See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/351067234

Flood Risk Mapping and its Effects on livelihood in Ghana using Sentinel-1 data.
A case study of Accra

Article · January 2021

CITATION READS

1 370

4 authors:

Elisha Njomaba James Nana Ofori

Czech University of Life Sciences Prague Universität Potsdam
8 PUBLICATIONS 3 CITATIONS 9 PUBLICATIONS 6 CITATIONS

SEE PROFILE SEE PROFILE

Ben Aikins Davina Nyame

University of Ghana University of Ghana
10 PUBLICATIONS 9 CITATIONS 1 PUBLICATION 1 CITATION

SEE PROFILE SEE PROFILE

All content following this page was uploaded by James Nana Ofori on 23 April 2021.

The user has requested enhancement of the downloaded file.

International Journal of Research and Scientific Innovation (IJRSI) | Volume VIII, Issue I, January 2021 | ISSN 2321–2705

Flood Risk Mapping and its Effects on livelihood in

Ghana using Sentinel-1 data. A case study of Accra
1
E. Njomaba ∗, J. N Ofori2, B.E. Aikins3 D. A. A Nyame4
1
Laboratory of Geo-information Science and Remote Sensing, Wageningen University & Research, Gelderland, Netherlands
2
Dept. of Plant Ecology and Nature Conservation, Wageningen University and Research, Droevendaalsesteeg 3, Wageningen,
Gelderland, Netherlands
3
Remote Sensing & GIS Unit, Bedeston Technologies, and Consulting Services, Accra, Tema, Ghana
4
Department of Geography & Resource Development, University of Ghana, Legon

Abstract: Flooding has been a major issue in Ghana, and Flooding has been a major issue in Ghana, especially the
several driving factors have been attributed to flooding in capital city, Accra since the early 1930s [4] with significant
Accra. Among the factors outlined throughout the literature floods recorded in 1955, 1960, 1963, 1973, 1986, 1995,
are climatic influences such as rainfall which results in a 1999, 2001, 2002, 2010, 2011, and 2015 (IFRC, 2016).
change in the natural hydrology of the area. Other causes
Flooding in Ghana has caused massive destruction since
include building in waterways and dumping of refuse in the
drainage area. This paper seeks to develop an operational 1995 when the first devastating floods even occurred
methodology using Synthetic Aperture Radar images to (Twumasi & Asomani 2002). The July 2015 floods
support the response agencies by providing timely affected 46,370 people, 187 houses were partially
information on flood risk areas over a period to prioritize destroyed, and 300 people died (IFRC, 2016). The
response activities within the Greater Accra Metropolitan estimated cost of property destroyed was about $55 million
Area (GAMA) area. The method used was a change detection [7].
algorithm where the after-flood mosaic was divided by the
before-flood mosaic, resulting in a raster layer showing the [9] have determined the possible role of climate change in
degree of change per pixel. Additional Dataset to refine the Accra’s flooding. This they say is because of changes in
flood layers were adopted such as Digital Elevation Model rainfall and temperature coupled with changes in climatic
(DEM) and Global Surface-water dataset. One of the major conditions that result in a change in the natural hydrology
findings of this study is that the Ga south district is the of the area through heavy precipitation and increasing peak
district with the highest average flooding extent compared to run-off discharge [19].
the other districts within GAMA and the main reason was
because of the Weija Dam being situated close to the district. To prevent the occurrence of natural hazards such as floods,
When the Dam is full and being spilled over, it causes Remote Sensing and Geographic Information System
surrounding environments to experience flooding. The technologies are being used as effective tools for risk
implication is that much attention should be given to the Ga assessment and hazard management. This technology
south district to curb the menace of flooding in GAMA.
provides effective and faster analysis than conventional
Keywords: Digital Elevation Model, Ground Range Detection, survey methods. Remote Sensing systems are often used
Interferometric Wide Swat, Multicriteria Decision Approach, for flood risk mapping and has offered different levels of
Synthetic Aperture Radar. accuracy [15]. However, since optical observations are
I. INTRODUCTION limited in their validity by clouds, which are most common
during floods, microwave SAR systems are preferred

O ver the past decades, flooding has become a global

issue, which affects economic and social
development. This global issue has in most cases led to the
because they provide more efficient analysis due to their
capabilities of penetrating the atmosphere [15].

loss of lives and damages in most countries including Most research into flood risk mapping has been based on a
Ghana. Aside from demographic growth, urbanization Multi-criteria decision approach using several parameters
trends, and climate change, the causes of floods are such as DEM, Slope, Aspect, soil, etc.
shifting, and their impacts are being felt. Most flooding In a study conducted by [17], a special MCDA (Multi-
happens in urban areas when there is prolonged rainfall and criteria Decision Approach) was developed, implemented,
water drainage facilities can no longer contain the surface and executed to identify the potentially flood-prone areas
running water. Flooding is triggered by a lot of factors. through creating maps of five classes of flood vulnerability
Floods which have serious environmental and social from very low to very high flood-prone zones.
implications have been mostly caused by a climate-related
factor such as severe rainfall and in some cases severe Other studies have also demonstrated flood risk monitoring
snowmelt as well (CRED, 2015). by generating inundation maps from hydrological models

www.rsisinternational.org Page 214

International Journal of Research and Scientific Innovation (IJRSI) | Volume VIII, Issue I, January 2021 | ISSN 2321–2705
2321

which requires DEM as well as computing infrastructure to the GAMA that experienced high risk of flooding over the
model the effects of obstacles on the flow of floodwater in study period.
flood plains [14].
II. MATERIALS AND METHODS
The contribution of this research is seen in a way that the
2.1 Study Area
use of satellites for monitoring flood risk overcomes the
challenges provided by the hydrologically based model. The Greater Accra Metropolitan area is located between
Most approaches in mapping flooding in Ghana have relied 505′27″ to 5028′2″north latitude and 0045′8″east longitude
on multi-criteria analysis using GIS-based
based models. This has to 0037′3″ west longitude along the Atlantic coast of Ghana
not been able to provide a spatial-based
based analysis of the (Figure 2). The size of the study area is approximately
situation. Optical images provide a more accurate way of 1585km2. The region is made up of 12 administrative
mapping flood risk zones. divisions. The
he area also has a population of about
4,000,000 (World Population
pulation Review, 2019).
Although optical images have great potential for mappmapping
during good weather conditions with image analysis
The topography of the study area varies from flat to gently
capacity and accuracy, their use is limited in Accra due to
undulating slopes that rise to 75m at the foothills with a
high cloud coverage during the rainy season. Since clouds
few isolated hills and rocks. In terms of climate, the region
always occur during the rainy season, n, the utilization of
lies in the dry equatorial climatic zones and it has
h two rainy
satellite optical data is infeasible in providing flood risk
seasons. The first starts in March and ends in July, while the
maps during the disaster.
second starts in September and ends in November. The
The main objective of this research was to use sentinel-1 rains are most intensive and normally cause floods in most
data to map flooding risk areas and their effects on parts of the region. Mean monthly
onthly temperature ranges from
livelihood in the GAMA of Ghana. Damage assessment
assessments 24.70C in August to 330C in March, with an annual average
on livelihoods such as urban areas
as and population were of 26.80C. Coastal savannah shrubs interspersed with
made. Specific objectives include (1) Mapping the extent thickets mainly covers the area. Grasses are short and grow
of Flooding within the Greater Accra Metropolitan Area barely beyond one meter with the trees growing to an
from 2015-2019;
2019; (2) Quantify the estimated number of average height of five meters (Farvacque-Vitkovic
(Farvacque et al.,
urban environments affected by flooding in the Greater 2008).
Accra Metropolitan Area; (3) identify the districts within

Figure 1: Study area showing districts.

www.rsisinternational.org Page 215

International Journal of Research and Scientific Innovation (IJRSI) | Volume VIII, Issue I, January 2021 | ISSN 2321–2705

2.2 Software/hardware/materials used in the ArcGIS desktop version. Also, the ArcGIS online
version was used to create a time slider where the flood extent
All image processing from data acquisition, pre-processing,
layers for the various years were animated.
and analysis was performed within the Google Earth Engine
cloud-based platform (GEE). GEE is very efficient for the 2.3 Methods/procedures followed.
analysis of large datasets like satellite images because all the
A detailed overview identifying each major research phase
processing is done within the cloud infrastructure provided by
with their required inputs and expected results is shown in the
Google. Other map creation and visualizations were executed
diagram below.

Figure 2 Workflow of methodology

2.4 Major Flood periods and Dates area were identified based on literature studies. According to
the Poverty & Equity Global Practice paper 156 of the World
Major Flood periods within the Greater Accra Metropolitan
Bank, some major flooding occurred in 2015 where about

www.rsisinternational.org Page 216

International Journal of Research and Scientific Innovation (IJRSI) | Volume VIII, Issue I, January 2021 | ISSN 2321–2705

53,000 people were affected in the Greater Accra where false positives were eliminated in flood layer extent.
Metropolitan Area. Other flood events occurred in 2016, The Global Surface Water dataset was used to mask out all
2017, 2018, and 2019, respectively. For the flood event areas covered by permanent water. This dataset has a
analysis, the 2018 flood event was used for demonstrating resolution of 30m. Also, to remove high slope areas over 5%
flood risk assessment in the Greater Accra Metropolitan Area. in the study area, a digital elevation model, DEM data was
The 2018 floods were also considered a major flooding event used. This operation helps to reduce the noise to the extent of
due to the heavy rains and thunderstorms that occurred on the flooding.
18th and 28th of June 2018 and resulted in 14 deaths,
2.5.3 Area calculation and exposed urban population
displacement of 34,076 others, and damaged properties
according to the National Disaster Management Organization For the computation of the area of the flood extent, the flood
(NADMO) (Ansah et al., 2020). layers as generated in the google earth engine were then
exported as raster layers and opened in the ArcGIS Desktop.
2.5 Datasets used, pre-processing and Analysis.
These layers were then converted into polygons for the
Sentinel-1 dataset based according to the flooding periods various years and a calculation of the entire polygon area for
defined for this project; before and after flooding were the specific years was done using the Geometry tool in
downloaded for the pre-processing and analysis of flood risk ArcGIS. To know the affected urban population, a landcover
mapping. This was done in Google Earth Engine. The data as classification of the study area was made where the urban
provided within GEE (Sentinel-1 Level-1) is a Ground Range areas from the landcover were extracted to assess potentially
Detected (GRD) and therefore has a resolution of 10m. The affected urban areas.
instrument mode is Interferometric Wide Swath (IW). For this
2.5.4 Time Animation of Flood events
project, the image used was acquired in ascending mode pass
direction due to the coverage of the study area which is Processing of the vector layers of flooding extent for the years
located within the Ascending mode. The images were then 2015, 2016, 2017, 2018, and 2019 was done where an
clipped to the study area boundary. addition field was created to handle the date of flooding (year
of flooding). These layers were then published on the ArcGIS
Information from the Sentinel-1 Level-1 Ground Range
online platform where the time animation widget was added to
Detected (GRD) imagery in Google Earth Engine has already
show in animation, the extent of flooding in the study area
undergone some pre-processing steps including the
between 2015 and 2019.
application of orbits, border noise removal which removes
low-intensity noise and invalid data on the scene edges, III. RESULTS
thermal noise removal, radiometric calibration which
computes backscatter intensity using sensor calibration The result as presented in this section gives an overview of the
flood extent layers from 2015 to 2019 within the GAMA area.
parameters, terrain correction and conversion of backscatter
The statistics in terms of which year as well as which district
coefficient into decibels, among others. The only pre-
has the highest flood records are also provided in the results.
processing steps applied was the smoothing filter to reduce the
And in terms of the risk assessment, the extent of the urban
speckle-effects of radar imagery.
environment affected by flooding, using the main flooding
2.5.1 Change Detection event (2018) as the basis for that calculation is also provided.
Results from this research show that the year 2018 experience
The change detection approach was applied in the analysis of
the highest flood extent of 54.57% within the area, followed
flood extent, where the after-flood mosaic is divided by the
by 2016, experiencing about 13.22% of the total flood extent
before-flood mosaic, resulting in a raster layer showing the
by year, followed by the year 2017 (11.72%), followed by the
degree of change per pixel. This approach was adopted from
year 2015 (11.07%) and finally the year with the lowest flood
the United Nations Office for Outer Space Affairs Knowledge
extent is 2019 (9.43%).
Portal on Space-based information for Disaster Management
and Emergency Response (2020). Other results showed that on average, the Ga South district
experienced the highest flooding extent over the period (2015-
In this regard, high values represent bright pixels and indicate
2019), followed by Ga west district, then the Kpone
high change, while low values represent dark pixels and show
Katamanso district, Accra Metropolis, Tema Metropolis, etc.
little change. A predefined threshold was then applied in this
the district that recorded the lowest rainfall extent was the
case, assigning 1 to values greater than the threshold and 0 to
Amasaman district. In terms of the urban areas affected, a
values less than the threshold value. The raster layer created
classification of the study area shows that 45.7% are urban
by this layer shows the potential flood extent. The thresholds
whiles 54.4 are non-urban areas (water, vegetation, wetlands
used were considered and adjusted due to either high rates of
among other landcover types). Out of the 45.7% urban areas,
false positives or false negatives.
about 25% urban areas were affected by flooding, using the
2.5.2 Refining the flood layer year 2018 as the base flood layer, since it recorded the highest
rainfall extent.
Additional datasets were used to refine the flood layer extent

www.rsisinternational.org Page 217

International Journal of Research and Scientific Innovation (IJRSI) | Volume VIII, Issue I, January 2021 | ISSN 2321–2705

Figure 3: Flood extent, 2015

Figure 4: Flood extent, 2016

Figure 5: Flood extent, 2017

www.rsisinternational.org Page 218

International Journal of Research and Scientific Innovation (IJRSI) | Volume VIII, Issue I, January 2021 | ISSN 2321–2705
2321

Figure 6: Flood extent, 2018

Figure 7: Flood extent, 2019

Total Flood Extent by year Flood Extent by Districts in

(%) GAMA

ASHAIMAN
DISTRICTS IN GAMA

9.43 11.07 LEDZOKUKU / KROWOR

13.22
LA DADE KOTOPON
11.72 ACCRA METROPOLIS
54.57
GA CENTRAL MUNICIPAL
GA WEST

0.0020.00
0.00 40.0060.0080.00
100.00

flood_2015 flood_2016 flood_2017

flood_2019 flood_2018 flood_2017
flood_2018 flood_2019
FLOOD EXTENT SQUARE KILOMETRES
Figure 8: Flood extent by year
Figure 9:: Flood extent by districts in GAMA

www.rsisinternational.org Page 219

International Journal of Research and Scientific Innovation (IJRSI) | Volume VIII, Issue I, January 2021 | ISSN 2321–2705

Figure 10: Classification of GAMA into urban and non-urban

Figure 11: Affected urban areas, 2018.

IV. DISCUSSION
Extent of urban areas The flood extent for the period between 2015 and 2019 shows
affected (2018) that flooding in 2018 was very intense as compared to the
other years. This shows that there was a heavy downpour of
54.4 0 25 rains in 2018 in comparison to the other years. The year with
45.7
the lowest flooding extent was 2019, which could be due to a
45.7 20.7 smaller number of rains experience within that period or due
to the changing climate.
Urban Non-urban
unaffected urban
A careful study of flooding by the districts within GAMA was
analyzed and it showed that the GA South district experienced
Affected urban the highest flooding extent over the period (2015-2019),
followed by Ga west district, then the Kpone Katamanso
Figure: 12 Extent of affected urban areas, 2018 district, Accra Metropolis, Tema Metropolis, etc. The district
that recorded the lowest rainfall extent was the Amasaman

www.rsisinternational.org Page 220

International Journal of Research and Scientific Innovation (IJRSI) | Volume VIII, Issue I, January 2021 | ISSN 2321–2705

district. the unavailability of rainfall data for the study period to make
a comparison between the flood extent years and the amount
In terms of the urban areas affected, a classification of the
of rainfall experience for the corresponding years makes it
study area shows that 45.7% are urban whiles 54.4 are non-
quite difficult to attribute the cause of high flooding extent in
urban areas (water, vegetation, wetlands among other
2018 to high amounts of rainfall received.
landcover types). Out of the 45.7% urban areas, about 25%
urban areas were affected by flooding, using the year 2018 as According to results from this study, about half of the urban
the base flood layer, since it recorded the highest rainfall areas within the study area gets flooded during heavy rains
extent. within GAMA. Per the land cover classification performed
with the 2018 satellite data, about 45% of GAMA is urban
The Ga south district is more flood-prone as compared to the
and the city is currently experiencing rapid growth and spatial
other districts. This can be attributed to the closeness to the
sprawl. Most of this growth is occurring in the peri-urban
Weija Dam [2]. The Weija Dam which was constructed in
areas particularly in the GA south, Ga west, and GA east
1978 has always served as the primary source of water in the
districts which initially were rural areas. The increase in urban
Greater Accra Region, supplying up to 80% of the water in
areas has a significant hydrological implication. It reduces the
Accra, unfortunately, due to human activities such as the
flood regularity of the city according to [18]. This is because
dumping of waste, its banks keep closing [11]. Whenever the
the infiltration rate is reduced by increase urban environments
Ga south district experiences severe rainfall, the Weija Dam is
and thus increase susceptibility to flooding. This explains why
filled with excess water beyond its capacity, and therefore the
the majority of urban areas experience flooding.
Ghana Water Company Limited (GWCL) begins spillage of
the Dam to prevent the collapse of the dam [6]. This action The results provide a general overview of flooding within the
however leads to severe flooding in the district especially GAMA zone from 2015 to 2019. This flood extent shows the
surrounding communities, particularly Tetegu, Oblogo, and total amount of flood experience as well as showing which
Weija Dam catchment area. district has the highest flooding in comparison to other cities.
This is very critical for the government and the national
On 10th July 2017, about four regions including Greater
response agencies to help in their response and rescue
Accra, East, West, and Central Regions were declared to have
missions during flood outbreaks. Flooding in Ghana,
flood emergencies with the potential to cause devastation.
especially in Accra always leads to the loss of lives and
According to the assessment conducted in Greater Accra,
properties which in most cases needs responses from
households located in the low-lying areas in Weija and Tetegu
government agencies. Being aware of districts that experience
in the Ga South district had been inundated with water
the highest flooding is going to better inform the response
because of spillage from the Weija Dam (IFRC, 2019).
team as to the allocation of resources to the affected districts
Another factor for high flooding occurrence is the low-lying of flooding.
nature of the Greater Accra Region in general. Accra is
Having an overview of the current flooding extent helps in
generally a low-lying area, though it rises to 100m above sea
forecasting the extent of flooding in the coming years which
level in some locations. Several catchment areas lie between
will eventually help the government in planning the cities in
the two ridges in the east and west party of the city [5].
response to flooding.
According to (Okyere, 2015), topography provides insight
into the direction of moving waters and areas which are The results give a general picture of districts and communities
potential receptacles of run-offs. The Greater Accra District is within the GAMA area that is mostly affected by flooding.
of importance when it comes to flooding because it is the What is lacking here in terms of the damage and response
capital city of Ghana and the hub of urban infrastructure as assessment of the flood events was to quantify in terms of the
well. According to [18], flooding problems can be attributed exposed population to flooding. This was missing because of
to the weak planning systems that the country faces. Accra has the unavailability of population data which to perform such
a massive development, most of which are residential analysis. The district-by-district analysis of flood extent
development. The problem now is that most of these generalizes the situation in the district. There however could
developments take place ahead of planning and thus proper be local situations that will probably be different from what
layout and infrastructure which should accompany residential the entire district portrays. Hence the analysis could have
development in developing areas have been decidedly absent taken into consideration smaller communities for the analysis
and this goes a long way to cause flooding during high instead of looking at entire districts within GAMA.
amounts of downpours.
The derived flood maps provide spatial and temporal
The year 2018 from the analysis happens to have experienced dynamics of the flooding area across the study area. The
the highest amount of flooding within the GAMA area. This utilization of the GEE for the image processing facility in this
could be attributed to climatic factors such as high amounts of study helped to develop an operational land cover map for
rainfall experienced within the year in question. Research damage assessment. The method can be used to monitor land
according to [2], suggests that climatic variability especially cover regularly as the analysis can easily be re-run while new
rainfall contributes immensely to the issue of flooding. Hence satellite data is ingested in GEE. The derived maps and charts

www.rsisinternational.org Page 221

View publication stats

International Journal of Research and Scientific Innovation (IJRSI) | Volume VIII, Issue I, January 2021 | ISSN 2321–2705

provide crucial information for local disaster management There should be sustained funding for NADMO activities so
agencies, which help to prioritize relief and rescue operations. that they can provide more relief items to affected victims, as
The use of SAR images provides significant advantages, as well as embarking on sensitization programs to build
their cloud-free and all-weather capabilities enable the communities real residence to floods.
production of regularly sampled flood extent information in a
REFERENCES
real-time delivery manner.
[1] Kwasi Appeaning Addo et al. “Impacts of Coastal Inundation Due
4.1 Remaining gaps/needs for further research. to Climate Change in a CLUSTER of Urban Coastal Communities
in Ghana, West Africa”. Remote Sensing 3(9) (2011), pp. 2029–
The study produced a flood map with some level of accuracy, 2050. DOI: 10.3390/rs3092029. URL:
however, there were also some uncertainties due to the change https://dx.doi.org/10.3390/rs3092029.
detection method used, where some changes could not be [2] E Dogbevi. 2017. URL:
https://www.ghanabusinessnews.com/2017/05/30/reconstruction-
because of flooding and hence classifying all changes because
costs-after-accra- floods-estimated-at-105m-world-bank/.
of flooding might not be entirely accurate. To overcome those [3] C Farvacque-Vitkovic et al. “Development of the Cities of Ghana:
uncertainties, local knowledge is crucial. Alternatively, Challenges, Priorities, and Tools”. Africa Region Working Paper
unmanned aerial vehicles (UAVs) are emerging tools for 110 (2008), pp. 1–157.
[4] Elissavet Feloni, Ioannis Mousadis, and Evangelos Baltas. “Flood
monitoring real-time floods in disaster management. This
vulnerability assessment using a GIS-based multi-criteria
could be applied to accurately map real-time flooding events approach—The case of Attica region”. Journal of Flood Risk
(Uddin et al., 2019). Further research should consider a Management 13(S1) (2020), pp. 1–15. DOI: 10.1111/jfr3.12563.
comparison of rainfall amounts experienced for the various URL: https://dx.doi.org/10.1111/jfr3.12563.
[5] N K Karley. 2009.
years with the flood extent for the years of study to accurately
[6] Li Lin et al. “Improvement and Validation of NASA/MODIS NRT
justify that rainfall or climatic conditions are the result of high Global Flood Mapping”. Remote Sensing 11(2) (2019), pp. 205–
flooding in a particular year compared to the other years. 205. DOI: 10.3390/rs11020205. URL:
https://dx.doi.org/10.3390/rs11020205.
V. CONCLUSIONS AND RECOMMENDATIONS [7] B Nyarko. “Application of a rational model in GIS for flood risk
assessment in Accra, Ghana”. J Spatial Hydrol 2 (2002), pp. 1–14.
The whole study area is vulnerable and has experienced [8] Xinyi Shen et al. “Inundation Extent Mapping by Synthetic
flooding throughout the study (2015 – 2019). The Ga South Aperture Radar: A Review”. Remote Sensing 11(7) (2019), pp.
district is the district that has the highest extent of flooding 879–879. DOI: 10.3390/rs11070879. URL:
https://dx.doi.org/10.3390/rs11070879.
averagely over the study period due to the spillage of the
Weija Dam within the district. Most communities such as UNCITED REFERENCES
Tetegu, Kaneshie, Odorkor, Accra, Weija, and Mallam among
others are considered flood-prone areas or areas liable to [9] 2016. URL: https://reliefweb.int/report/ghana/ghana-floods-
emergency-plan-action- epoa-mdrgh011.
flooding. Other conclusions were made that due to the [10] 2020. URL: http://www.un-spider.org/advisory-
topography of Accra which happens to be a low-lying area, support/recommended- practices/recommended-practice-google-
the majority of such communities are therefore liable to earth-engine-flood-mapping: Assessed on 22nd.
flooding. Other causes of flooding are the rapid urbanization [11] A Annor. 2020. URL:
https://www.ghanaiantimes.com.gh/displaced-weija-residents-
in the study area which leads to the construction of buildings count-their-losses-one-week-after-dam-spillage/.
in waterways. [12] S. O. Ansah et al. “Meteorological Analysis of Floods in Ghana”.
Advances in Meteorology 2020 (2020), pp. 1–14. DOI:
Recommendations 10.1155/2020/4230627. URL:
https://dx.doi.org/10.1155/2020/4230627.
Based on the findings arising from this study, several [13] A Doreen. 2020. URL: https://www.graphic.com.gh/news/general-
recommendations are hereby put forward for policy news/building-on- watercourses-human-activities-cause-Accra-
considerations. The government should intervene in floods.html.
displaying the political will to prevent further development of [14] Ifrc. 2017.
[15] B Okyere. 2015. URL:
buildings in the area very close to the Weija Dam. These are https://www.ghanaweb.com/GhanaHomePage/NewsArchive/Floo
legally acquired buffer zones that some developers have ding-in-Accra-Ghana- How-can-we-help-the-victims-362650.
encroached upon, therefore authorities should invoke the [16] S Peter. 2009. URL:
relevant legislation to prevent further loss of lives and https://www.modernghana.com/news/223780/1/flooding-in-accra-
research-report.html.
property in the areas. Again, structures identified to cause a [17] G Tetteh. 2017.
flood or in waterways need to be demo0lished. And there [18] The Human Cost of Weather-Related Disasters. Geneva,
should be no political interference in exercising such Switzerland, 1995.
authority. [19] K Uddin, M A Matin, and F J Meyer. 2019. URL:
https://doi.org/10.3390/rs11131581.
Efforts should be made to construct drainage facilities and
complete all projects aimed at reducing flood risk in the study
area. This requires increased funding for such projects to be
performed.

www.rsisinternational.org Page 222