SOFTWARE REVIEW
HYDROLOGICAL PROCESSES
Hydrol. Process. 19, 1123–1130 (2005)
Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/hyp.5818
The Terrain Analysis System: a tool for hydro-geomorphic
applications
Reviewer:
John B. Lindsay*
Department of Geography,
University of Western Ontario,
London N6A 5C2, Canada
*Correspondence to:
John B. Lindsay, School of
Environment and Development,
Mansfield Cooper Building,
University of Manchester, Oxford
Road, Manchester M13 9PL, UK.
E-mail:
John.Lindsay@manchester.ac.uk
Copyright  2005 John Wiley & Sons, Ltd.
Introduction
Much progress has been made in extending the hydro-geomorphic
modelling toolbox as a result of developments in the field of dig-
ital terrain analysis, also known as geomorphometry (Wilson and
Gallant, 2000; Pike, 2002). The enhancement of computer tech-
nologies and the widespread availability of digital elevation mod-
els (DEMs) since the early 1990s have impacted environmental
modelling greatly (Pike, 2000). Geographical information systems
(GISs) are now widely applied in hydrology and geomorphol-
ogy to automate basin, hillslope, and stream network analyses.
Several commercial GIS packages have incorporated more com-
mon terrain attributes (e.g. slope, aspect, curvature, and wetness
index) and terrain analysis procedures (e.g. basin and stream net-
work extraction). These software packages are, however, often pro-
hibitively expensive. Furthermore, whilst commercial GIS packages
often contain basic terrain analysis capabilities, very few possess
the flexibility and advanced functionality needed by researchers.
Several terrain analysis packages have been developed for this
purpose, e.g. ANUDEM (Hutchinson, 1989), TARDEM/TauDEM
(Tarboton, 1997), LandSerf (Wood, 1998), TAPES (Moore et al.,
1991; Gallant and Wilson, 1996), LANDLORD (Florinsky et al.,
2002), and DiGeM (Conrad, 2002), some of which are freely avail-
able. Unfortunately, many terrain analysis packages lack the visu-
alization and general spatial analysis capabilities of a GIS, or
have been developed as extensions to existing commercial soft-
ware. Additionally, although many terrain analysis packages pos-
sess the advanced functionality needed by researchers, they are
often difficult to use and, therefore, remain inaccessible to non-
experts.
The Terrain Analysis System (TAS) was designed to meet the
research needs of government and academic scientists while being
simple enough in operation to be used for student instruction. TAS
is a stand-alone GIS that possesses much of the spatial analysis
functionality typically found in GIS packages. However, it is also
capable of advanced modelling of catchment processes. The follow-
ing is a brief description of TAS’s functionality for applications in
hydrology and geomorphology.
Received 21 August 2004
1123 Accepted 08 November 2004
 J. B. LINDSAY

2
3
Figure 1. The TAS graphical user interface. Two images, a DEM and a contributing-area image, are displayed with vector basins
and streams overlayed

Functionality found that filling impacts on DEMs more than

Analysis in TAS is organized into five broad cat-
egories: preprocessing, image processing, statisti-
cal analysis, GIS analysis, and terrain analysis.
These five categories are reflected in the menu
structure of the program (Figure 1). The terrain
analysis capabilities of TAS are accessed from
the Pre-processing and Terrain Analysis menus.
The Pre-processing menu contains several sub-
programs used for preparing DEMs for analysis.
These subprograms include routines for ‘burning-
in’ streams and for removing flats, spikes, and
depressions from DEMs. Depression filling is per-
haps the most widely implemented algorithm for
depression removal and is often found in GIS
packages. Nonetheless, Lindsay and Creed (2005)
any other depression-removal method and, there-
fore recommended more widespread use of alter-
natives. Six depression-removal algorithms are
available in TAS, including the Impact Reduction
Approach described by Lindsay and Creed (2005).
There are also subprograms for creating triangu-
lar irregular networks (TINs) and for interpolating
DEMs based on TINs under the Pre-processing
menu.
Table I describes each of the subprograms con-
tained under the Terrain Analysis menu. Sev-
eral of these algorithms have never been imple-
mented elsewhere. Terrain attributes are based on
local neighbourhoods, extended neighbourhoods,

Copyright  2005 John Wiley & Sons, Ltd. 1124 Hydrol. Process. 19, 1123–1130 (2005)
 SOFTWARE REVIEW
Table I. Subprograms within TAS’s Terrain Analysis menu
Submenu Subprogram name
Surface Derivatives Slope
Max downward slope
Aspect
Plan curvature
Profile curvature
Tangential curvature
Local Neighbourhoods Shaded relief
Flow direction
# of downslope neighbours
# of upslope neighbours
# of inflowing cells
Max downslope elev. change
Local elev. percentile
Difference from mean elev.
Standard deviation of elev.
Extended Catchment/dispersal area
Neighbourhoods
Watershed
Downslope flowpath length
Max upslope flowpath length
Min upslope flowpath length
Average upslope flowpath length
Average upslope elev.
Average upslope slope
Edge contamination
Viewshed
Visibility index
Relative Landscape Elev. above pit cell
Position
Relative to peaks and pits
Relative to min and max elev.
Relative to channel and divide
Elev. above target cell
Copyright  2005 John Wiley & Sons, Ltd. 1125
Description
Slope gradient (Zevenbergen and Thorne, 1987)
Slope to the lowest neighbour in a 3 × 3 window
(Burrough and McDonnell, 1998)
Direction of maximum downward gradient
(Zevenbergen and Thorne, 1987)
Along-slope curvature (Zevenbergen and Thorne,
1987)
Downslope curvature (Zevenbergen and Thorne,
1987)
Curvature in an inclined plane (Mitášová and
Hofierka, 1993)
Hill shaded image (Horn, 1981)
Direction of flow calculated using one of seven flow
algorithms
Number of neighbours in a 3 × 3 window that are
of lower elevation
Number of neighbours in a 3 × 3 window that are
of higher elevation
Number of neighbours in a 3 × 3 window that flow
into the centre cell
Maximum elevation drop to a neighbour in a 3 × 3
window
Elevation percentile within a user-specified window
Difference from mean elevation in a 5 × 5 window
Standard deviation of elevations in a 5 × 5 window
Spatial pattern of contributing area from one of
seven algorithms
Areas draining to digitized/inputted points or
stream links
Downslope distance along flowpath to outlet
Maximum flowpath distance to ridge
Minimum flowpath distance to ridge
Average flowpath distance to ridge
Average elevation of contributing area
Average slope of contributing area
Cells that are connected to an edge cell by a
flowpath
Area that can be seen from a point (Franklin and
Ray, 1994)
Spatial pattern of the number of visible cells
(Franklin and Ray, 1994)
Absolute elevation above nearest downslope
depression cell
Elevation relative to nearest upslope peak and
downslope depression cell
Elevation relative to DEM minimum and maximum
Elevation relative to nearest channel and divide cells
Absolute elevation above nearest downslope target
cell
(continued overleaf )
Hydrol. Process. 19, 1123–1130 (2005)
 J. B. LINDSAY

Table I. (Continued )

Submenu Subprogram name Description

Compound Terrain Wetness (topographic) index
Attributes
Stream power index
Sediment transport capacity
Mass accumulation
Network wetness index
Landform classification
Composite relief model
Depression Analysis Find depressions
Depth in sink
Stochastic shape analysis
Find depression in contours
Stream Network Derive network
Analysis
Strahler stream order
Shreve stream magnitude
Channel link classification
Stream index number
Average link slope
Link length
Find main channel
Convert streams to vector
Basin Morphometry Shape and relief indices
Drainage density
Hypsometric analysis
None Drainage pathway analysis
None Derive contours
None Profile
None Flowpath and stream profile
Beven and Kirkby (1979) TOPMODEL index
Spatial pattern of stream power (Moore et al., 1991)
Spatial pattern of erosion/deposition (Moore et al.,
1991)
Mass routed throughout a basin with loading and
efficiency terms
Minimum downslope wetness index along flowpaths
(Lane et al., 2004) and/or areas of possible
overland flow reinfiltration
Landform classes based on the Pennock et al.
(1987) scheme
Combination of any terrain attribute or image with
a shaded relief image
All cells contained within depressions.
Depression depth for each depression cell (Antonić
et al., 2001)
Probability of depression occurrence (Lindsay,
2004)
Depression contours in a vector contour coverage
(Lindsay, 2004).
Stream network derived from contributing area and
slope images
Ordered stream network based on the
Strahler/Horton method
Ordered stream network based on the Shreve
method
Exterior and interior links and source, link, and
outlets nodes
Unique identifier assigned to each link in a stream
network
Average slope of each link in a stream network
Length of each link in a stream network
The main channel in a stream network identified by
examining dominant contributing area at
bifurcations
Raster stream network converted to vector format
Numerous shape and relief indices for basins
Total channel length divided by total basin area
Hypsometric plot and the hypsometric integral
Downslope flowpath delineated from target cells
Vector contour coverage from a specified DEM
Elevation profile corresponding to line segments
digitized onto a displayed DEM
Flowpath drawn onto a displayed DEM from a
digitized point and a longitudinal profile
corresponding to the flowpath

Copyright  2005 John Wiley & Sons, Ltd. 1126 Hydrol. Process. 19, 1123–1130 (2005)
 SOFTWARE REVIEW
or combinations of terrain attributes (i.e. com-
pound attributes). Algorithms involving upslope
or downslope calculations (i.e. those within the
Extended Neighbourhood and Stream Network
Analysis menus) rely on the steepest descent, or
‘D8’, flow-routing algorithm (O’Callaghan and
Mark, 1984) because of the need for unique,
non-diverging flowpaths. Several other algorithms
involving extended neighbourhoods (e.g. many of
the subprograms under the Compound Terrain
Attributes menu), however, allow the user to select
a flow-routing algorithm from a list of possibili-
ties. Although most commercial GIS packages only
implement the D8 flow-routing algorithm, TAS
also offers the FD8 (Freeman, 1991), Rho8 (Fair-
field and Leymarie, 1991), FRho8 (Gallant and
Wilson, 2000), FD8-Quinn (Quinn et al., 1995),
D∞ (Tarboton, 1997), and ADRA (Lindsay, 2003)
algorithms. The ADRA flow-routing algorithm can
extract channel networks during the calculation of
contributing area because it incorporates a chan-
nel initiation model directly into the calculation.
TAS is currently the only GIS that contains this
algorithm.
Common tasks for which TAS has been applied
include landform classification, watershed extrac-
tion, basin morphometric analysis, stream network
extraction and analysis, and wetland delineation.
Table II. General spatial analysis functions of TAS
Menu Function
Pre-processing Interpolation
Depression removal
Flats removal
Spikes removal
Burn in streams
Pixel aggregation
Cropping
Datum transformation
Coordinate transformation
Rectification
Image processing Rotate and reflect
Filter
Contrast stretch
Resample
RGB colour composite
Split RGB colour composite
Copyright  2005 John Wiley & Sons, Ltd.
TAS can classify landforms based on the Pen-
nock et al. (1987) scheme, but the relative land-
scape position subprograms (Table I) and fuzzy
membership facilities (Table II) are also useful for
developing custom landform classifications. TAS’s
watershed extraction sub-program is very flexi-
ble. Watersheds can either be extracted for dig-
itized/specified points or, given a stream network,
can be delineated for sub-basins (i.e. areas drain-
ing to stream links), hillslopes (i.e. areas drain-
ing to either side of links), Strahler basins (i.e.
areas draining to links of each Strahler order), and
Shreve basins (i.e. areas draining to links of each
Shreve magnitude). TAS provides several functions
for basin analysis, including numerous relief and
shape indices and hypsometric analysis. Stream
network analysis includes facilities to order net-
works and derive the associated network statistics
(e.g. number and length of channels per order,
bifurcation ratio, and length ratio), to classify inte-
rior/exterior links and source/link nodes within
networks, and to measure average link slope and
length. A longitudinal profile can be generated
simply by digitizing a point on a displayed DEM.
In addition to the terrain analysis functionality
described above, TAS also possesses extensive spa-
tial analysis capabilities (Table II). The descrip-
tive statistics and raster calculator subprograms
Menu Function
Statistical Analysis Descriptive statistics
Image histogram
Convert to Z-scores
Compare distributions
Image correlation
Cross-tabulation
Generate random field
Fuzzy membership
GIS Analysis Reclass
Overlay
Clump (group)
Spread (distance)
Cost surface analysis
Buffer
Area
Raster calculator
Raster/vector conversion
1127 Hydrol. Process. 19, 1123–1130 (2005)
 J. B. LINDSAY
Figure 2. TAS’s raster calculator
(Table II) are of general interest for hydro-geomor-
phic applications. The descriptive statistic subpro-
gram is sophisticated and can calculate numerous
basic statistics (e.g. mean, minimum, maximum,
diversity, etc.) over user-defined regions. This is
particularly useful for terrain analysis because it
allows statistics to be calculated for multiple sub-
basins within a region or for links within a stream
network (e.g. average basin slope or average chan-
nel slope). TAS’s raster calculator (Figure 2) is
also quite advanced and allows complex calcula-
tions involving both numerical and logical func-
tions within the same statement. Raster calculator
scripts can be used to automate multiple state-
ments that must be executed consecutively. The
raster calculator is one of the most useful subpro-
grams for spatial modelling.
Visualization
The TAS graphical user interface allows multiple
images to be displayed simultaneously (Figure 1),
which greatly facilitates visual analysis of multi-
ple terrain attributes. Displayed images can also
be combined with shaded-relief images to enhance
visualization of terrain. In these ‘composite-relief
models’, variations in colour correspond to the dis-
played attribute and tonal variations correspond
to hill shading. Vector data may be overlaid onto
raster images to enhance data visualization and
interpretation further. Although vector data are
primarily used for display purposes, some TAS
algorithms do perform analysis on vectors.
Copyright  2005 John Wiley & Sons, Ltd.
TAS is distributed with a library of nearly 50
standard colour palettes, many of which have
been created specifically to display topography
and other terrain attributes (e.g. wetness index,
flow direction, and contributing area). Nonetheless,
users will eventually find that they need to create
custom palettes for specific purposes. The palette
manager is a useful subprogram for creating and
modifying colour palettes.
Importing and Exporting Data
DEMs are the main input data to TAS, but
the program does utilize many other types of
spatial data, including satellite imagery. Raster
import/export functions include facilities to read
and write IDRISI images, ArcView raster for-
mats, GRASS images, Surfer grids, and device-
independent bitmaps. The program can also read.
hgt files, the Shuttle Radar Topography Mis-
sion (SRTM) DEM format. TAS reads and writes
IDRISI and GRASS ASCII vector files and deli-
mited XYZ vector point files. Graphical output
(i.e. displayed images with vector overlays) can be
saved as Windows metafiles (.wmf), which can
be read by most graphical packages and several
word-processing programs.
The TAS raster format consists of a single-
precision floating-point binary file (.tas extension)
and an ASCII header file (.dep extension). The
binary file consists of a line for each row in the
image, with numbers ordered from top to bottom
and from left to right. Thus, the TAS binary file
1128 Hydrol. Process. 19, 1123–1130 (2005)
 SOFTWARE REVIEW
format is identical to the ArcView binary floating-
point raster export format. The TAS header file
contains information about the minimum and
maximum values (for display purposes only), the
coordinates of the image’s boundaries, and the
number of rows and columns.
Program Development and Availability
TAS was originally developed to house the special-
ized algorithms needed for my doctoral research
(Lindsay, 2004); however, the capabilities of the
program have been extended considerably past the
scope of the dissertation. TAS operates on IBM
PCs under Microsoft Windows platforms (i.e.
98, 2000, NT, and XP). Hardware requirements
vary depending on the size of DEM or image being
processed, but the program itself requires approx-
imately 6 MB of RAM.
TAS has been made freely available for down-
load from the University of Manchester Geogra-
phy research page: http://www.sed.manchester.ac.
uk/geography/research/tas/. TAS is also housed
on a mirror site maintained by the Catchment
Research Facility (University of Western Ontario):
http://publish.uwo.ca/∼icreed/tas.htm. Some tuto-
rials, which have been used for student instruction
in the past, are also available for download at
these Websites. Program development is ongoing
(the current version is 1·6·6) and the TAS home-
page is updated frequently. The author welcomes
comments and feedback.
Concluding Remarks
TAS marries the visualization and spatial anal-
ysis capabilities commonly found in raster GIS
packages with an extensive list of subprograms
specifically designed for research in hydrology
and geomorphology. Although TAS is a powerful,
research-grade environmental modelling program,
its ease of use and free availability also make
it ideal for student instruction. TAS is suited to
research and instruction in physical geography,
hydrology, geomorphology, climatology, environ-
mental science, and watershed modelling. Since
TAS was posted to the Internet in September
2003, the program has been downloaded in at
least 19 countries and is becoming widely used by
researchers.
Copyright  2005 John Wiley & Sons, Ltd. 1129
Acknowledgements
I would like to thank Irena Creed and Paul
Bates for their valuable comments and assistance
in preparing this review. Funding has been pro-
vided by a Natural Sciences and Engineering
Research Council of Canada (NSERCC) Postgrad-
uate Scholarship and the Ontario Graduate Schol-
arship (OGS) programme.
References
Antonić O, Hatic D, Pernar R, 2001. DEM-based depth in
sink as an environmental estimator. Ecological Modeling 138:
247– 254.
Beven KJ, Kirkby MJ, 1979. A physically based, variable con-
tributing area model of basin hydrology. Hydrological Sciences
Bulletin 24: 43–69.
Conrad O, 2002. Digitales Gelände-Modell (DiGeM) Terrain Anal-
ysis Software. http://www.geogr.uni-goettingen.de/pg/saga/
digem/ (15 August 2004).
Fairfield J, Leymarie P, 1991. Drainage networks from grid dig-
ital elevation models. Water Resources Research 27(5): 709–717.
Florinsky IV, Eilers RG, Manning GR, Fuller LG, 2002. Predic-
tion of soil properties by digital terrain modelling. Environmental
Modelling & Software 17: 295–311.
Franklin WR, Ray CK, 1994. Higher isn’t necessarily better:
visibility algorithms and Experiments. In Advances in GIS
Research: Sixth International Symposium on Spatial Data Han-
dling, Waugh TC, Healey RG (eds). Taylor & Francis: Edin-
burgh; 751–770.
Freeman TG, 1991. Calculating catchment area with divergent
flow based on a regular grid. Computers & Geosciences 17(3):
413– 422.
Gallant JC, Wilson JP, 1996. TAPES-G: a grid-based terrain
analysis program for the environnemental sciences. Computers
& Geosciences 22(7): 713–722.
Gallant JC, Wilson JP, 2000. Primary topographic attributes. In
Terrain Analysis: Principles and Applications, Wilson JP, Gal-
lant JC (eds). John Wiley: New York; 51–85.
Horn BKP, 1981. Hill shading and the reflectance map. Proceed-
ings of the IEEE 69(1): 14– 47.
Hutchinson MF, 1989. A new procedure for gridding elevation
and stream line data with automatic removal of spurious pits.
Journal of Hydrology 106: 211–232.
Lane SN, Brookes CJ, Kirkby MJ, Holden J, 2004. A network-
index-based version of TOPMODEL for use with high-resolution
digital topographic data. Hydrological Processes 18: 191– 201.
Lindsay JB, 2003. A physically based model for calculating
contributing area on hillslopes and along valley bottoms. Water
Resources Research 39(12): 1332. DOI: 10·1029/2003WR002576.
Lindsay JB, 2004. Coping with topographic depression in digital
terrain analysis. PhD thesis, University of Western Ontario.
Hydrol. Process. 19, 1123–1130 (2005)
 J. B. LINDSAY
Lindsay JB, Creed IF, 2005. Removal of artifact depressions
from DEMs: towards a minimum impact approach. Hydrological
Processes 19: in press.
Mitášová H, Hofierka J, 1993. Interpolation by regularized spline
with tension: II. Applications to terrain modeling and surface
geometry analysis. Mathematical Geology 25(6): 657– 669.
Moore ID, Grayson RB, Ladson AR, 1991. Digital terrain mod-
eling: a review of hydrological, geomorphological, and biological
applications. Hydrological Processes 5: 3–30.
O’Callaghan JF, Mark DM, 1984. The extraction of drainage
networks from digital elevation data. Computer Vision, Graphics,
and Image Processing 28: 323– 344.
Pennock DJ, Zebarth BJ, deJong E, 1987. Landform classifica-
tion and soil distribution in hummocky terrain, Saskatchewan,
Canada. Geoderma 40: 297– 315.
Pike RJ, 2000. Geomorphometry—diversity in quantitative sur-
face analysis. Progress in Physical Geography 25(1): 1–20.
Pike RJ, 2002. A bibliography of terrain modeling (geomorphom-
etry), the quantitative representation of topography— Supplement
4 ·0 . US Geological Survey, Open- file Report 02–465; 158.
Copyright  2005 John Wiley & Sons, Ltd. 1130
Quinn PF, Beven KJ, Lamb R, 1995. The ln(a/tanb) index: how
to calculate it and how to use it within the TOPMODEL frame-
work. Hydrological Processes 9: 161–182.
Tarboton DG, 1997. A new method for the determination of flow
directions and upslope areas in grid digital elevation models.
Water Resources Research 33(2): 309–319.
Wilson JP, Gallant JC, 2000. Digital terrain analysis. In Ter-
rain Analysis: Principles and Applications, Wilson JP, Gallant JC
(eds). John Wiley: New York; 1–27.
Wood J, 1998. Modelling the continuity of surface form using
DEMs. In The 8th International Symposium on Spatial Data
Handling, Vancouver, Canada, 11–15 July.
Zevenbergen LW, Thorne CR, 1987. Quantitative analysis of
land surface topography. Earth Surface Processes and Landforms
12: 47–56.
Hydrol. Process. 19, 1123–1130 (2005)