [Accessed 2 Atmospheric Administration) 2009.
August 2011]. Understanding flood recurrence
[56] DigitalGlobe 2011c.WorldView-2 intervals [online]. Davenport: National
[online]. Longmont: DigitalGlobe. Weather Service. Available from:
Available from: www.digitalglobe.com/ http://test.crh.noaa.gov/dvn/?n=Un
index.php/88/WorldView-2 [Accessed 2 derstandingfloodrecurrenceintervals.
August 2011]. [Accessed 3 July 2009].
[57] ENPAT (Environmental Potential Atlas) [71] RapidEye 2011.Satellite imagery
2001. Environmental Potential Atlas product specification [online].

8 PositionIT – January/February 2013