nbul 2004 
sults 
ze are the 
nd Spatial 
ich to the 
| Images, 
se Change 
Area,Geo- 
: Active 
compute 
GIS in 21 
f Wuhan 
tection of 
rchives of 
formation 
1antitative 
lonitoring 
rammetric 
hematical 
Xf. Spatial 
gration of 
RSS 97 
he Israeli 
's of the 
formation 
detection 
"Road- 
stem for 
insactions 
IS. 
;graphical 
CHANGE-DETECTION IN WESTERN KENYA — THE DOCUMENTATION OF 
FRAGMENTATION AND DISTURBANCE FOR KAKAMEGA FOREST AND 
ASSOCIATED FOREST AREAS BY MEANS OF REMOTELY-SENSED IMAGERY 
T. Lung, G. Schaab * 
Department of Geoinformation, Karlsruhe University of Applied Sciences, Moltkestr. 30, D-76133 Karlsruhe 
Germany, gertrud.schaab@fh-karlsruhe.de 
KEY WORDS: Remote Sensing, Multisensor Imagery, Landsat, Classification, Land Cover, Change-Detection, GIS 
ABSTRACT: 
In order to understand causes and effects of disturbance and fragmentation on flora and fauna, a time series on land cover change 1s 
needed as basis for the BIOTA-East Africa project partners working in western Kenya. For 7 time steps over the past 30 years Land- 
sat data were collected for Kakamega Forest and its associated forest areas. Preprocessing involved georeferencing and radiometric 
corrections. In a first step the time series is evaluated via a threshold analysis distinguishing between “forest” and “non-forest”. Even 
though a temporally changing pattern of forest losses and replanting is observed, in total no major change in forest-covered area is 
revealed. Therefore, a DAN multispectral classification is performed distinguishing between classes at the ecosystem level. 
Ground truthing for the historical imagery is done with the help of maps showing vegetation types or land cover. Actual land cover 
verification is based on amateur photographs taken from an aeroplane as well as on terrain references. For classification the maxi- 
mum-likelihood decision rule is applied considering bands 3, 4, 5, 7 plus 7/2 for TM/ETM+ imagery and |, 2, 3 and 4 for MSS-data, 
repectively. If available, scenes from both the rainy and dry seasons are made use of. From planned 17 land cover classes 12 can be 
realized, of which 6 belong to forest formations. A shortcome is that plantation forest of Maesopsis eminii (planted mixed in with 
other indigenous tree species) cannot be separated. Nevertheless, the classification results form a solid basis for a consistent and 
detailed evaluation of forest history between 1972 and 2001. Analyses presented include graphs of change in land cover class areas 
over time as well as such allowing for true change detection with transitions between the different classes. 
I. INTRODUCTION 2. THE STUDY AREA 
Detecting land cover and its changes is needed in order to un- 
derstand global environmental change. Of particular interest are 
the effects of tropical rain forest fragmentation which varies 
regarding intensity and extent. The increase in fragmentation is 
related either to natural or human sources (e.g. WADE et al., 
2003; Geist & Lambin, 2001). Both influence biodiversity, 
which is known to be immensly rich in tropical forests (Gaston, 
2000). In general, the BIOTA (Biodiversity Monitoring Tran- 
sect Analysis in Africa) project, funded by the German Ministry 
of Education and Rescarch (BMBF), analyses changes in Afri- 
can biological diversity related to land cover and environmental 
changes with the aim to develop recommendations for a sus- 
tainable biodiversity management (see www.biota-africa.org). 
The research acticities of BIOTA-East Af- 
Kakamega Forest is known to be the species-richest forest in 
Kenya. It inhabits a large number of rare animals and even some 
endemic plant species (KIFCON. 1994). The forest is sour- 
rounded by small forest fragments which might have been 
connected to the main forest in former times. The Nandi Forests 
are of comparable size but placed on an escarpment some 200 to 
300 m higher in elevation. There has been an ongoing debate 
wether all these left forest islands have once formed one larger 
forest area if not even have belonged to the congo-guinean rain 
forest belt (see e.g. Kokwaro, 1988). The forests are placed in 
one of the world's most densely populated rural areas (Blackett, 
1994: mean population density of 600 inh./km?) which is inten- 
sily used for subsistence agriculture. Due to continualy increas- 
  
    
  
  
  
  
  
rica (see Kóhler, 2004) are currently focus- 
ing on Kakamega Forest and fragments 
close by, located in western Kenya. Informa- | > thon [| Kakamega Forest and Associated Forest Areas | 
tion retrieval regarding land cover change | e LT meo di 
and thus forest fragmentation and distur- v Suis X 
bance is performed for the wider area of aet. e bis. Forest Y. nom 
Kakamega Forest and associated forest areas =" Kenya | 4 is { se 
(see Figure 1). Here, the results of satellite T | raramea @ CS 
data analysis form a major contribution of front | «meme CE es À ei 
geo-information processing for documenting SS | c AE C A 
anthropogenic influences on these rain da ramon e, Poen 
forests over time. They will allow the differ- | À 5 UE 
ent biological and/or ecological subprojects | S RIO o É MEM B Soph andi ; 
of BIOTA-East Africa to conclude on | S. y S LÁ) 
changes in forest ecosystems or biodiversity | eee vo 7 
changes in forest ecosy 0 y V 
related to global change. With the study sites Ws = / 
Spread over a relatively small area there is Figure 1: Location of the study area and the different forests covered by remote 
the need and the chance to offer rather de- 
tailed information, both in space and time. 
sensing analyses in western Kenya (34?3775" — 3579725" east of Gr. 
0°32°24” north - 0°2°52” south of the equator) 
  
165 
* corresponding author