sensing is called the normalized burn ratio (NBR), computed as 
the difference between near-infrared (NIR) and middle-infrared 
(MIR) reflectance divided by their sum (Roy et al, 2006). 
Though the near-infrared wavelengths in colour infrared 
photographs are useful in mapping vegetation mortality and soil 
moisture, it has been shown that mid-infrared bands (band 5 = 
1.55-1.75 Am and band 7 = 2.08-2.35 Am) of Landsat 
Thematic Mapper (TM) contribute new information for 
classifying burn severity (White et al, 1996; Miller and Yool, 
2002). Most researchers agree that fire severity is a measure of 
the amount of soil organic matter lost due to burning, decrease 
in surface cover, and volatilization or transformation of soil 
components to soluble mineral forms (Wells & Campbell, 1979, 
cited by Miller and Yool, 2002). Research by Clark (2000, cited 
by Miller and Yool, 2002) has shown that pre- and post-fire 
differences of mid-infrared to near-infrared ratio (TM band 
7/TM band 4) provided the highest contrast fire scar in 
comparison to TM band 4, PCA, Kauth—-Thomas, NDVI, and 
MSAVI and, therefore provides the best enhancement for 
classifying changes due to fire. NBR values range between -1 
and 1 as does the NDVI (Miller and Yool, 2002). 
The capabilities of remote sensing and geographic information 
science (GIS) for mapping and monitoring changes in land are 
burgeoning (Alo and Pontius Jr, 2008). These techniques have 
been used increasingly as tools to measure the spatial and 
temporal patterns of land cover in many regions of the world 
(e.g. Musaoglu et al, 2005; Yuan et al, 2005 in Alo and Pontius 
Jr, 2008; Kusimi 2008; Yiran et al, 2011). However, mere 
monitoring of the land-cover changes does not necessarily shed 
light on the underlying causes of the land transformations. 
Scientists and policy makers need to understand the 
fundamental processes of land transformation in order to 
establish effective conservation and management strategies ( Alo 
and Pontius Jr, 2008). In this regard, the study was undertaking 
using remote sensing and social survey techniques to access the 
causes and socio-economic and environmental effects of the 
annual bushfires on the people and the vegetative cover. 
2. STUDY AREA 
The study was conducted in the Krachi District which is 
located in the northern part of Volta Region in Ghana between 
latitude 7°20' - 8°20'N of the Equator and longitude 0°15'W — 
0°24'E of the Greenwich Meridian. The District falls within two 
climatic regions, the tropical continental in the north and the 
wet semi-equatorial climate in the south with a mean annual 
rainfall of 115 — 150 cm. Mean monthly temperatures range 
from 24°C in August to about 36°C in March. The southern part 
of the District is covered by the wet semi—deciduous forest 
whiles the northern sector the vegetation is the savannah 
grassland (Dickson and Benneh, 1995). The main drainage 
system in the district is the Volta Lake with a number of rivers 
draining into it (Fig.1). The lake covers the western part of the 
map and also partitions the district primarily into the Kete 
Krachi arm to the west and the Dambai arm to the east. 
The District has experienced a steady growth in population 
from 29,980 (1970) to 65,530 (2000) (Ghana Statistical Service, 
2002). Over 80% of inhabitants engage in agriculture which 
includes farming, fishing and pastoralism. Generally, farming is 
on subsistence level and crops grown include cereals like maize, 
sorghum and root crops such as yam and cassava. The 
environmental problem of the farming activity is shifting 
cultivation and bush fallowing which employ the slash and burn 
40 
approach in preparing land for cultivation. Lands are often set 
ablaze by farmers to minimise the cost of preparing lands. 
9200w oo too COTE O'Z00E 
dr 
Inset Ghana Map Showing Study Area 
Legend 
9 Sampling Sites 
vo River 
7777 Read 
Lake 
Boundary 
Volta Region | my a pope 
  
U290W 
Fig.1: Map of the Study Area 
3. RESEARCH MATERIALS AND METHODS 
Field data collection involved interviews with opinion leaders 
such as chiefs, queen mothers, assemblymen and women, 
District Assembly staffs, Fire Service and National Disaster and 
Management Organization (NADMO) (e.g. Gangemi et al, 
2003). A total of 102 respondents were interviewed in 17 towns 
and Table 1 illustrates the number of respondents selected from 
each community and Fig.1 shows the spatial distribution of the 
communities. The District was divided into East Krachi and 
West Krachi and a simple random sampling method without 
replacement was used to select the communities (e.g. Kusimi 
and Yiran, 2011). The number of people selected in each 
community was a function of population size. Focus groups 
discussions were also organized among farmers and Fulani 
herdsmen (e.g. Gangemi et al, 2003). Proceedings were 
recorded and later transcribed to augment information from 
interviews. 
Satellite images employed for the determination of change in 
vegetative cover/fire severity are Landsat ETM+ of 2002 and 
2003. These were the only free downloadable scenes of the 
study area. The study area falls into two scenes of row and path 
054/193 and 055/193. The images were downloaded from the 
Global Land Cover Facility website www.glcforg and 
glovis.usgs.gov. Landsat ETM+ 2002 images were acquired on 
Table 1: Nu 
Administra 
Town 
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
10" Decer 
the images 
acquisitior 
cloud cov 
cropped t: 
area. Land 
and Land: 
burn ratio 
following 
Yool, 200 
_ba 
NBR = Da 
Afterward 
image (Fi 
from the 
denoted a: 
With in 
wildernes: 
infrastruc 
sheds and 
or even ce 
at risk (M 
4.1. Caus 
One of 
District a 
and there 
practice i