Mapping landscapes of Africa using remote sensing data: detecting spatio-temporal environmental dynamics from the satellite images

Mapping landscapes of Africa using remote sensing data:
detecting spatio-temporal environmental dynamics
from the satellite images
Polina Lemenkova
June 13, 2024
Polina Lemenkova Mapping landscapes of Africa using remote sensing data: detecting spatio-temporal environmental dynamics f
June 13, 2024 1 / 40

Introduction
Introduction
Concept: Landscapes of Africa: land surface
where diverse environmental processes interplay
Understanding landscape dynamics: modelling
and mapping complexity of factors that affect
the shape of the Earth
Dynamics: spatio-temporal changes caused by
human and natural forces
Applications of landscape ecology and
environmental monitoring of Africa:
Landscape monitoring
Land management (urban planning)
Sustainable development (food resources,
agriculture)
Nature: Factors affecting formation of
landscapes: geologic-tectonic setting, climate
processes, anthropogenic activities => different
relief, soil and vegetation
June 13, 2024 2 / 40

Introduction
Concept
EO Data as an Opportunity
Possibilities: Technical revolution in the amount of Earth data (satellite imagery, digital
raster and vector datasets, tabular data) enabled smooth and quick access to monitoring of
the large regions of the Earth, such as African continent
Access: Open datasets and repositories available online enable to access, download, examine,
and map RS and cartographic spatial data for mapping Africa
Coverage: EO data provide information on spatial events across local, regional, and global
scales of African landscapes: e.g. risk assessment, tsunami caused by monsoons, earthquakes
(East African rift), landslides
Machine Learning for Processing big Earth data
Problem: Extracting knowledge and information from EO data requires advanced technical
tools => need for ML/DL algorithms and scripting approaches to data handling
Paradigm: ML and scripts present advanced methods of EO data analysis that automate
data processing, modelling, mapping and visualization.
Hierarchy: ML is a branch of Artificial Intelligence (AI) => GIS learns from data and
extracts information. Identification of landscape patterns and detecting land cover changes
June 13, 2024 3 / 40

Introduction
Objectives and Goals
Machine Learning in Earth data processing for mapping Africa
Mapping: Environmental analysis of African landscapes through processing of RS data
Cartography: Generic Mapping Tools (GMT): special syntax enabling to quickly operate with
cartographic data using shell scripts
Examples: Automation of data processing: programming languages (Python and R),
scripting GIS and scripting toolsets (GRASS GIS, GMT)
Remote sensing
Using scripting algorithms for RS data (satellite images, e.g., Landsat) to extract
environmental information for mapping variability of landscapes across Africa
Research Techniques
This research includes algorithms of Generic Mapping Tools (GMT) that has a scripting
syntax enabling to quickly operate with spatial data using shell scripts
June 13, 2024 4 / 40

AF
Research Highlights: Variability of African landscapes
Complexity: variability of regional
geographic setting on the African
continent
Scale: Handling multi-source
spatial data for local, regional and
continental scales
Data-driven research: Variety of
RS data (Landsat TM/ETM+
OLI/TIRS, Sentinel-2A, SPOT),
weather data as time series,
climate forecasts, ecological
monitoring, seismic observations
(hazard assessment).
Methods: Processing spatial data
by advanced tools and methods
(scripts, programming algorithms) Landscape ecology of diverse regions of Africa reflects a
variety of impact factors: tectonic-geologic setting,
climate change and anthropogenic activities
June 13, 2024 5 / 40

AF
Research Progress: Current state of the project
Current state: mapping of 27 African countries is published in 32 journal articles
No Name Publication Link
1 Algeria [9] doi: 10.3390/asi6040061
2 Botswana [12] doi: 10.3390/ijgi11090473
3 Burkina Faso [25] doi: 10.24425/agg.2023.146157
4 Cameroon [23] doi: 10.47246/CEJGSD.2020.2.2.4
5 Central Africa Republic [28] doi: 10.3390/min13050604
6 Chad [5] doi: 10.2478/boku-2023-0005
7 D.R.C. Congo [31] doi: 10.3390/app122412554
8 Egypt [27] doi: 10.3390/info14040249
9
Ethiopia
[11, 15] doi: 10.24425/jwld.2022.141573; doi: 10.5755/j01.erem.78.1.29963
10 [21, 19] doi: 10.4314/sinet.v44i1.9; doi: 10.2478/abmj-2021-0010
11 Ghana [14, 17] doi: 10.24425/gac.2022.141169; doi: 10.32909/kg.20.36.2
12 Guinea [4] doi: 10.5281/zenodo.10673287
13 Ivory Coast (Côte d’Ivoire) [32] doi: 10.3390/jimaging8120317
14 Kenya [8] doi: 10.2478/trser-2023-0008
15 Malawi [13, 18] doi: 10.5937/tehnika2202183L; doi: 10.5281/zenodo.5772315
16 Mali [24] doi: 10.2478/arsa-2023-0011
17 Mozambique [3] doi: 10.3390/coasts4010008
18 Nigeria [29] doi: 10.3390/jmse11040871
19 Rwanda [10] doi: 10.5281/zenodo.7113414
20 Sierra Leone [1] doi: 10.5281/zenodo.11473307
21 Somalia and Seychelles [30] doi: 10.2478/jaes-2022-0026
22 South Africa [2] doi: 10.4000/11pyj
23 South Sudan [7] doi: 10.3390/analytics2030040
24 Sudan [26] doi: 10.3390/jimaging9050098
25 Tanzania [16] doi: 10.3176/earth.2022.05
26 Tunisia [6] doi: 10.3390/land12111995
27 Uganda [20] doi: 10.5281/zenodo.5082861
28 Zambia [22] doi: 10.5937/ZemBilj2101117L
June 13, 2024 6 / 40

Data
Data
Understanding complexity and variability of landscapes requires processing of large amounts of diverse datasets
Multi-source data => effective, rapid yet accurate processing
High-resolution data => at print-quality mapping and modelling. Accuracy of digital Earth data is of crucial importance,
because it directly controls research output.
Table 1: Data used in this project: their sources, types and precision
No Data Origine Type, resolution
1 Landsat 8-9 OLI/TIRS USGS satellite images
2 ETOPO1 NOAA 1 arc-min GRM grid
3 ETOPO5 NOAA 5 arc min GRM grid
4 GEBCO BODC 15 arc sec DEM raster grid
5 SRTM NASA 15-sec DEM raster grid
6 Geology USGS Vector layers for diverse African countries
7 Social data FAO Vector layers for administrative borders and population
8 Gravity Scripps IO CryoSat-2, Jason-1 grid
9 EGM96 Scripps IO Geopotential data: geoid
10 EGM-2008 Scripps IO Geoid model
11 TerraClimate NOAA Climate characteristics of African landscapes
12 GLOBE dataset NOAA Topography model
13 GEBCO IHO-IOC Geographic gazetteer
*30 arc second = ca. 1km cell size, 15 arc sec = ca. 500 m cell size, etc.
June 13, 2024 7 / 40

Methods
Methods
Cartographic tasks by scripts: Generic Mapping
Tools (GMT)
Figure 1: Using GMT as integrated tools for cartographic workflow
Plotting a map by GMT becomes a rapid procedure, fully
controlled by the cartographer
Such automatisation by GMT facilitates mapping and
decreases time of cartographic workflow
Various GMT techniques and modules have been
used in this research for rapid and accurate mapping
à gmt psxy for plotting graphs and adding
annotations on raster map
à gmt makecpt for making color palette
à gmt grdimage for generating raster image
à gmt psscale for adding legend for adding scale
à gmt grdcontour for adding isolines
à gmt psbasemap for adding grid and essential
cartographic marks
à gmt psbasemap for adding scale and
directional rose
à echo UNIX utility for adding text labels
à gmt pstext for adding subtitle
à gmt logo for adding GMT logo
à gmt psconvert for convert PS to image (by
GhostScript interpreter)
à gmt legend adding legend on maps with
cartographic annotations
à gmt pshistogram for statistical analysis of
topographic data distribution (relief)
à gmt psrose for plotting rose diagrams
à gmt img2grd converting geospatial formats
à gmt grdcut selecting study area and clipping
the region from the world map
These and many other modules of GMT
were used in this research.
June 13, 2024 8 / 40

DZ
Algeria
Algeria – the largest country in
Africa and the Mediterranean
Basin.
Contrasting environment:
south - a significant part of
Sahara Desert;
center – Tell Atlas Mts. (Saharan
Atlas) + vast plains, highlands
(e.g., Hoggar Mts) & mountain
ranges
coastal areas: hilly &
mountainous landscapes, a few
harbours.
Diverse topographic setting +
climate setting = contrasting
precipitation patterns
Salt lakes (saline sabkhas) in the
semi-desert environment on
border of Sahara
Figure 2: Topographic map of Algeria
June 13, 2024 9 / 40

DZ
Algeria, Sahara: Sabkhas Salt Lakes
Figure 3: Sabkha - coastal sandy flats of northern Algeria where evaporite and
saline minerals accumulate. Sabkhas fluctuate seasonally and over years
(subject to climate change)
Figure 4: Workflow used to map sabkhas of Algeria using RS
data
Sabkha of Algeria - in semiarid to
arid climate (border of Sahara +
impact of Mediterranean).
June 13, 2024 10 / 40

BW
Botswana
Figure 5: Topographic map of Botswana with specific geographic features.
Relief: mostly flat, gently rolling tableland
Climate: dominated by the Kalahari Desert (ca. 70% of its
land surface)
Hydrology: Okavango Delta – one of the world’s largest
inland river deltas (UNESCO World Heritage Site) is
located in Botswana.
Figure 6: Right: climate variables over Botswana (here: droughts)
Environment Okavango Delta – oasis in an arid climate of
Botswana => large wetland system
Ecology rich wildlife, swamp-dominant rare species.
Habitats: salt inlands, lagoons.
June 13, 2024 11 / 40

BF
Burkina Faso
Food insecurity & Poverty ($1,000) per capita - IMF ’23 Reasons for food instability:
Environmental hazards Burkina Faso experiences one of the most contrasting climatic variations in the world with
environmental hazards varying from severe floods to extreme drought => agriculture instability
Climate issues: tropical with two very distinct seasons – rainy and dry seasons. During dry period - effects from harmattan
(hot dry Saharan winds )
Agricultural vulnerability: insect attacks (locusts and crickets), which destroy crops
Migration: people migrate to other countries to find better jobs and support for their families
Topography: flat relief with 2 major types of landscapes influenced by geologic setting.
Peneplain: larger part of the country (gently undulating landscape and a few isolated hills) from Precambrian massif.
Sandstone massif: on the southwest of the country with the highest peak, Ténakourou
Figure 7: Data capture from the EarthExplorer repository of the USGS: Landsat-8 OLI/TIRS satellite image, central Burkina Faso, Ouagadougou.
June 13, 2024 12 / 40

CM
Cameroon
Figure 8: Topographic map of Camerron
Cameroon – Africa in miniature
All major climates and vegetation of the
continent:
Landscapes: a mixture of coastal
ecosystems, desert plains and mountains
and savannah in the central regions,
tropical rain forests in the south
Social-environmental problems of Cameron
Population growth, urbanisation and
industrialisation
Land degradation, water / air pollution
Waste management (plastic pollution)
Anglophone crisis: colonial history of
Cameroon (France and UK shared
Cameroon unequally – French area 80% of
the country => two nations in one country
June 13, 2024 13 / 40

Data
Data Integration
Mapping central African regions (complex geology)
Multi-source data mixing and integration using GMT adjustments:
Data mining from open sources (e.g. Landsat, GEBCO, ETOPO1, geological data, etc) for geographic analysis
Predictive environmental analytics: forecasting and prognosis in hazard risk assessment based on climate data and FAO
information on food security
Mapping geophysical setting in central African countries, volcanic activity, seismicity, earthquakes
Data analysis by GMT has potential to advance development of cartography
It enables scientific discoveries, based on spatial analysis in geology
Scripting techniques in cartography
GMT-based mapping provides accurate visualization of the complex terrain as in African countries
Solutions of GMT:
Advanced level of cartographic functionality: variety of projections, data processing, import/export, etc
High-quality cartographic output: colour palettes, layouts, grids, vector lines, advanced design approaches
Flexibility of coding / shell scripting from the console
Embedded data analysis and visualisation: mapping, modelling, statistical analysis, logical queries, import/export, vector
and raster processing
June 13, 2024 14 / 40

CAR
Central African Republic (CAR)
Figure 9: Snippet of the GMT script for cartographic mapping
Figure 10: Topographic map and location of the CAR to analyse correlation
between topography and geophysical setting
June 13, 2024 15 / 40

TD
Chad
Figure 11: Topographic map of Chad
Chad: social-environmental problems
Poverty: Chad – the 7th poorest country in the world, with 80% of the
population living below the poverty line (Source: UN’ Human
Development Index)
Arid climate, droughts: Landlocked country with arid climate.
Terrestrial ecoregions: savannah, xeric woodlands, south Saharan
steppes
Lake Chad: important source of food and natural resources; yet subject
to seasonal and climate fluctuations
RS and cartography for landscape analysis
June 13, 2024 16 / 40

CG
Congo (D.R.C)
Figure 13: Topographic map of Congo D.R.C.
Climate hazards: Congo: high precipitation
and the highest frequency of thunderstorms
in the world.
Environment: Congolese rainforests – the 2nd
largest rain forest in the world after the
Amazon
Biodiversity of Congo: lush jungle rainforest
+ savannah + grasslands + mountainous
terraces
Geomorphology: Rwenzori Mountains – the
source of the Nile River with unique (for
Africa) alpine meadows
Figure 14: Advantages of R for satellite image processing (libraries)
Figure 15: Workflow for research methodology
June 13, 2024 17 / 40

CG
Congo (D.R.C.)
Figure 16: Snippet of the R code for unsupervised
classification by k-means clustering (Basoko, Congo DRC)
.
Automatisation in RS data
analysis for Congo: diverse
geospatial libraries of R
Figure 17: k-means clustering classification algorithm in RStudio using the
RStoolbox package for satellite image processing (Landsat images covering
Congo DRC)
Automatic image processing by R to study
changes in vegetation of Congo, to analyse
their correlation with landforms
(topography), detect correlation with
climate settings
June 13, 2024 18 / 40

EG
Egypt
Figure 18: Qena Bend as study area on the map of Egypt
Cartographic scripting by GMT is essential for
3D modelling and geomorphological mapping
to capture the correlation between geological,
topographic and environmental variables
Console-based mapping techniques are
script-based and data-driven
Figure 19: 3D perspective view of Qena Bend, Nile River, Upper Egypt
Features of Qena Bend:
Qena Bend – a remarkable geomorphological feature in southern Egypt.
Qena Bend presents a great loop of the Nile River towards the east
Qena Bend is bounded by limestone cliffs on both sides
Qena Bend is one of the most structurally-complicated regions of the
Nile Valley, where it intersects from the east with some tectonic trends:
Wadi Qena and Qena–Safaga Shear Zone
The geodynamic formation of Qena Bend is still not fully understood.
June 13, 2024 19 / 40

ET
Ethiopia
Figure 20: Geomorphic features of Ethiopia: slope, aspect, DEM and hillshade
relief map mapped using R
Geologic mapping
Advanced cartographic tools => mapping relief =>
understanding the links between geology, geomorphology,
seismicity and topography of Ethiopia
Lnks among geomorphology, tectonics and geology =>
insights into geologic evolution of the African continent
and current topography of the East African Rift region
Important geologic-geographic
features of Ethiopia
The major part of Ethiopia is located in
the Horn of Africa, – the easternmost part
of the African landmass
Great Rift Valley divides Ethiopia into two
parts of landscapes with a vast highland
complex of mountains and dissected
plateaus
Great Rift Valley – branch of the East
African Rift which stretches in SW-NE
direction from the Afar Triple Junction.
Afar Triple Junction is a unique example
on the Earth of continental rifting (as
opposed to seafloor spreading)
In brief, the geologic reason for Afar Triple
junction – extension of lithosphere crust,
caused by mantle upwelling => seismically
active region
Great Rift Valley is surrounded by
lowlands, steppes, or semi-desert.
June 13, 2024 20 / 40

GH
Ghana
Figure 21: Thematic maps of Ghana plotted using GMT
Hydrology of Ghana
White Volta River + its tributary Black Volta, flowing to Lake Volta – the world’s third-largest reservoir by
volume and largest by surface area,
=> hydroelectric Akosombo Dam – major electricity producer.
Ecoregions of Ghana
5 terrestrial ecoregions of Ghana: Eastern Guinean forests, Guinean forest–savanna mosaic, West Sudanian
savanna, Central African mangroves, and Guinean mangroves.
Mapping using diverse data (geologic, topographic, land cover types) helps get better insights into the
underlying processes of landscape formation which affect the distribution of vegetation types.
June 13, 2024 21 / 40

GN
Guinea
Figure 22: Topographic map of Guinea
Environmental alerts of Guinea:
Reduced wildlife due to human encroachment and
hunting (e.g., large mammals)
overfishing =>t hreat to the nation’s marine life.
Threatened species: the African elephant, Diana
monkey, and Nimba otter-shrew.
A nature reserve Mount Nimba Strict Nature
Reserve is established on Mt. Nimba: protected
area and UNESCO World Heritage Site
Figure 23: Landscape types of Guinea
Environmental features of Guinea:
Guinean Forests of West Africa Biodiversity
hotspot - southern part of Guinea
Dry savanna woodlands (north-eastern region)
Guinea is home to 5 ecoregions: Guinean montane
forests, Western Guinean lowland forests, Guinean
forest-savanna mosaic, West Sudanian savanna,
and Guinean mangroves.
June 13, 2024 22 / 40

CI
Côte d’Ivoire (Ivory Coast)
Figure 24: Satellite image processing and DEM of Ivory Coast visualised using
Python (fragment)
Environmental highlights of Ivory Coast:
The most biodiverse country in West Africa (over
1,200 animal species, 223 mammals, 702 birds,
161 reptiles, 85 amphibians, 111 fish species,
4,700 plant species, of which 3,927 species of
terrestrial plants)
The world’s largest exporter of cocoa beans;
The largest economy in the West African
Economic and Monetary Union (40% total GDP)
9 national parks (the largest – Assgny National
Park); 6 terrestrial ecoregions
Economy grows fast; high-income country; highly
diversified agriculture (cocoa, cashews, coffee,
rubber, cotton, palm oil, bananas, etc)
Figure 25: Topographic map of Ivory Coast, West Africa
Environmental problems and alerts in
Ivory Coast:
High deforestation rates due to high
agriculture activities and economic
activities: logging of timber, mineral
exploration
Loss of biodiversity – declining variety
of flora and fauna
June 13, 2024 23 / 40

KE
Kenya
Figure 26: Map of Kenya performed using GMT and GEBCO data
East African Rift
Tectonics: active continental rift zone in East
Africa which affected countries => active and
dormant volcanoes (incl. Mt. Kilimanjaro, Mt.
Kenya etc).
Figure 27: Geological map of Kenya. Source:
Multi-source EO data
Analysis of geological data to understand
topographic structure which affects distribution of
vegetation and landscape formation through relief
forms.
EO data present an integration of geospatial data,
collection of RS resources (Landsat images,
topographic and geologic data, climate,
vegetation) => information extraction and
knowledge mining
Data heterogeneity: multi-source data of different
formats and scales => data integration and
analysis.
June 13, 2024 24 / 40

TZ
Tanzania
Figure 28: Script used for 3D mapping
Figure 29: 3D mapping of Tanzania
3D modelling as important cartographic
approach in EO data analysis
x 3D topographic scene effectively illustrates the relief
representation
x Advanced GMT functionalities specify different
cartographic elements on the map for each layer by a
combination of the GMT modules (grdview, grdcontour,
pscoast, pstext).
x Cartographic processing includes
å visualisation of layers (raster grids vector lines)
å text annotations
å subplots, hierarchical sub-levels of the elements
å coordinate axis labels and grids, translucency of layers
å general layout setting
x The outcome of the GMT cartographic processing:
print-quality maps
x 3D modelling of Earth data creates potential for
cartographic opportunities in visual data analytics and
interpretation of the seafloor structure
x GMT enables flexible, accurate and rapid visualisation of
the 2D and 3D EO data
x Data selection: 3D modelling of terrain landscapes is
limited by massive data processing (3D arrays of cells)
=> ETOPO5 presents better
June 13, 2024 25 / 40

MW
Malawi
Figure 30: Topographic and seismic maps of Malawi (based on IRIS
data on deep earthquakes).
Geographic features of Malawi:
The Great Rift Valley runs through the
country from N to S
Lake Malawi – (a.k.a. Lake Nyasa) - 3/4 of
Malawi’s eastern boundary; one of the major
Rift Valley ancient lakes
Lake Malawi - unique meromictic lake (layers
of water that don’t intermix and don’t
circulate) with specific hydrology –
permanently stratified water layers
Lake Malawi - 4th
largest freshwater lake in
the world by volume, the 9th
largest lake in
the world by area and the 3rd
largest and 2nd
deepest lake in Africa.
Lake Malawi National Park created to protect
unique ecosystems of the lake
Problems in Malawi:
Tanzania–Malawi dispute about the borders
of the Malawi Lake
Deforestation - serious environmental issue
Scarcity of natural resources (landlocked)
Extreme poverty and rapidly growing
population dependent on resources
June 13, 2024 26 / 40

ML
Mali
Figure 31: Topographic map of Mali plotted using GMT scripts and GEBCO
data. The green rotated square shows the location of the Landsat satellite
images used for image analysis. These images cover the area of Inner Niger
Delta.
Figure 32: Scheme of information extraction from RS data to analyse
vegetation indices for ecological monitoring of Niger Inner Delta, Mali. For
time series analysis, images are collected in various years.
June 13, 2024 27 / 40

ML
Mali: Inner Niger Delta
Figure 33: Landsat 8-9 OLI/TIRS images: showing Niger Inner Delta, Mali in
natural colours.
Image processing by R
Analysing land cover types by R-based satellite image
processing => visualising heterogeneity of landscape
features formed under various environmental setting:
Example of data analysis by R: computing vegetation
indices in Mali; clustering for image classification.
Data analysis in R is based on a full range of algorithms
used in RS data processing.
Figure 34: Unsupervised classification by k-means clustering: Inner Niger Delta
June 13, 2024 28 / 40

MZ
Mozambique
Figure 35: Workflow used for RS data classification using Random Forest (RF)
in GRASS GIS: A case of Mozambique
Environmental problems in Mozambique
Deforestation & loss of natural habitats; disrupted coastal
ecosystems, land cover change, land degradation
Mangroves and wetlands are being lost and converted into
rice farms, aquaculture and housing
Climate change and high population growth
Destructive fishing practices (e.g. use of dynamite & fine
mesh nets) => declining fish stocks;
Figure 36: RS data classification using Random Forest in GRASS GIS
Climate setting in Mozambique
Tropical climate with two seasons: wet (Oct-Mar)
and dry (Apr-Sep); => responses of vegetation.
Climatic conditions vary with altitude.
Rainfall is heavy along the coast and decreases in
the north and south of Mozambique.
June 13, 2024 29 / 40

ML
Deep Learning / Machine Learning
Figure 37: Conceptual scheme of ML/DL algorithms
ML in cartography and geoinformatics
Ù Rapid development of ML/DL => Cartographic
knowledge encoding with algorithms of data handling
Ù Using cartographic patterns as classification/generalisation
rules for learning models (as seed) for processing RS data
Ù DL is a new, rapidly evolving field of ML and consists in
advanced algorithms of computer vision and analysis.
Ù ML algorithms include, among others, methods of
Artificial Neural Networks (ANN), Random Forest (RF),
Convolution Neural Networks (CNN) and Recurrent Neural
Networks (RNN), Support Vector Machines (SVM)
Figure 38: Methodological scheme of ANN
ML opportunities for satellite image processing
Ù ML techniques (Python, R, GRASS GIS) can help visualise
land cover patterns in landscapes using classification of RS
data
Ù Land cover types of the diverse landscapes are inherently
linked with regional topographic, geologic and climate
setting
Ù ML enables to get insights into land cover patterns
(variability, dynamics using time series analysis,
fragmentation) formed in various environmental conditions
in diverse ecoregions of Africa
June 13, 2024 30 / 40

NG
Nigeria
Figure 39: Topographic map of Nigeria with shown study area (region of southern coastal
Nigeria with mangroves as aquatic plants)
Environmental problems in Nigeria
Deforestation due to extensive and rapid
clearing of forests (reasons: expanding
agriculture, logging, urbanisation, and
infrastructure development)
Niger Delta pollution due to oil exploration /
petroleum industry => Nigeria’s Delta
region is one of the most oil-polluted
aquatories in the world
Decline of mangroves in coastal areas due to
oil pollution (mangroves are precious
ecosystems and unique saltwater-tolerant
plants)
Land cover changes (important types:
tropical forest zones in the south, wetlands,
mangroves and freshwater swamps in the
Atlantic coasts)
Waste management including sewage
treatment => water pollution, soil
contamination (untreated wastes go to
groundwater)
Population growth => non-systematic
industrial planning, increased urbanisation,
increased of urban spaces
June 13, 2024 31 / 40

NG
Nigeria
Figure 40: Snippet of GRASS GIS script for image processing
RS data processing by GRASS GIS for
analysis of land cover changes in the
coastal regions of Nigeria (monitoring
the decline of mangroves).
Figure 41: Landsat OLI/TIRS images classified using GRASS GIS
June 13, 2024 32 / 40

RW
Rwanda
Figure 42: Geophysical maps of Rwanda: seismicity (deep-focus earthquakes according to IRIS), inundations of the geoid model and free-air gravity (Faye’s)
Geographic features and particularities of Rwanda
High-altitude relief: the entire country is located in uplands and highlands. The lowest point is the Rusizi River (950 m a.s.l)
Mountains of Rwanda - part of the Albertine Rift Mountains (branch of the East African Rift) => complex geologic and
tectonic setting, high seismicity (”Land of a thousand hills”).
Temperate tropical highland climate => montane forests, terraced agriculture
Volcanoes National Park => rare species (1/3 of the worldwide mountain gorilla population) living in bamboo forests; also:
rhinos, lions, chimpanzees, and other protected animal species.
Nyungwe Forest: rare species endemic to Albertine Rift
3 terrestrial ecoregions => Albertine Rift montane forests, Victoria Basin forest-savanna mosaic, and Ruwenzori-Virunga
montane moorlands.
June 13, 2024 33 / 40

SS
South Sudan
Figure 43: Image segmentation performed using GRASS GIS for Landsat images covering South Sudan
Highlights of South Sudan
` Civil war => destruction and displacement > 2M people died, and > 4 M became refugees or displaced persons
` Origine of Nile – White Nile passes through the country, passing by Juba. Name ”White” due to clay sediment contained in
the water => water color pale
` Rich wildlife => tropical forests – habitat for rare species: bongo, giant forest hogs, red river hogs, forest elephants,
chimpanzees, forest monkeys, etc.
` Bandingilo National Park protected area – the 2nd
-largest wildlife migration in the world
` Ecoregions of South Sudan => tropical forest, swamps, and grassland
June 13, 2024 34 / 40

SD
Sudan
Figure 44: Aerial images showing the confluence of the Blue
Nile and White Nile which meet in Khartoum to form the Nile
Highlights of Sudan
T 5 Cataracts of Nile: 5 out of 6 Cataracts of
Nile are located in Sudan (from 2nd
to 6th
,
with only the 1st
being in Egypt)
T Land cover changes in Sudan:
Desertification, deforestation, soil erosion
=> displacement of vegetation
T Environmental problems: water & air
pollution, waste management, trash and
plastic disposal
T Human activities: unsustainable agricultural
expansion => soil desiccation, lowering of
soil fertility and water table
T Threats to wildlife through poaching and
illegal hunting
T Natural hazards: sandstorms in northern dry
regions (”haboob”).
Figure 45: Cataracts of Nile on the topographic map of Sudan
Cataract-related landscape formation
Region of northern Sudan is tectonically active => Nubian Swell
has diverted the Nile’s course => formation of the cataracts,
shallow depth of Nile => black soil/deposits in arid areas of
cataracts => shallow alkaline pans => swampy environment in
cataracts’s areas, northern Sudan
June 13, 2024 35 / 40

TN
Tunisia
Figure 46: RS data processing to compared two gulfs of Tunisia (Gulf of Gabes and Gulf of Hammamet): A GRASS GIS approach
June 13, 2024 36 / 40

UG
Uganda
Figure 47: Topographic map of Uganda
Highlights of Uganda
T Tectonic setting: The Great Rift Valley of East Africa Rift
System (EARS) – a unique combination of graben basins
forming a complex Afro-Arabian rift system.
T Geographic location: Uganda is located in the eastern part
of the EARS => diverse geomorphology, consisting of
volcanic hills, mountains, dense hydrologic network of
lakes and river basins.
T Geology: The Albertine Graben is a geologically important
region of Uganda and one of the most petroliferous rifts in
Africa.
T Mountainous geomorphology: Rwenzori mountains – the
highest non-volcanic, non-orogenic mountains in the world.
T Biodiversity high biodiversity of Rwenzori Mountains
National Park (World Heritage Site) => vegetation
ranging from tropical rainforest to alpine meadows.
T Landscape mixture: 5 overlapping vegetation zones in the
Ruwenzori Mts: the evergreen forests, the bamboo; the
heather; the alpine; and, the nival zone
T Climate change: impact of global warming and rise in T◦
on the Ruwenzori’s glaciers => snow and glaciers melt
T Environmental issues: deforestation. Reasons: population
growth, agricultural expansion, and the demand for natural
resources (timber and charcoal).
June 13, 2024 37 / 40

ZM
Zambia
Figure 49: Topographic location of Zambia
Zambia: geographic and environmental features
Climate: humid subtropical or tropical wet and dry
(Source: Köppen climate classification).
Lication: Landlocked country in southern Africa which
influences types of vegetation
Biodiversity: ecosystems in Zambia include forest, thicket,
woodland and grassland vegetation
Geomorphology: high plateaus with some hills and
mountains, dissected by river valleys
GMT-based mapping
Scripting techniques have the advantage of
increased speed, accuracy and precision of
mapping which enables rapid EO and RS data
processing
Validity and utility of data models in GMT
Embedded mathematical algorithms of data
modelling and cartographic principles
GMT => deeper understanding of the underlying
processes affecting the Earth’s landscapes across
Africa
With increasing rise and constant updates in EO
data availability in environmental analysis, it is
essential to apply data-driven mapping and
script-based cartographic visualisation
ML facilitates cartographic workflow
Processing large volumes of heterogeneous and
rapidly arriving digital geodata using scripts and
ML supports GIS.
GMT-based mapping uses scripts which optimises
cartographic workflow
June 13, 2024 38 / 40

Con
Conclusions
Monitoring landscapes of Africa using EO data and integration of advanced cartographic tools
3 African landscape are very complex. In this research proposal, selected examples of these ecosystems were presented to
illustrate current state of the research
3 Next steps in the proposed research include publishing 3-4 articles (reviews) that will summarise the variability of African
landscapes with focus on climate change issues, environmental dynamics, anthropogenic impacts, topographic-geologic
setting which shape the face of the African countries and help get more insights in its unique landscapes
3 In the era of big EO data, these research aims at integrated monitoring of African ecosystems using advanced cartographic
tools such as GMT scripts, R and Python for image processing (computing vegetation indices, image segmentation) and ML
algorithms for image classification
3 ML methods of GRASS GIS were used with a case of Random Forest for the cases of Mozambique land cover types
3 The demands from environmental landscape analysis for effective methods of EO data processing is explained by the
overwhelming increase of data that should be processed rapidly yet effectively for the scale of African continent
3 The study presented cartographic challenges faced in large-scale environmental approach: A case of diverse landscape and
ecoregions of Africa
3 Benefits of scripting tools and programming algorithms used for cartographic data processing are discussed
3 Advanced software ensure processing of multi-source data with diverse formats (export, converting and import via GDAL),
diverse geospatial data type (raster, vector and tabular), high processing speed, and mapping accuracy
3 Cartographic development opportunities in the era of big EO data are related to the machine-based data processing, that is
scripting and algorithms of automatisation of satellite image processing and classification
3 Modelling landscape of Africa starts with diversified steps of processing and includes advanced algorithms of data handling
3 GRASS GIS, GMT, Python and R – all these software present a smart methodological combination for RS data processing
3 The use of such technologies in landscape analysis results in the accurate processing large amounts of geospatial data
including satellite images which enable to reveal complex heterogeneous characteristics of the African landscapes
June 13, 2024 39 / 40

Thanks
Thanks
Thank you for attention !
Questions ?
June 13, 2024 40 / 40

References
Author’s publications
[1] P. Lemenkova. “Landscape Fragmentation and Deforestation in Sierra Leone, West Africa, Analysed Using Satellite Images”. In: Transylvanian
Review of Systematical and Ecological Research 26 (1 June 2024), pp. 13–26. issn: 2344-3219. doi: 10.5281/zenodo.11473307. url:
https://magazines.ulbsibiu.ro/trser/trser26/trser_26.1-2024-contents.pdf.
[2] P. Lemenkova. “Exploitation d’images satellitaires Landsat de la région du Cap (Afrique du Sud) pour le calcul et la cartographie d’indices de
végétation à l’aide du logiciel GRASS GIS”. In: Physio-Géo 20 (May 2024), pp. 113–129. issn: 1958-573X. doi: 10.4000/11pyj. url:
https://journals.openedition.org/physio-geo/17018.
[3] P. Lemenkova. “Random Forest Classifier Algorithm of Geographic Resources Analysis Support System Geographic Information System for Satellite
Image Processing: Case Study of Bight of Sofala, Mozambique”. In: Coasts 4 (1 Feb. 2024), pp. 127–149. doi: 10.3390/coasts4010008.
[4] P. Lemenkova. “An automated algorithm of GRASS GIS to retrieve the data on land cover types in Guinea, West Africa, from Landsat-8 OLI/TIRS
images”. In: Ovidius University Annals of Constanta - Series: Civil Engineering 25 (1 Dec. 2023), pp. 19–36. doi: 10.5281/zenodo.10673287.
[5] P. Lemenkova. “Using open-source software GRASS GIS for analysis of the environmental patterns in Lake Chad, Central Africa”. In: Die
Bodenkultur: Journal of Land Management, Food and Environment 74 (1 Nov. 2023), pp. 49–64. doi: 10.2478/boku-2023-0005.
[6] P. Lemenkova. “Monitoring Seasonal Fluctuations in Saline Lakes of Tunisia Using Earth Observation Data Processed by GRASS GIS”. In: Land 12
(11 Oct. 2023), p. 1995. doi: 10.3390/land12111995.
[7] P. Lemenkova. “Image Segmentation of the Sudd Wetlands in South Sudan for Environmental Analytics by GRASS GIS Scripts”. In: Analytics 2 (3
Sept. 2023), pp. 745–780. doi: 10.3390/analytics2030040.
[8] P. Lemenkova. “Mapping Wetlands of Kenya Using Geographic Resources Analysis Support System (GRASS GIS) with Remote Sensing Data”. In:
Transylvanian Review of Systematical and Ecological Research 25 (2 July 2023), pp. 1–18. doi: 10.2478/trser-2023-0008.
[9] P. Lemenkova. “A GRASS GIS Scripting Framework for Monitoring Changes in the Ephemeral Salt Lakes of Chotts Melrhir and Merouane, Algeria”.
In: Applied System Innovation 6 (4 June 2023), p. 61. doi: 10.3390/asi6040061.
[10] P. Lemenkova. “Command-Line Cartographic Data Processing for Geophysical Plotting of Rwanda Using GMT and R Scripts”. In: Romanian Journal
of Physics 67 (7–8 Sept. 2022), pp. 1–21. doi: 10.5281/zenodo.7113414.
[11] P. Lemenkova. “Evapotranspiration, vapour pressure and climatic water deficit in Ethiopia mapped using GMT and TerraClimate dataset”. In:
Journal of Water and Land Development 54 (7–9 Sept. 2022), pp. 201–209. doi: 10.24425/jwld.2022.141573.
[12] P. Lemenkova. “Mapping Climate Parameters over the Territory of Botswana Using GMT and Gridded Surface Data from TerraClimate”. In: ISPRS
International Journal of Geo-Information 11 (9 Aug. 2022), p. 473. doi: 10.3390/ijgi11090473.
[13] P. Lemenkova. “Cartographic Scripting for Geophysical Mapping of Malawi Rift Zone”. In: Tehnika 77 (2 May 2022), pp. 183–191. doi:
10.5937/tehnika2202183L.
[14] P. Lemenkova. “Mapping Ghana by GMT and R scripting: advanced cartographic approaches to visualize correlations between the topography,
climate and environmental setting”. In: Advances in Geodesy and Geoinformation 71 (1 May 2022). Article no. e16, pp. 1–20. doi:
10.24425/gac.2022.141169.
June 13, 2024 40 / 40

References
[15] P. Lemenkova. “Seismicity in the Afar Depression and Great Rift Valley, Ethiopia”. In: Environmental Research, Engineering and Management 78 (1
Apr. 2022), pp. 83–96. doi: 10.5755/j01.erem.78.1.29963.
[16] P. Lemenkova. “Tanzania Craton, Serengeti Plain and Eastern Rift Valley: mapping of geospatial data by scripting techniques”. In: Estonian Journal
of Earth Sciences 71 (2 Apr. 2022), pp. 61–79. doi: 10.3176/earth.2022.05.
[17] P. Lemenkova. “Geophysical Mapping of Ghana Using Advanced Cartographic Tool GMT”. In: Kartografija i Geoinformacije 20 (36 Dec. 2021),
pp. 16–37. doi: 10.32909/kg.20.36.2.
[18] P. Lemenkova. “Mapping Earthquakes in Malawi Using Incorporated Research Institutions for Seismology (IRIS) Catalogue for 1972–2021”. In:
Malawi Journal of Science and Technology 13 (2 Dec. 2021), pp. 31–50. doi: 10.5281/zenodo.5772315.
[19] P. Lemenkova. “Evaluating the Performance of Palmer Drought Severity Index (PDSI) in Various Vegetation Regions of the Ethiopian Highlands”.
In: Acta Biologica Marisiensis 4 (2 Dec. 2021), pp. 14–31. doi: 10.2478/abmj-2021-0010.
[20] P. Lemenkova. “Application of scripting cartographic methods to geophysical mapping and seismicity in Rwenzori mountains and Albertine Graben,
Uganda”. In: Makerere University Journal of Agricultural and Environmental Sciences 10 (1 July 2021), pp. 1–21. doi: 10.5281/zenodo.5082861.
[21] P. Lemenkova. “Scripting methods in topographic data processing on the example of Ethiopia”. In: SINET Ethiopian Journal of Science 44 (1 June
2021), pp. 91–107. doi: 10.4314/sinet.v44i1.9.
[22] P. Lemenkova. “Mapping environmental and climate variations by GMT: a case of Zambia, Central Africa”. In: Zemljište i biljka 70 (1 May 2021),
pp. 117–136. doi: 10.5937/ZemBilj2101117L.
[23] P. Lemenkova. “Object Based Image Segmentation Algorithm of SAGA GIS for Detecting Urban Spaces in Yaoundé, Cameroon”. In: Central
European Journal of Geography and Sustainable Development 2 (2 Dec. 2020), pp. 38–51. doi: 10.47246/CEJGSD.2020.2.2.4.
[24] P. Lemenkova and O. Debeir. “Time Series Analysis of Landsat Images for Monitoring Flooded Areas in the Inner Niger Delta, Mali”. In: Artificial
Satellites 58 (4 Dec. 2023), pp. 278–313. doi: 10.2478/arsa-2023-0011.
[25] P. Lemenkova and O. Debeir. “Environmental mapping of Burkina Faso using TerraClimate data and satellite images by GMT and R scripts”. In:
Advances in Geodesy and Geoinformation 72 (2 Oct. 2023), pp. 1–32. doi: 10.24425/agg.2023.146157.
[26] P. Lemenkova and O. Debeir. “Multispectral Satellite Image Analysis for Computing Vegetation Indices by R in the Khartoum Region of Sudan,
Northeast Africa”. In: Journal of Imaging 9 (5 May 2023), p. 98. doi: 10.3390/jimaging9050098.
[27] P. Lemenkova and O. Debeir. “Recognizing the Wadi Fluvial Structure and Stream Network in the Qena Bend of the Nile River, Egypt, on Landsat
8-9 OLI Images”. In: Information 14 (4 Apr. 2023), p. 249. doi: 10.3390/info14040249.
[28] P. Lemenkova and O. Debeir. “Coherence of Bangui Magnetic Anomaly with Topographic and Gravity Contrasts across Central African Republic”.
In: Minerals 13 (5 Apr. 2023), p. 604. doi: 10.3390/min13050604.
[29] P. Lemenkova and O. Debeir. “Computing Vegetation Indices from the Satellite Images Using GRASS GIS Scripts for Monitoring Mangrove Forests
in the Coastal Landscapes of Niger Delta, Nigeria”. In: Journal of Marine Science and Engineering 11 (4 Apr. 2023), p. 871. doi:
10.3390/jmse11040871.
June 13, 2024 40 / 40

Bibliography
[30] P. Lemenkova and O. Debeir. “Satellite Altimetry and Gravimetry Data for Mapping Marine Geodetic and Geophysical Setting of the Seychelles and
the Somali Sea, Indian Ocean”. In: Journal of Applied Engineering Sciences 12(25) (2 Dec. 2022), pp. 191–202. doi: 10.2478/jaes-2022-0026.
[31] P. Lemenkova and O. Debeir. “R Libraries for Remote Sensing Data Classification by k-means Clustering and NDVI Computation in Congo River
Basin, DRC”. In: Applied Sciences 12 (24 Dec. 2022), p. 12554. doi: 10.3390/app122412554.
[32] P. Lemenkova and O. Debeir. “Satellite Image Processing by Python and R Using Landsat 9 OLI/TIRS and SRTM DEM Data on Côte d’Ivoire,
West Africa”. In: Journal of Imaging 8 (12 Nov. 2022), p. 317. doi: 10.3390/jimaging8120317.
June 13, 2024 40 / 40

Mapping landscapes of Africa using remote sensing data: detecting spatio-temporal environmental dynamics from the satellite images

More Related Content

Similar to Mapping landscapes of Africa using remote sensing data: detecting spatio-temporal environmental dynamics from the satellite images (20)

More from Universität Salzburg (20)

Recently uploaded (20)

Mapping landscapes of Africa using remote sensing data: detecting spatio-temporal environmental dynamics from the satellite images