The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B7. Beijing 2008
wide-range of applications and this development has been
identified by professionals in the geo-information arena as a
significant landmark that will tremendously revolutionize
medium-scale topographic mapping. The near-global SRTM
digital elevation model (DEM) product was processed and
compiled at a resolution of 90m by the Consultative Group for
International Agriculture Research Consortium for Spatial
Information (CGIAR-CSI) and hosted on a Web portal for free
public access and download. Although this product presents
attractive promise for terrain analysis for an impressively wide
range of applications, several researchers have proposed a
thorough evaluation of its vertical accuracy. For instance
Gorokhovich and Voustianiouk (2006) in their study to assess
the accuracy of the SRTM-based elevation data noted that an
overall assessment of the accuracy of the product requires
additional regional studies involving ground truth control and
accuracy verification methods with higher level of precision.
Both the vertical accuracy of the product and its applicability in
different contexts have been extensively investigated in many
studies (Koch, A. and Lohmann, P., 2000; Miliaresis, G. and
Paraschou, C. V. E., 2005: Kocak G. et al, 2005: Koch, A. et al,
2005, Kleusberg, A. and Klaedtke, H. G., 1999). The major
objectives of these studies have been to:
• determine the absolute and relative vertical accuracies
of the SRTM digital elevation data for defined study
sites;
• determine the relationship between SRTM vertical
errors and certain topographic derivatives from the
SRTM DEM such as slope and aspect;
• evaluate the impact of spatial structure on the
accuracy of contour maps derived from the SRTM
DEM;
• investigate the effects of geostatistical processing
models on the SRTM DEM derivatives.
The enormous attention directed towards regional-level
assessment of SRTM vertical accuracy suggests that extracting
topographical data by contour interpolation from SRTM
elevation data necessitates accuracy assessment of the
interpolation result to ascertain its suitability for digital
topographical mapping especially at a scale of 1/25,000. In this
paper, we present a framework for accuracy assessment of
interpolating contours from 1:50,000 topographical maps and
SRTM height data for topographical mapping at the scale of
1:25,000. Quantitative statistical and geostatistical tests were
performed on the two spatial data sources for different terrain
configurations and contexts to determine their suitability for
topographical mapping in different scenarios. In particular, this
study consisted of:
(1) measuring the vertical accuracy of the DEM derived
from the 1:50,000 topographic map and that of the
90-m resolution CGIAR-CSI SRTM digital elevation
data against higher precision GPS measurements
within the same site;
(2) interpolating digital elevation models from an
existing topographic map covering the same area and
comparing measurements from the two sources;
(3) implementing a processing strategy to minimize
errors emanating from contour interpolation using
SRTM data as a base.
Using an on-going 1:25,000 state-wide topographical mapping
project undertaken by one of the State Governments of Nigeria
as a case study, we demonstrate the applicability of the
proposed contour interpolation accuracy assessment model for
arriving at the best strategy for the topographical mapping
process. The rest of the paper is organised as follows. Section 2
describes the site chosen for this study. Section 3 is dedicated to
a description of the materials and the methodology adopted for
the study. Section 4 presents a discussion of the results of the
GIS analysis and geostatistical processing methods adopted
while Section 5 concludes the paper.
2. STUDY SITE
The site chosen for this study lies between Latitudes 5°30’ E
and 5° 45’ E and Longitudes 7° 15’N and 7°30’N. The site is
situated in Ondo State, a state in the South-Western part of
Nigeria. Figure 1 shows a digital elevation model (DEM) of the
site presented as a color-coded image. The chosen site
corresponds to the area covered by one map sheet at the scale of
1/50,000 (27.8Rm X 27.5Km). The site covers low-level, mid
level and high-altitude terrain with elevations ranging from
182m to 594m above mean sea level. The mean elevation of the
site is 284m above mean sea level. As depicted in Figure 1, the
South-Eastern part of the area is generally low-lying with
elevations ranging from 182m above mean sea level to 285m
above mean sea level. The North-Eastern part is characterized
by rugged terrain interspersed with hills having elevations
ranging from 490m to 594m above mean sea level. The rest of
the area is generally at mid-level altitudes with elevations
ranging from 285m to 388m above mean sea level. The entire
area is well-drained with a good network of rivers and streams
all flowing Southwards. The predominant vegetation in this area
is of tropical rain forest type with dense canopy cover in some
areas and low, widely separated trees in others. The hilly areas
present exposed rock surfaces in some areas and low grasses
and shrubs in others. This site was deliberately chosen by
reason of the fact that it presents different topographic
conditions (from low to high terrain characteristics) for the
purpose of conducting statistical and geostatistical analyses
presented in this study.
£t «vatio» Voily Ar»«
Figure 1. The study site as oolor-coded Digital Elevation Model
3. DATA, MATERIALS AND METHODS
3.1 Source Data
The accuracy tests conducted in this study employed three
major sources of spatial data (90-m resolution CGIAR-CSI
SRTM digital elevation data, 1/50,000 topographical map of the
study site and GPS point data randomly distributed over the
study site).
1348