[
] 68
M
OST GEOLOGICAL HAZARDS
in Cameroon are linked to
the existence of the Cameroon Volcanic Line (CVL), a
1,600km long chain of Cenozoic-Present volcanoes
that traverses the country in a SW-NE direction from the Gulf
of Guinea island of Annobon across Cameroon to eastern
Nigeria. Such hazards include toxic gas emissions from volcanic
crater lakes, landslides, floods and hazards from volcanic erup-
tions.
In the twentieth century and early twenty-first century, the
country recorded a series of natural disasters located within and
around the corridors of the CVL. These include:
• The Lake Manoun toxic gas disaster (1984) in which 37
people were killed
• The Lake Nyos toxic gas disaster (1986) which claimed
approximately 1,700 to 2,000 lives
• The Bafaka Balue landslide (1995) with 3 lives lost
• The Mount Cameroon volcanic eruption (1999) where lava
consumed an enormous amount of forest (about 800
hectares) with its rich bio-diversity and cut off 83m of tarred
road, disrupting economic activities.
Limbe subdivision is located between latitude 3
o
90’ and 4
o
05’N and longitude 9
o
29’ and 9
o
06’E in the southwestern part
of Cameroon. This is a coastal region with an approximate 50.5
km of coastline. The subdivision consists of more than 25
villages with the city of Limbe, the capital.
The project
Over the years, this region has experienced persistent landslides
and floods with the worst being in June 2001, taking some 30
lives, rendering over 2,000 people homeless, and destroying
properties and social amenities such as roads and telephone
lines worth millions of CFA francs. The occurrence of disasters
in this region varies periodically, resulting from oceanographic
processes, heavy rains and seismic activity from MC, particu-
larly lava flows leading to fire outbreaks from volcanoes.
Delineating the potential risk zone is an attempt to provide deci-
sion makers and planners with geospatial information on areas
where they need to concentrate concerted effort in order to
develop mitigation plans.
Geospatial tools were employed in the execution of this
project for data collection and analysis. Extensive use was made
of existing feature datasets considering their source, quality, and
date of publication. Landsat 7 ETM+ satellite imagery was also
used extensively for categorizing, interpreting and digitizing
features. Unfortunately, only the April 2003 image was cloud-
free and therefore was put to full use for this project. Images
were orthorectified and georeferenced by EarthSat to its
GeoCover dataset, a commonly used and available system.
1
Image interpretation
The ETM+ scanner measures the sun’s reflectance in seven
spectral channels known as bands. These include the visible
spectrum (bands 1-3), near infrared (bands 4 and 5), thermal
(band 6) and medium infrared (band 7). These bands were
selectively combined into colour-composite images to enhance
visibility. Different band combinations were used to maximize
the visibility of landslide scars and different land uses under a
range of conditions. To improve the detection of narrow land-
slides, scars and less prominent features, the panchromatic band
(14.5m resolution) was merged with the other bands (28.5m
resolution) to create a ‘pan-sharpened’ image. This greatly
improved the visibility of older landslide scars before catego-
rization and digitization.
Topographic map update
A topographic map update was performed by image analysis
and manipulation. Pixel transformation was carried out by
computing the normalized difference vegetative index (NDVI) to
improved visual interpretation of enhanced false colour compos-
ite enabling discrimination between different vegetations and
non-vegetative surfaces. This visual enhancement of the image
permitted easy digitization.
The topographic map of the subdivision was updated with
focus on new roads and land use/land cover patterns. Based on
an accumulated pattern of NDVI, four main groups were cate-
gorized:
1) Non-vegetative areas (settlement, bare ground or outcrop,
road)
2) Natural vegetative area (evergreen forest, deciduous forest)
3) Agricultural area (CDC plantations, farmland etc.)
4) Water surfaces.
The Cameroon Development Corporation (CDC) has colossal
palm and rubber plantations in the subdivision, accounting for
over 30 per cent of total land use. The landscape has undulat-
ing relief with heights varying from 2m to 362m above sea level.
Most human settlements are around the coast, with a few
villages situated above 200m. Low-lying areas adjacent to the
sea include Mbojo, Mutowoh, Wovia, Ngeme, Man’O War Bay,
Sonara Bota-Land and a host of other villages on the west coast
of the subdivision. Areas around the city of Limbe with remark-
able hills include Towel, Miletwo, Coconut Island, Mabetal New
Layout, Ambas Bay, and Mount Etindi. Over the years the city
Delineation of potential risk zones,
Limbe subdivision, Cameroon
Buh Wung Gaston, GIS and Remote Sensing Department, Limbe Botanic Garden, Cameroon




