USING LANDSAT THEMATIC MAPPER DATA TO DETECT AND MAP 
VEGETATION CHANGES IN KUWAIT 
Andy Yaw Kwarteng and Dhari Al-Ajmi 
Kuwait Institute for Scientific Research 
Remote Sensing Group, Environmental and Earth Sciences Division 
P.O. Box 24885 
13109 Safat, Kuwait 
Fax: 965-481-5202 
Email: andyk@sunpci.kisr.edu.kw 
Commission VII, Working Group 6 
KEY WORDS: Vegetation, Change_Detection, Mapping, Processing, Landsat. 
ABSTRACT 
Remotely sensed satellite images are invaluable data sets for land-cover analysis on either local, regional or global scales. Equally 
important is to employ digital enhancement techniques sensitive to minor vegetation changes, especially in areas with low 
vegetation cover. In this study, Landsat Thematic Mapper (TM) digital image data acquired on February 4, 1987, and February 28, 
1993, respectively, were geometrically and radiometrically calibrated and used to detect and map vegetation changes in the desert 
environment of southern Kuwait. The normalized difference vegetation index (NDVI), an index related to photosynthetically active 
green biomass, was used as input to a selective principal component analysis procedure. The change image or the second principal 
component (PC2) mapped 19.82% of total information related to vegetation between the two dates. The increase in vegetation 
between the 1993 and 1987 images was supported by recorded rainfall data that increased by about three times for the same time of 
the year for the two dates. Vegetation increase was observed mainly within the Kuwait City limits and cultivated farms where 
plants are irrigated. The majority of the desert areas with active and smooth sand sheets displayed insignificant or no vegetation 
changes between the two dates. 
1. INTRODUCTION 
Mapping of land-cover from space offers an unequaled and 
inexpensive technique to monitor vegetation changes on local, 
regional and global scales. The principal image sensor for 
global or broadscale land-cover analysis has been the 
Advanced Very High Resolution Radiometer (AVHRR) 
onboard National Oceanic and Atmospheric Administration 
(NOAA) series of polar orbiting meteorological satellites. 
Several studies using NOAA AVHRR data to monitor 
vegetation cover in diverse environments have been reported in 
the literature (Townshend and Justice, 1986; Choudhury and 
Tucker, 1987; Goward et al., 1991; Justice et al., 1991). The 
popularity of AVHRR is due to the availability and the 
relatively inexpensive data sets. However, a major 
disadvantage is the large pixel resolution of 1.1 km per pixel. 
For regional and detailed local studies, Landsat Thematic 
Mapper (TM) and Satellite Pour d'Observation de la Terré 
(SPOT), with spatial resolutions of 30 m and 10-20 m, 
respectively, have provided invaluable information related to 
land-cover analysis (Dymond et al., 1992; Lauver and Whistler, 
1993; Schriever and Congalton, 1995). 
Desertification in an arid environment, such as Kuwait, entails 
land degradation, whereby desert scrubs are gradually replaced 
by sand and/or bare soil. The process could be natural/climatic 
induced or anthropogenic. The use of remote sensing and other 
methods to map seasonal and yearly vegetational changes are 
critical to studying desert dynamics and could help in the 
recommendation of the best techniques to combat 
desertification. Such techniques facilitate the demarcation of 
desert boundaries and areas threatened by desertification, 
398 
mapping of arable lands, as well as areas susceptible to erosion 
and generation of sand and dust storms. Vegetation, 
undoubtedly, is the single most effective and environment 
friendly way to combat desertification, stabilize soils and 
reduce wind erosion in arid environments. In arid an 
environment with low vegetation cover, it is imperative to use 
sophisticated digital enhancement techniques so that satellite 
images can map minute seasonal and inter-annual vegetation 
variations. A method that is sensitive to slight vegetation 
changes would be very important and serve as an early warning 
for areas under threat from desertification. Remote sensing 
analysis of land-cover will continue to provide more useful 
information, especially with the launching of satellites with 
higher spatial resolution of 1-3 m by 1997 (Fritz, 1996). 
Described herein is a study that utilized multitemporal Landsat 
TM bands to detect and map vegetation changes in the arid 
environment of Kuwait. The study area was extracted from 
Landsat TM digital image data of path 165, row 40, that were 
acquired on February 4, 1987, and February 28, 19%, 
respectively. The images were selected to include the principal 
types of vegetation, namely, undershrubs, perennial shrubs and 
spring ephemerals. The normalized difference vegetation 
index (NDVI) images were used as an input into an automatic 
change detection procedure that mapped vegetation differences 
between the two dates. 
The study was conducted in the southeastern part of the State 
of Kuwait, extending south of Kuwait City to the Kuwait-Saudi 
Arabian border. Like most parts of the Arabian Shieló, 
Kuwait is characterized by a desert type of environment with 
scanty rainfall, and a dry, hot climate. Summer is very hot, 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B7. Vienna 1996 
ua. A o wx "m- ma omma nud. Lut — CÓ hxl JC 5 lano uua Aat 
p RN OS S Jue ÉPgà EAN dae — paa CL SAR