land

Review
A Review on Land Use and Land Cover Change in
Ethiopian Basins
Motuma Shiferaw Regasa 1 , Michael Nones 1, * and Dereje Adeba 2

1 Department of Hydrology and Hydrodynamics, Institute of Geophysics Polish Academy of Sciences,

01-452 Warsaw, Poland; [email protected]
2 Department of Hydraulic and Water Resources Engineering, College of Engineering and Technology,
Wollega University, PO Box 395 Nekemte, Ethiopia; [email protected]
* Correspondence: [email protected]

Abstract: Land Use Land Cover (LULC) changes analysis is one of the most useful methodologies to
understand how the land was used in the past years, what types of detections are to be expected in the
future, as well as the driving forces and processes behind these changes. In Ethiopia, Africa, the rapid
variations of LULC observed in the last decades are mainly due to population pressure, resettlement
programs, climate change, and other human- and nature-induced driving forces. Anthropogenic
activities are the most significant factors adversely changing the natural status of the landscape
and resources, which exerts unfavourable and adverse impacts on the environment and livelihood.
The main goal of the present work is to review previous studies, discussing the spatiotemporal
LULC changes in Ethiopian basins, to find out common points and gaps that exist in the current
literature, to be eventually addressed in the future. A total of 25 articles, published from 2011
to 2020, were selected and reviewed, focusing on LULC classification using ArcGIS and ERDAS



imagine software by unsupervised and maximum likelihood supervised classification methods.


Key informant interview, focal group discussions, and collection of ground truth information using
Citation: Regasa, M.S.; Nones, M.; ground positioning systems for data validation were the major approaches applied in most of the
Adeba, D. A Review on Land Use and studies. All the analysed research showed that, during the last decades, Ethiopian lands changed
Land Cover Change in Ethiopian from natural to agricultural land use, waterbody, commercial farmland, and built-up/settlement.
Basins. Land 2021, 10, 585. https://
Some parts of forest land, grazing land, swamp/wetland, shrubland, rangeland, and bare/ rock out
doi.org/10.3390/land10060585
cropland cover class changed to other LULC class types, mainly as a consequence of the increasing
anthropogenic pressure. In summary, these articles confirmed that LULC changes are a direct result
Academic Editor: Francisco
of both natural and human influences, with anthropogenic pressure due to globalisation as the main
Manzano Agugliaro
driver. However, most of the studies provided details of LULC for the past decades within a specific
Received: 10 May 2021 spatial location, while they did not address the challenge of forecasting future LULC changes at the
Accepted: 31 May 2021 watershed scale, therefore reducing the opportunity to develop adequate basin-wide management
Published: 1 June 2021 strategies for the next years.

Publisher’s Note: MDPI stays neutral Keywords: Africa; Ethiopia; geographic information systems; land use land cover change; re-
with regard to jurisdictional claims in mote sensing
published maps and institutional affil-
iations.

1. Introduction
Land use is defined as how the land is utilised by people and their habitats, usu-
Copyright: © 2021 by the authors. ally with an accent on a functional role of land for economic activities, whereas land
Licensee MDPI, Basel, Switzerland. cover is a physical characteristic of the Earth’s surface [1,2]. Land use land cover (LULC)
This article is an open access article dynamics are a well-known, accelerating, and substantial process, mostly driven by hu-
distributed under the terms and man activities, that is contributing significantly to forest fragmentation, land degradation,
conditions of the Creative Commons
and biodiversity loss [3,4]. Land use land cover changes (LULCCs) analysis is one of the
Attribution (CC BY) license (https://
most used techniques to understand how the land was used in the past years, what types
creativecommons.org/licenses/by/
of detections are to be expected in the future, as well as the driving forces and processes
4.0/).

Land 2021, 10, 585. https://doi.org/10.3390/land10060585 https://www.mdpi.com/journal/land

Land 2021, 10, 585 2 of 18

behind these changes [3,4]. Besides natural variations, the increasing human population is
driving modifications of the Earth’s land surface that are unprecedented, and evidence is
present on a global scale [5]. Therefore, there is the need for better evaluation of changes
in the land cover (namely, the biophysical attributes of the Earth’s surface) and land use
for human purposes to understand the past variations and depict future trends for the
coming decades.
LULCCs are so persistent that, when aggregated globally, they expressively affect
strategic aspects of Earth’s system functioning. They directly impact biotic diversity
worldwide, contribute to local and regional climate change as well as global climate
warming, and are one of the primary sources of soil degradation. By altering ecosystem
services at the local and regional scales, LULCCs affect the ability of biological systems
to support and adapt to human needs [6–8]. Indeed, the major modifications of LULC
worldwide could be related to the intense agricultural development and the growing
population [9].
Similar to the rest of the world, East Africa (Horn of Africa) is not an exception to
these land use land cover changes (e.g., [10–13]). In particular, very rapid changes are
clearly recognisable in Ethiopia, due to the population pressure, resettlement programs,
climate change, and other human and nature-induced driving forces. Similar to other
countries, anthropogenic activities are the most significant factors adversely changing the
natural status of the Ethiopian landscape [12], involving detrimental and adverse impacts
on the natural environment and livelihood [14–16]. The land is a critical resource for the
livelihood of East Africans, and there has been a steady decline in the size of land holdings
per household. Following the demand for land, LULCCs in this region have resulted in
a decline of natural forests to human settlements, urban centres, farmlands, and grazing
lands [17]. Between 1990 and 2015, the East African forest cover decreased annually by
about 1%, while the human population increased at an average annual rate of around
2% [10]. As pointed out by Dibaba et al. [18], factors such as biophysical, socioeconomic,
institutional, technological, and demographic, contributed to LULCCs, which leads to a
decline in the agricultural yield and a loss of biodiversity in the entire upper Blue Nile
Basin, but significantly in the Finchaa sub-basin in the Oromia Regional State, Ethiopia.
The authors also pointed out that extended aridity and persistent drought, land and soil
degradation, as well as the decline of water resources in general, are the major consequences
of LULCCs at the regional scale.
Forest disturbance and the subsequent conversion to other LULC classes (such as
grazing land, agricultural land, bare land, pasture, or settlement areas) could modify the hy-
drologic cycle at the local scale, involving significant effects on water yields, water quality,
and streamflow dynamics [19,20]. The rapid rate of deforestation is mainly occurring be-
cause of several reasons such as unsustainable large and small scale agriculture, forest fire,
migration and population growth, illegal logging for construction purposes, charcoal and
fuelwood production for cooking, and poor resource management [21], namely, defor-
estation, which is connected to the increased occurrence of shifting cultivators, triggering
mechanisms that invariably involve changes in land development and new policies by the
national governments that push migrants into sparsely occupied areas [8]. Focusing on
the Horn of Africa, the main forest types that have undergone this decrease are tropical
rain and dry forests, tropical shrubs, tropical maintain the forest, and mangrove forests,
while there have been intensive efforts to establish plantation forests [10]. Land policy in
developing countries such as Ethiopia is considered a crucial part of the overall develop-
ment policy that the national governments need for assuring rapid economic growth and
poverty mitigation, regardless of the natural resources management [22].
Ethiopia is historically passed significant dynamics in LULC for many decades. How-
ever, nowadays, LULCCs and degradation are increasing at an alarming rate, playing a
significant role in the increasing rate of soil erosion. The need for more cultivated lands
has negatively affected the presence of forest and grasslands, eventually fostering soil
erosion [23]. Environmental conversions and changes can be mainly attributed to various
 Land 2021, 10, 585 3 of 18

adverse human actions such as the expansion of farm plots at the expense of agricultural
lands, massive fuelwood and charcoal production, overgrazing, and encroachment of farm-
steads into vegetated lands. According to Tefera [12], ecologically, Ethiopia is characterised
by a rich but shrinking diversity in biological resources such as forest, woody and grassy
lands, shrubs, varied wildlife, and fertile soil. It is also renowned for its massive mountain
ranges, high flat plateaus, deep gorges, river valleys, lowland plains, extensive wetlands,
and deserts. Landscape degradation by soil erosion has increased considerably in the
Ethiopian highlands since the deforestation of the natural mountain forests and the cultiva-
tion of large areas, resulting in serious danger to the Ethiopian population [24,25]. This also
affects the water balance of an area by changing the balance between rainfall, evaporation,
infiltration, and runoff. Based on the observed trends, it is clear that a systematic analysis
of LULCC is crucial to exactly comprehend the extent of the changes and take necessary
measures to scale down the soil erosion [16], rate of changes, and protect the land cover
resources sustainably.
The main objective of this article is to review the actual literature on LULC in Ethiopian
basins, to point out what are the existing situations and the research gap that should be
1, 10, x FOR PEER REVIEW 4 of 20
addressed in the future.

2. Case Study: Ethiopia

production has beenEthiopia
growingisbylocated
about in theduring
2.3% north-eastern part1980–2000,
the period of the Africanwhilecontinent,
the pop-in the so-called
Horn of Africa, which ◦ and 18◦ north latitude, 33◦ and 48◦ east longitude,
ulation was growing at an average ratelies
of between
2.9% per 3year, leading to a decline in per capita
within the
agricultural production bytropics (Figure
about 0.6% per1). The[15].
year totalAccording
area of thetocountry is 1,119,683
this report, km2 , while the area
the percent-
2
age of people inoccupied
Ethiopia by waterbodies
who were absolutelyis 7,444 kmin[26].
poor Ethiopia
the year 2001 wasis a country where
around 44%, about more than
but
the level of poverty shows significant variation among rural, urban areas and across re- sorts of natural
80 million people, containing 50.46% male; the country is grappling with all
gional states. Inand
this manmade
country, theproblems such as famine,
income distribution seemsenvironmental degradation,
to be more unequally distrib-erratic rainfalls,
uted in rural and urban areas, compared to other sub-Saharan African regions. To tackleand widespread
the prevalence of malaria and HIV/AIDS, poor but improving governance,
poverty.
this situation, in Aboutthe
recent times, 84% of the people
Ethiopian Ministrylive
of in rural areas,
Agriculture andassuring their livelihoods with
Rural Develop-
ment (MoARD)subsistence
announced agriculture,
one of the most which is a sector
detailed nowadays suffering
agro-ecological arrangementsfromofthe
thelack of essential
inputs and a very variable rainfall pattern. Poverty is more
country by taking also soil moisture regimes into account, in addition to altitude and tem- than common in Ethiopia,
perature. though slightly declining over time [12].

Figure 1. Maps of Africa (on the left) and Ethiopia (on the right).
Figure 1. Maps of Africa (on the left) and Ethiopia (on the right).
3. Methodology In terms of geography, the prominent features of Ethiopia are the extensive high
The present worksurging
lands, is based on a review
plateaus, conducted
and deep on peer-reviewed
river canyons, and the Greatarticles pub- Rift System,
East African
lished in the lastdividing
10 years, from
the 2011into
country to 2020, which were focused
the central/western part, on the issue
mostly of land use
mountainous, and the southern
land cover change in Ethiopian basins, focusing on nonurban environments. The search
was based on the exact phrase of ‘land use land cover changes in Ethiopia’ in the Web of
Google scholars, which searches within each article’s title, abstract, keywords, years of
publication, and ‘keywords plus’, a series of additional relevant keywords selected by
well-known databases such as Web of Science, Scopus, and Google Scholars. Based on
Land 2021, 10, 585 4 of 18

highlands, which are surrounded by lowlands [27]. As indicated by Tefera [12], about 45%
of the country is highland, with an average altitude greater than 1500 m and peaks of
around 4000 m, in which about 88% of the country’s population is located. Overpopulation,
extensive croplands, and frequent incision by ravines and gullies characterise the highlands.
On the basis of altitude, its influence on temperature, and rainfall, Ethiopia is traditionally
classified into four broad agro-climatic zones. These are termed wurch (cold moist),
dega (cool humid), woina dega (semi humid), and qolla (arid and semiarid). The wurch
region encompasses all areas located around 3200 m above the mean sea level, with an
average annual rainfall of over 22 mm. The dega zone consists of areas with altitudes and
an average annual rainfall ranging from 2400 to 3200 m, and 1200 to 2200 mm, respectively.
The woina dega zone covers areas within the altitudinal range of 1500 to 2400 m, having an
average annual rainfall of 800 to 1200 mm. The qolla zone refers to areas lying below the
altitude of 1500 m, where the average annual rainfall is around 800 mm [12]. In addition to
these four regions, the Ethiopian physical environment can further be classified into eleven
more detailed groups, still depending on average altitude and annual rainfall: bereha
(namely, desert), dry qolla, moist qolla, wet qolla, moist woina dega, wet-woina dega,
moist dega, wet dega, moist wurch, wet wurch, and high wurch.
The Ethiopian economy is among the most vulnerable in sub-Saharan Africa, and it
is heavily dependent on the agricultural sector, which has suffered from the recurrent
droughts that are reflected in extreme fluctuations of outputs. For example, agricultural
production has been growing by about 2.3% during the period 1980–2000, while the
population was growing at an average rate of 2.9% per year, leading to a decline in per
capita agricultural production by about 0.6% per year [15]. According to this report,
the percentage of people in Ethiopia who were absolutely poor in the year 2001 was around
44%, but the level of poverty shows significant variation among rural, urban areas and
across regional states. In this country, the income distribution seems to be more unequally
distributed in rural and urban areas, compared to other sub-Saharan African regions.
To tackle this situation, in recent times, the Ethiopian Ministry of Agriculture and Rural
Development (MoARD) announced one of the most detailed agro-ecological arrangements
of the country by taking also soil moisture regimes into account, in addition to altitude
and temperature.

3. Methodology
The present work is based on a review conducted on peer-reviewed articles published
in the last 10 years, from 2011 to 2020, which were focused on the issue of land use land
cover change in Ethiopian basins, focusing on nonurban environments. The search was
based on the exact phrase of ‘land use land cover changes in Ethiopia’ in the Web of Google
scholars, which searches within each article’s title, abstract, keywords, years of publication,
and ‘keywords plus’, a series of additional relevant keywords selected by well-known
databases such as Web of Science, Scopus, and Google Scholars. Based on these criteria,
17 articles were selected and systematically reviewed to find out the strength and the
research gaps in the study of LULCC in Ethiopian watersheds (Table 1).

Table 1. Summary of the revised literature.

Article Num. Authors Title Year

Land Use Change Analysis Using Remote Sensing
1 Ayana, A.A.; Kositsakulchai, E. and Markov Modelling in Fincha 2012
Watershed, Ethiopia
Drivers and Implications of Land Use/Land Cover
2 Dibaba, W.T.; Demissie, T.A.; Miege, K. Dynamics in Finchaa Catchment, 2020
North-western Ethiopia
Modelling the Effects of Land Use Change and
Ayana, A.B.; Edossa, D.C.;
3 Management Practices on Runoff and Sediment 2014
Kositsakulchai, E.
Yields in Fincha Watershed, Blue Nile
Land 2021, 10, 585 5 of 18

Table 1. Cont.

Article Num. Authors Title Year

Betrua, T.; Tolera, M.; Sahleb. K.; Trends and Drivers of Land Use/Land Cover
4 2019
Kassac, H. Change in Western Ethiopia
Effects of Land Use and Land Cover Change on Soil
Mariye, M.; Mariyo, M.; Changming, Y.;
5 Erosion Potential in Berhe District: A Case Study of 2020
Teffera, Z.L.; Weldegebrial, B.
Legedadi Watershed, Ethiopia
Drought Vulnerability Drives Land Use and Land
6 Biazina, B.; Sterk, G. Cover Changes in the Rift Valley Dry Lands 2012
of Ethiopia
Dynamics of Land Use, Land Cover Change Trend
7 Hailua, A.; Mammo, S.; Kidane, M. and Its Drivers in Jimma Geneti District, 2020
Western Ethiopia
Land Use–Land Cover Dynamics of Huluka
8 Gebreslassie, H. 2014
Watershed, Central Rift Valley, Ethiopia
The Impact of Land Use/Land Cover Change on
9 Tolessa, T.; Senbeta, F.; Kidane, M. Ecosystem Services in the Central Highlands 2017
of Ethiopia
Land Use and Land Cover Changes and Associated
Alemu, B.; Garedew, E.; Eshetu, Z.;
10 Driving Forces in North-western Lowlands 2015
Kassa, H.
of Ethiopia
Analysis of Land Use/Land Cover Changes in the
Fisseha, G.; Gebrekidan, H.; Kibret, K.;
11 Debre-Mewi Watershed at the Upper Catchment of 2011
Yitaferu, B.; Bedadi, B.
the Blue Nile Basin, Northwest Ethiopia
Land Use/Cover Change Analysis and Local
Community Perception towards Land Cover Change
12 Mussa, M.; Teka, H.; Mesfin, Y. 2017
in the Lowland of Bale Rangelands,
Southeast Ethiopia
Land Use Land Cover Change Trend and Its Drivers
13 Alemayehu, F.; Tolera, M.; Tesfaye, G. 2019
in Somodo Watershed Southwestern, Ethiopia
Land Use/Land Cover Dynamics in Response to
Tolessa, T.; Dechassa, C.; Simane, B.;
14 Various Driving Forces in Didessa 2020
Alamerew, B.; Kidane, M.
Sub-Basin, Ethiopia
Land Use/Land Cover Dynamics in Nonno District,
15 Tefera, M.M. 2011
Central Ethiopia
Effect of Land Use–Land Cover Change on the
16 Dinka, M.O.; Klik, A. Regimes of Surface Runoff—The Case of Lake 2019
Basaka Catchment (Ethiopia)
Analysing the Rate of Land Use and Land Cover
Change and Determining the Causes of Forest Cover
17 Othow, O.O.; Gebre, S.L.; Gemeda, D.O. 2017
Change in Gog District, Gambella Regional
State, Ethiopia
Berihu, L.B.; Tsunekawa, A.;
Exploring Land Use/Land Cover Changes, Drivers,
Haregeweyn, N.; Meshesha, B.T.; Adgo,
18 and Their Implications in Contrasting 2019
E.; Tsubo, M.; Masunaga, T.; Fenta, A.A.;
Agro-Ecological Environments Of Ethiopia
Sultan, D.; Yibeltal, M.
Evaluation and Prediction of Land Use/Land Cover
Gashaw, T.; Tulu, T.; Argaw, M.; Worqlul,
19 Changes in the Andassa Watershed, Blue Nile 2017
A.W.
Basin, Ethiopia
Impact of Land Use Land Cover Change on Stream
20 Andualem, T.G.; Gebremariam, B. Flow and Sediment Yield: A Case Study of Gilgel
Abay Watershed, Lake Tana Sub-Basin, Ethiopia
The Impact of Land Use Change on the Hydrology
21 Getachew, H.E.; Melesse, A.M. 2012
of the Angereb Watershed, Ethiopia
Land 2021, 10, 585 6 of 18

Table 1. Cont.

Article Num. Authors Title Year

Analyses of Land Use and Land Cover Change
Dynamics Using GIS and Remote Sensing During
22 Meshesha, T.W.; Tripathi, S.K.; Khare, D. 2016
1984 and 2015 in the Beressa Watershed Northern
Land 2021, 10, x FOR PEER REVIEW Central Highland of Ethiopia 6 of 20

Land Use/Land Cover Changes and Their

23 Miheretu, B.A,; Yimer, A.A. Environmental Implications in the Gelana
Sub-Watershed
Land Use/Land of Northern
Cover Changes Highlands
and Their of Ethiopia
Environmental
23 Miheretu, B.A,; Yimer, A.A. LandImplications in theDynamics
Use/Land Cover Gelana in the Central Rift
24 Molla, M.B. Sub-Watershed Valley Region ofHighlands
of Northern Ethiopia: Case of Arsi
of Ethiopia 2015
Negele District
Land Use/Land Cover Dynamics in the Central Rift Valley
24 Molla, M.B. 2015
RegionLand
of Ethiopia:
Use and Case
Landof Arsi Change
Cover Negele District
and Implication
Land Usetoand
Watershed
Land Degradation
Cover Change by Using
and GIS and to
Implication Remote
25 Sewnet, A.; Abebe, G. 2018
25 Sewnet, A.; Abebe, G. Watershed Degradation Sensing in the
by Using GISKoga
and Watershed,
Remote Sensing in 2018
North-western Ethiopia
the Koga Watershed, North-western Ethiopia

In
In terms ofgeographical
terms of geographicaldistribution,
distribution,
thethe considered
considered articles
articles covercover a large
a large part ofpart
theof the
country, spanning
country, spanning different
different agro-climatic
agro-climaticzones
zones(Figure
(Figure2).2).Since nono
Since articles that
articles referred
that referred to
to LULCCs
LULCCs inin theeastern
the easternpart
part were
were found,
found,the
theoutcomes
outcomes summarised
summarised in the
inpresent studystudy
the present
can be
can be considered
considered representative
representativeofofa general trend
a general of LULCCs
trend of LULCCs in Ethiopia, except
in Ethiopia, for the
except for the
Ogaden Region.
Ogaden Region.

Figure 2. Main locations of the studies. Some reviewed articles referred to the same watershed,
Figure 2. Main
hence are locationsonly
here reported of the studies.
once. Some reviewed
The Ethiopian borders articles referred
are indicated to athe
with same
black watershed, hence
line.
are here reported only once. The Ethiopian borders are indicated with a black line.
Due to not being yet standardised, the LULC classification method is highly subjec-
tive and can change from one study to the other (Table 2).

Table 2. Characteristics of the study area for the selected case studies in Ethiopia.
Land 2021, 10, 585 7 of 18

Due to not being yet standardised, the LULC classification method is highly subjective
and can change from one study to the other (Table 2).

Table 2. Characteristics of the study area for the selected case studies in Ethiopia.

Article Ethiopian
Study Site Coordinates Study Period Model Method
Num. Region
ERDAS
Fincha between 9◦ 100 05 ”–10◦ 000 59 ” supervised maximum
1 Oromiyaa 1985–2005 Imagine
watershed N and 37◦ 000 16 ”–37◦ 330 20 ” E likelihood method
software
Fincha between 9◦ 100 05 ”–10◦ 000 59 ” supervised maximum
2 Oromiyaa 1987–2017 ArcGIS
watershed N and 37◦ 000 16 ”–37◦ 330 20 ” E likelihood method
combination of
ERDAS supervised classification
Fincha between 9◦ 100 05 ”–10◦ 000 59
”
3 Oromiyaa 2005 Imagine based on the minimum
watershed N and 37◦ 000 16 ”–37◦ 330 20 ” E
software distance algorithm
method
hybrid supervised
classified and intensive
Benishangul between 9◦ 330 –10◦ 540
N and
4 Asosa zone 1978–2016 not described on–screen digitizing
Gumuz 34◦ 080 –35◦ 210 E
(visual image
interpretation) method
Legedadi between 9◦ 010 –9◦ 130 N and ERDAS supervised maximum
5 Oromiyaa 1985–2013
watershed 38◦ 600 –39◦ 070 E Imagine 9.2 likelihood method
ERDAS
supervised maximum
6 Rift Valley Oromiyaa 1965–2010 Imagine and
likelihood method
ArcGIS
ERDAS
Jimma
imagine 14.1 supervised maximum
7 Geneti Oromiyaa 1973− 2019
and ArcGIS likelihood method
district
10.3
combination of
Huluka between 7◦ 16.40 –7◦ 30.70 N and ERDAS unsupervised and
8 Oromiyaa 1973–2009
watershed 38◦ 47.70 –38◦ 44.30 E Imagine supervised maximum
likelihood method
Chillimo ERDAS supervised maximum
9 Oromiyaa 38◦ 100 E–9◦ 050 N 1973–2015
Forest imagine.10 likelihood method
between 13◦ 390 46.5
”–14◦ 260 34.9 ” N and 36◦ 270 4.7
Kafta ”–37◦ 330 7.1 ”E; between
combination of
Humera, Tigray, Amhara, 12◦ 170 33.63 ”–13◦ 50 52.52 ” N
unsupervised and
10 Metema and Benis- and 1985–2010 ArcGIS
supervised maximum
and hangulGumuz 35◦ 450 21.34 ”–36◦ 450 31.31 ” E;
likelihood method
Shorkole between 10◦ 260 18.98
”–11◦ 140 25.65 ” N and
34◦ 450 21.33 ”–35◦ 450 31.30 ” E;
Debre- visual interpretation and
between 11◦ 200 –11◦ 210 N and ArcGIS 9.2
11 Mewi Amhara 1957–2008 digitised based on false-
37◦ 240 –37◦ 250 E software
watershed colour composites method
combination of
ERDAS
Raitu between 6◦ 200 0 ”–7◦ 250 0 ” N unsupervised and
12 Oromiyaa 1986–2016 Imagine and
district and 41◦ 300 00 ”–42◦ 000 00 ” E supervised classification
ArcGIS 10.5
methods
both unsupervised and
Somodo between 7◦ 460 00 ”–7◦ 470 00 ” ERDAS
13 Oromiyaa 1985–2017 maximum likelihood
watershed N and 36◦ 470 00 ”–36◦ 480 00 ” E Imagine
supervised method
ERDAS
both unsupervised and
Didessa IMAGINE 2010
14 Oromiyaa 1974–2014 maximum likelihood
sub-basin and ArcGIS
supervised method
10.5
Nonno between 2◦ 540 –15◦ 180 N and ENVI 4.3 and
15 Oromiyaa 1984–2007 supervised classification
district 32◦ 420 –48◦ 180 E ArcGIS 9.3
Land 2021, 10, 585 8 of 18

Table 2. Cont.

Article Ethiopian
Study Site Coordinates Study Period Model Method
Num. Region
Lake unsupervised and
ERDAS
16 Basaka Oromiyaa 1973–2015 maximum likelihood
Imagine
catchment supervised method
between
ERDAS maximum likelihood
17 Gog district Gambella 7◦ 270 38 ”–8◦ 180 57 ” N and 1990–2017
Imagine supervised method
34◦ 140 59 ”–35◦ 330 49 ” E
Guder, Aba
visual interpretation
Gerima and
18 Oromiyaa 1982–2017 ArcGIS 10.4 on-screen digitisation
Debatie
technique
watersheds
ERDAS
both unsupervised and
Andassa between 11◦ 080 –11◦ 320 N and Imagine 2014
19 Amhara 1985–2045 supervised classification
watershed 37◦ 160 –37◦ 320 E and ArcGIS
methods
10.3
ERDAS
Gilgel Abay between 10◦ 560 –11◦ 510 N and
20 Amhara 1986–2011 Imagine and not mentioned
watershed 36◦ 440 –37◦ 230 E
ArcGIS
between ERDAS
Angereb maximum likelihood
21 Amhara 12◦ 360 22 ”–12◦ 430 34.8 ” N and 1985–2011 Imagine and
watershed supervised method
37◦ 250 2 ”–37◦ 300 28 ” E ArcGIS
ArcGIS10.2.2
between 9◦ 400 –9◦ 410 and maximum likelihood
22 Beressawatershed Amhara 1984–2015 and ERDAS
39◦ 370 –39◦ 320 E supervised method
Imagine14
between
Gelana ERDAS maximum likelihood
23 Amhara 11◦ 340 44 ”–11◦ 450 4 ” N and 1964–1984
sub-basin Imagine9.1 supervised method
39◦ 340 11 ”–39◦ 450 2 ” E
ERDAS
Arsi Negele between 7◦ 090 –7◦ 410 N and
24 Oromiyaa 1973–2010 Imagine 8.7 and Not mentioned
district 38◦ 250 –38◦ 540 E
ArcGIS
ArcGIS, IDL,
Koga between 11◦ 100 –11◦ 250 N and hybrid unsupervised and
25 Amhara 1973–2011 and ERDAS
watershed 370◦ 20 –370◦ 170 E supervised Method
Imagine

In the present analysis, land classes’ names which may have a similar meaning or
approach were considered as one. For example, the brushland class cover was considered
as a shrubland class cover. Moreover, in some cases, a single class was classified as a
combination of both, such as in the case of [28–30], where the settlement and agricultural
land cover classes were classified together and defined as the ‘settlement/agricultural’
class. For such a case, we selected only the dominating class. Therefore, the agricultural
class was chosen instead of the built-up/settlement one.
For those articles that did not describe the LULCCs for each class between the final
and initial year, the percentage of LULCCs was as follows:

( LULCt1 − LULCt0 )
LULCC = ∗ 100
A
where A is the total area of the study region, while t0 and t1 are the initial and final years,
respectively.
To compare the different articles, the LULCC percentages of each watershed were
determined as
PLCCC in P1 + PLCCC in P2 + PLCCC in P3 + . . . + PLCCC in P25
LULCCW =
total number o f articles in which the LULC was used

where W indicates the watershed, while PLCCC and P represent the percentage of land
class cover change and the article number (Table 1), respectively.
Land 2021, 10, 585 9 of 18

For example, agricultural LULCs were classified in all articles except in the work of
Ayana et al. [31]. Therefore, the magnitude of such LULCCs was determined depending
on the 16 articles, as

Ag1 + Ag2 + Ag4 + Ag5 + Ag6 + Ag7 + Ag8 + Ag9 + Ag10 + . . . + A25
124
where Ag is the agricultural land use, while the subscripts indicate the article number.
As another example, the percentage of waterbody LULCCs was calculated as follows:

Wb1 + Wb2 + Wb5 + Wb7 + Wb10 + Wb16 + Wb17 + Wb20 + Wb21 + Wb22 + Wb25
11
where Wb is the waterbody land use. From this, it can be seen that the waterbody LULCCs
are addressed in articles 1, 2, 5, 7, 10, 16, 17, 20, 21, 22, and 25 (see Table 1).

4. Article Analysis
In this section, we report briefly the main outcomes of each study (see Table 2),
aiming to provide the readers with a few more details needed for critical comparison.
According to Mariye et al. [2], the maximum likelihood classification method was used
to study LULC of the Lege Dadi watershed, for 1995, 1997, and 2013, by using the ERDAS
Imagine 9.2 software. The study was aimed to investigate the effects of LULCC on soil
erosion potential in the Berhe District, a small portion of the Lege Dadi Basin. To achieve
the objective of the study, Landsat satellite images were downloaded from the United
States Geological Survey (USGS) official website (earthexplorer.usgs.gov). The results
of classification show six LULC classes: water body (including human-made reservoir),
cultivated land, settlement, grazing land, forest (composed mainly of Eucalyptus globules
plantation), and bare land. Comparing the different years, the authors observed that cul-
tivated land, settlement, and forest increased, while a decrease was observed looking at
the areas covered by water bodies, grazing land, and bare land. Besides using remote im-
agery, the authors of this study conducted a series of focus group discussions with farmers,
development agents, cabinet members, elders, and knowledgeable community represen-
tatives to obtain further information about the long-term experience of LULC practices
in the watersheds. In their report, the settlement area and cultivated land were increased
significantly, whereas grazing land and bare land classes were reduced, confirming what
was retrieved from satellite data.
Similarly, in [29], a maximum likelihood classification method was used in combina-
tion with a geographic information system (GIS) to study the drought vulnerability drives
of LULCCs in the Rift Valley drylands of Ethiopia using aerial photographs, satellite im-
agery, rainfall, and ground measured data. Specifically, aerial and satellite images refer to
1965, 1986, and 2000, while ground-based measurements were taken in 2010. The input
data, such as aerial photographs and satellite images, were obtained from the Ethiopian
Mapping Authority and Global Land Cover Facility (glcf.umd.edu/data/landsat), respec-
tively. Additionally, in this case, focused group discussions with selected stakeholders and
semi-structured interviews with key informants, as well as questionnaires, were considered
as a support methodology to assess LULCCs in the area. The report shows five LULC
classes, such as dense acacia woodland, scattered acacia with grass undergrowth, grassland,
cultivated land, and bare lands. Both dense acacia vegetation and scattered acacia were
consistently decreased from 1965 to 2010; grassland cover was increased from 1965 to 2000
years and slightly decreased from 2000 to 2010 years; bare land was increased from 1965
to 1986 and then decreased from 1986 to 2010 years. The cultivated land was increased
from 1986 to 2010 years. From this analysis, it is clear that agriculture is becoming the
predominant LULC class in the basin.
Gebreslassie [16] studied the LULC dynamics of the Huluka watershed, Central Rift
Valley, Ethiopia, from 1973 to 2009. In this research, both supervised and unsupervised clas-
sification was used to classify LULC via the ERDAS Imagine 8.4 software. During the study,
Land 2021, 10, 585 10 of 18

key informant interviews, focal group discussions, and field data collection were combined
to validate the results. The finding shows six LULC classes, namely, cultivated land, wood-
land, open land, grassland, shrubland, continuous NF, fragmented NH, and plantation
forest. Of these LULC classes, only cultivated and open lands had shown continuous and
progressive expansion, mainly at the expense of grass, shrub, and forest lands. In detail,
the 25% and 0% of cultivated and open land of the watershed in 1973 expanded to 84% and
4% in 2009, respectively, while the 29%, 18%, and 22% of grass, shrub, and forest land of
the watershed in 1973 decreased to 3.5%, 4% and 1.5% in 2009, respectively.
To evaluate the trends and drivers of LULC in western Ethiopia, Betru et al. [28]
classified the region into four classes (forest, agriculture, shrub/grass, and settlement),
comparing the years 1978, 1986, 1991, 1999, 2010, 2013, and 2016. They adopted a hybrid
method, combining the outputs of supervised classified and intensive on-screen digitizing
(visual image interpretation) techniques to produce LULC maps. To determine LULCCs,
multi-sensor and multitemporal Landsat images were accessed freely from the USGS
website. To validate the results, key informant interviews and focal group discussion
was conducted to collect historical information of the last decades. The report shows
that 74% of forest was maintained, while the remaining was changed to shrub/grassland
(21%) and agriculture (5%) to 1986 and 1991; nearly 95% of the forest gain in this period
was from shrub/grassland. Between 1978 and 1986, most of the forest was degraded to
shrub/grasslands for the need of construction materials, fuelwood, and charcoal; similar to
the period from 1986 to 1991, 32% of forest was lost from 1991 to 2010, and the conversion
during this period was to shrub/grass and agriculture. Further, 51.7% and 44.7% of the
forest land was converted to agriculture and shrub/grasslands, respectively, due to the
large expansion of commercial agricultural practice between 2010 and 2016. In the same
period, also 25% of the forest land was recovered from shrub/grassland.
In [30], the authors focused on the Fincha watershed. The basin was classified into
agricultural land, forest land, grazing land, waterbody, swamp, and shrubland classes for
the years 1985, 1995, and 2005. The goal of this work was to evaluate LULCC combining
remote sensing and Markov Modelling by using freely available Landsat data. The ERDAS
Imagine software and a number of methodologies such as supervised maximum likelihood
classification, LULC detections, and spatial matrix analysis were adopted to evaluate
LULCCs for the 20-year study period, analysing separately the decades 1985–1995 and
1995–2005. The finding shows that agricultural land and water bodies increased in the area
by around 54% and 93%, respectively, while great losses were observed in the case of forest
land, grazing land, swamp area, and shrubland, by 51%, 31%, 51%, and 25%, respectively.
Looking at the same basin, Dibaba et al. [18] applied a supervised classification with
the maximum likelihood classifier in ArcGIS to classify the land classes in the years 1987,
200, and 2017, combining Landsat images, digital elevation model, and field data. The aim
of this study was to compare the changes between the different years, in addition to
understanding their drivers and implications. The research pointed out that, during the
last 30 years, agricultural land, commercial farm, built-up and waterbody classes increased
by 16%, 5%, 1.7%, and 1.7%, respectively, while forest land, rangeland, grazing land,
and swampy areas decreased by 12%, 8%, 3%, and 1%, respectively. To support the research
outcomes, the authors carried out key informant interviews, focal group discussions,
and field data collection, focusing on socioeconomic aspects.
Likewise, in [31], Ayana et al. classified the LULC of the Fincha watershed for the
year 2005, using Landsat ETM+ images, to model the effects of LULCCs and management
practices on the runoff and sediment yields in the Fincha watershed. The authors combined
supervised classification based on the minimum distance algorithm method, a digital
elevation model, LULC data, soil information, and weather data. Their results show six
classes: agricultural land, forest, grazing land, waterbody, swamp area, and shrubland.
In accordance with evidence pointed out by other authors ([18,30]), more than half of the
watershed was covered by agricultural land and the remaining was covered, in decreasing
order, by waterbody, grazing land, forest, shrubland, and swamp.
Land 2021, 10, 585 11 of 18

Hailu et al. [32] studied the dynamics and drivers of LULCC in the Jimma Geneti
District, western Ethiopia, from 1973 to 2019, using satellite images from the USGS website.
A supervised maximum likelihood classification method was adopted to classify LULC
change within the period, while key informant interviews and focal group discussions
were carried out to validate the results. In total, the authors classified six LULC classes:
bare land, cultivated land, forest, settlement area, waterbody, and wetland. During the
study period, cultivated land, settlement area, and water bodies increased, whereas forest
land, bare land, and wetlands decreased.
In his work, Tolessa et al. [33] studied the impact of LULCCs on ecosystem services
in the central highlands of Ethiopia, from 1973 to 2015, by using multispectral Landsat
imaginary (Landsat MSS, TM, ETM+, and Land OLS). Analysing these satellite images via
the maximum likelihood classification method, they were able to classify the region into
five classes (settlement, cultivated land, bare land, shrubland, and forest). The authors
performed ground control points to assist the supervised image classification. The report
shows the cultivated land and shrubland expanded significantly between 1973 and 2015,
while the forest decreased. No significant changes were observed on both settlement area
and bare land.
Alemu et al. [34] analysed the LULCC implications and drivers in the North-western
Lowlands of Ethiopia during the period 1985–2010 by combining supervised and unsu-
pervised methods of remotely sensed images. Similar to other studies, key informant
interviews, group discussions, and ground control points were performed to validate the
classification results. The study areas were classified into six LULC classes, namely, agri-
cultural land, bare land and settlement, woodland, shrubland, grassland, and water body.
By comparing three reference years (1985, 1995, 2010), the authors pointed out an increment
of the area covered by agriculture, bare land and settlement, grassland, and water bodies,
while woodland and shrubland declined.
In [35], the authors concentrated on the Debre-Mewi watershed, which is the upper
catchment of the Blue Nile Basin, in northwest Ethiopia. In this case, three reference
years were observed (1957, 1982, and 2008), and aerial photographs and multispectral
Landsat satellite images were compared to assess LULCC. To validate the outcomes,
information derived from key informant groups and focal group discussions, as well as
field data, were collected. The report shows that the area was classified into four LULC
classes: natural forest, shrub and bushland, grazing land, and cultivated and settlement
land in 1957. Additional classes were added for the 1982 (Eucalyptus plantation) and 2008
(Eucalyptus plantation, rock outcrop) analyses. In fact, after the 1960s, most individual
farmers started cultivating plantations around their homesteads, as a source of fuelwood,
construction material, and income generation, and some of the severely degraded cultivated
and grazing lands were converted to rock outcrops. For this reason, during the last decades
cultivated and settlement areas increased significantly, whereas natural forest, shrub and
bushland, and grazing land declined rapidly.
According to [36], a combination of unsupervised and supervised classification meth-
ods was employed to classify Landsat images of the Bale rangelands, in southeast Ethiopia,
to study LULCC at the regional scale, and how the local community perceived such
changes. The authors performed key informant interviews and field data collection to
support and validate the results. In this case, the study area was classified into seven LULC
classes: woodland, bushland, shrubby grassland, grassland, cultivated land, bare land,
and settlement, using 1986, 2001, and 2016 as reference years. During these 30 years,
cultivated land, settlement, bushland, and bare land expanded by 14%, 15%, 13%, and 22%,
respectively, while woodland, grassland, and shrubby grassland declined by 34%, 24%,
and 3%, respectively.
Alemayehu et al. [3] studied the trend of LULCCs in the Somodo watershed, south-
western Ethiopia, trying to address the main drivers of such changes. They applied both
unsupervised and maximum likelihood supervised classification methods using ERDAS
imagine 9.1. Landsat images, freely obtained from the USGS website, and key informants,
Land 2021, 10, 585 12 of 18

focal group discussions, and field data techniques were combined to derive coherent infor-
mation. The Somodo watershed was classified into four LULC classes: agriculture, forest,
grass, and home garden agroforestry. The study addressed the changes that occurred be-
tween 1985 and 2017, pointing out that the area covered by forest and agriculture decreased
by 61 ha (13%) and 5 ha (1%), respectively. In contrast, home garden agroforestry/settlement
and grassland increased by around 50 (7%) and 16 (6%), respectively. The authors calculated
that assuming the same existing rate of LULCCs, in 2029, agriculture and forestland are
predicted to increase by 91 ha and 21 ha, respectively, while grassland and home garden
agroforestry/settlement will decrease by 100 ha and 11.79 ha, respectively.
The study in [9] was conducted on LULC dynamics in the Didessa sub-basin, trying to
understand the various driving forces that shaped the landscape in the period 1974–2014,
looking at decadal changes. The analysis was performed via both the unsupervised and
supervised classification methods, using the ERDAS Imagine 2010 classifier within the
ArcGIS software. As made in similar studies, the imagery data used for land cover change
were obtained freely from the USGS website, while key informant interviews and focal
group discussions were carried out for better understanding the observed dynamics and
validate the satellite data. In this case, the study area was classified into seven LULC classes:
wetland, shrubland, settlement, grassland, forest, cultivated land, and commercial land.
The authors’ analysis shows that, during the four decades (1974–2014), agricultural land,
settlement, and commercial land increased, while wetland, grassland, forest, and shrubland
rapidly decreased.
Tefera [12] studied LULC dynamics in the Nonno District, located in central Ethiopia,
for the period 1987–2007. The research combined three satellite datasets: Landsat Thematic
Mapper with 30 m of spatial resolution (image of 1984), Landsat Enhanced Thematic
Mapper Plus with 30 m resolution (image of 2002), and SPOT image of 2007 with 5 m
resolution. Key informant interviews specifically made with older peoples, focal group
discussions, and field data were combined to validate the satellite-derived LULCC analysis.
The Nonno District was classified into woodland, shrubland, grassland, cultivated land,
settlement site, and town. During the observed period, woodland and grassland decreased
their extension, while shrubland, cultivated areas, and settlements expanded. In particular,
woodland and farmland were the two most decreased and increased land use types in the
district, respectively.
Focusing on the Lake Basaka catchment, Dinka and Klik [20] studied how LULCC
affected the regimes of surface runoff during the period 1973–2015. To do that, they adopted
a common methodology, processing Landsat data with the ERDAS Imagine software,
and using both unsupervised and supervised classification methods. The catchment area
was classified into seven classes: farmland, forest (comprising dense woods), shrubland,
grassland, bushy woods (open) land, wetland, and water body. The report indicates that the
Lake Basaka catchment experienced significant LULCCs: about 86% of forest coverage and
46% of grasslands were lost, and the territory was transformed to open bushy woodlands,
farms, lakes, and wetlands.
To analyse the rate of LULCCs during the period 1990–2017 and determine the causes
of changes in forest coverage in the Gog District, Gambella Regional State, Othow et al. [21]
used the maximum likelihood technique of the supervised classification of the ERDAS
Imagine 2014 software. The area was classified into six classes: water body, forest, farmland,
bushland, bare land, and grassland. The authors used free thematic maps (USGS website)
and performed key informant interviews, focal group discussions, and field data collection
to validate the result. Between 1990 and 2017, bare land, forest, and water bodies declined,
whereas farmland, brushland, and grasslands increased.
In their work, Berihun et al. [10] looked at LULCCs and drivers for contrasting the
alterations of the agro-ecological conditions of the Guder, Aba Gerima, and Debatie water-
sheds. For this study, they used images acquired in 1982 and 2017, which were visually
interpreted via the ArcGIS 10.4 software, with support information coming from key infor-
mant interviews and field observations. These three catchments were classified into six
Land 2021, 10, 585 13 of 18

LULC calluses: bare land, cultivated land, forest, grazing land, settlement, and plantations.
In 1982, the forest was the dominant LULC class in Guder and Aba Gerima watershed,
while the Debatie watershed was mostly covered by bushes. During the observed period
forest land, bushland, and grazing land decreased, while cultivated land increased sig-
nificantly in all the studied watersheds. As major drivers of such variations, the authors
identified both the population growth and the associated changes in farming practices.
The study in [13] was on the Andassa watershed, which is located within the Blue Nile
Basin. In this case, a hybrid classification technique was applied, combining unsupervised
and supervised classification methods, via ERDAS Imagine 2014 and ArcGIS 10.3 software.
The watershed was classified into five LULC classes: cultivated land, forest, shrubland,
grassland, and built-up area. This study was intended to analyse first the LULC changes
from 1985 to 2015 and then use a model validated on that period to predict the future state
in 2030 and 2045, through a cellular automata Markov (CA–Markov) model. The result
shows cultivated land and built-up area were of LULC class were increased while forest,
shrubland, and grassland were decreased. Similarly, the increase of cultivated land and
built-up area, and the withdrawal of forest, shrubland, and grassland were forecasted as
continuing in 2030 and 2045.
Andualem and Gebremariam [37] focused on the impact of LULCCs on streamflow
and sediment yield in the Gilgel Abay watershed, Lake Tana sub-basin. In their work,
three reference years were selected (1986, 2000, 2011), and the study area was classified
using five LULC classes: cultivated area, water body, grassland, forest, shrubland, by using
both ERDAS Imagine and ArcGIS software. During the last 25 years, the land covered
by agricultural activities increased significantly, whereas waterbody, grassland, forest,
and shrubland decreased.
Getachew and Melesse [38] used a maximum likelihood supervised method with
ERDAS Imagine and ArcGIS to evaluate the effect of LULCC on the hydrology of the
Angereb Watershed. The area was classified into pasture land, forest, built-up, rangeland,
agriculture, and water body, considering the years 1985 and 2011. To validate the analysis,
the authors used ground truth data acquired during field campaigns. Comparing the 2011
situation with 1985, it is possible to observe that the built-up and agricultural land increased
by 860% and 20.72%, respectively, while pasture land, forest, and rangeland decreased
by around 301%, 29%, and 1.4%, respectively, and water body remained almost constant.
This study showed one of the most significant examples of the increasing human pressure
in Ethiopia, which required the transformation of the region from a natural environment
to an anthropised one, with most of the region covered by agricultural activities and
settlements.
To evaluate LULC dynamics in the Beressa watershed, located in the north central
Highlands, Meshesha et al. [39] used images from 1984 and 2015, analysed through the
ArcGIS10.2.2 and ERDAS Imagine14 software. They classified the basin into six classes
by using a maximum likelihood classification method and validated the outcomes with
information derived from focus group discussions and informal interviews with local
citizens. The study pointed out that, during the studied period, farmland, settlement,
forest land, and water body expanded, whereas grazing land and bare land were reduced.
The situation of the Gelana sub-watershed, northern Highlands, in 1964, 1986, and 2014,
was studied by Miheretu et al. [40]. To classify the study area, a maximum likelihood super-
vised classification method was implemented in ERDAS Imagine9.1, while reference data
points were collected using a GPS for accuracy assessment. The Gelana sub-watershed was
classified into seven LULC classes: forest, shrubland, cultivated and rural settlement land,
grassland, bare land, urban built-up area, and wetland. The result revealed that from 1964
to 2014, shrubland, cultivated and rural settlement, grassland, bare land, and urban built-
up area expanded at a rate of 24%, 7%, 31%, 248%, and 1423%, respectively. On the other
hand, forest and wetland decreased by around 64% and 55%, respectively. Additionally,
in this case, a dramatic increment of urban and built-up areas is recognisable.
Land 2021, 10, 585 14 of 18

The study reported in [41] was conducted in the Central Rift Valley Region, to as-
sess the spatial and temporal LULCC that affected the Arsi Negele District in the pe-
riod 1973–2010. Remote sensing images, analysed via ERDAS Imagine 8.7 and ArcGIS,
were used for LULC classification, and the outcomes were validated using key informant
interviews and field evidence. In this case, the study area was classified into five LULC
classes: bare land, grazing land, cultivated land, shrubland, and Acacia woodland. Dur-
ing the study’s 37-year period, bare land, cultivated land, and shrubland expanded by 18%,
37%, and 47%, respectively, while grazing land and Acacia woodland declined by 53% and
36%, respectively.
Sewnet and Abebe [42] studied the Koga watershed, located in northwest Ethiopia.
Their research looked at understanding the implication of LULCC on the watershed degra-
dation, observed during the period 1973–2011. For doing that, they used four reference
years (1973, 1986, 1995, 2011), and a hybrid unsupervised and supervised classification
approach. The classification results obtained by ERDAS and ArcGIS were integrated
with field evidence measured using GPS systems, and structured household question-
naires, focus group discussions, and interview with key informants. The Koga watershed
was classified into cultivated and settlement areas, forest, brushland, grassland, wetland,
and water body. The authors’ analysis indicated an increment in areas covered by culti-
vated and settlement, forest, and water bodies, while bushland, grassland, and wetland
decreased significantly.
Table 3 summarises the results reported above and allows for a comparison between
the outcomes reported in the different articles.

Table 3. Land use land cover change considered in the reviewed articles.

Article Total Percentage (%)

LULC Class
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
agricultural land + + + + + + + + + + + − + + + + + + + + + + + + 23+, 1− 95.83%+, 4.17%−
forest − − − + − − − − − − − − − − − − − − − − + − − + 3+, 21− 12.5%+, 87.5%−
grazing − − − − + − + − − + − − − + − − − − − + − − 5+, 17− 22.73%+, 77.27%−
63.64%+, 27.27%−,
waterbody + + − + + + − − C + + 7+, 3−, C
9.01%
swamp/wetland − − − − + − = 1+, 6− 14.29%+, 85.71%−
shrub − − − + − − − − + + + − − + + − 6+, 10− 37.5%+, 62.5%−
commercial farm + + 2+ 100%
Built-
+ + + + + + + + + + + + + 13+ 100%
up/settlement
33.33%+,
rangeland − C − 1+, 1−, 1C 33.33*,
33.33%C
45.45%, 36.36%−,
bare/rock outcrop − C − C + + + − − − + + 5+, 4−, 2C
18.18%C

In Table 3, + indicates an increased land cover class, − indicates a decreased land

cover class, and C indicates a constant land cover class, while the free space indicates that
no LULCCs were considered in the specific article.
Following the methodology reported in Section 3, it is possible to observe that,
among Ethiopian basins, there is a clear increment in areas devoted to agriculture, commer-
cial farms, and settlements, as well as water bodies and bare/rock outcrop land (Figure 3).
The increment of this latter class is due to the construction of new human-made reservoirs,
mainly for hydropower and agricultural use. During the last decades, the increasing
human pressure affected negatively the natural environment, as is visible in the significant
decline of forests, swamps, and wetlands.
 Following the methodology reported in Section 3, it is possible to observe that,
among Ethiopian basins, there is a clear increment in areas devoted to agriculture, com-
mercial farms, and settlements, as well as water bodies and bare/rock outcrop land (Figure
3). The increment of this latter class is due to the construction of new human-made reser-
Land 2021, 10, 585 voirs, mainly for hydropower and agricultural use. During the last decades, the increasing 15 of 18
human pressure affected negatively the natural environment, as is visible in the signifi-
cant decline of forests, swamps, and wetlands.

Figure 3. Land use land cover changes percentage in Ethiopia.

Figure 3. Land use land cover changes percentage in Ethiopia.

Agricultural land use increased in all the analysed studies except in [32], while forest
land use decreased in all the study cases, except for the increment pointed out by [16].
In detail, agricultural land, waterbody, commercial farm, bare/rock outcrop and built-
up/settlement increased by around 32%, 17%, 3%, 21% and 191%, respectively. On the
contrary, forest, grazing land, swamp/wetland, shrubland, and rangeland decreased by
around 19%, 18%, 32%, 7%, and 5%, respectively (Figure 3).

5. Discussion
The results of the reviewed articles indicated that LULCCs for the past decades,
as derived from the analysis of satellite imagery, were in accordance with field evidence
(e.g., ground truth data and focus group discussion). In fact, most of the authors used
techniques such as key informant interviews, focal group discussions, and field data
collection to study the socio-economy and to validate the results obtained from Landsat
data. In the majority of the analysed works, the key informant interviews were conducted
involving the elder peoples, aged greater than 60 years old, to derive more consistent
information on the history of the study area. Focal group discussions were performed with
household farmers and local peoples, regardless of their age and social position. During the
field data collection, the authors used GPS information to validate the results.
The outcomes presented in this review article agree with the analysis performed by
Lambin et al. [8], who reviewed different studies covering a wider spatial scale. They ar-
gued that the pathways of LULCCs are a result of globalisation processes, intended as the
[cide with the incorporation of a region into an expanding world economy, as is visible
also for the Ethiopian case study. In this case, the expansion of internal and transnational
markets influenced the LULC, involving deforestation, rangeland modifications, agricul-
tural intensification, and urbanisation, since economic growth and persistent urbanisation
are unavoidable global phenomena that initiate urban encroachment into agricultural
lands [43,44].
Theoretical and numerical modelling can represent a very helpful tool for meeting
land management needs, and for better assessing and projecting the future role of LULCCs
in the functioning of the Earth’s system [45]. To be effective, such models should be able
to reproduce the main drivers of land use change, accounting for their scale dependency,
predicting both the location and the quantity of LULCC, incorporating all possible bio-
Land 2021, 10, 585 16 of 18

physical feedbacks [44–48]. Numerical modelling approaches can benefit from the recent
development in computational resources and the availability of remotely sensed data.
Cloud computing services such as Google Earth Engine can provide information on the
long-term LULCCs over a wide area [49], creating an extensive dataset to calibrate and
validate numerical models.
The reviewed articles addressed LULCCs with high detail, providing significant
evidence at the watershed scale. In fact, all the studied basins experienced a general
trend towards ‘more people more erosion’, with implications in terms of land degradation
and hydrological response. However, there is a lack of (i) detailed investigations of the
implications of LULCCs on land erosion and basin-wide hydrology and (ii) studies focused
on forecasting future trends of LULCCs. Therefore, there is the need to tackle both these
aspects in detail to develop adequate strategies for land management and monitoring
systems needed for assuring a sustainable Ethiopia for the next decades.
In this sense, a few studies tried to simulate the future evolution of LULCCs across
Ethiopian basins, mainly using cellular automata and Markov chain models, which permit
to account for both physical and socioeconomic drivers of LULC dynamics [13,50–52].
Despite the challenges associated with data and model validation, these authors have
shown that such kind of studies is needed to support governmental strategies, both in
rural and urban areas. They pointed out that the increase in built-up areas is an indication
of the rapid population growth, and this may remain a challenge unless environmentally
friendly policies on land use will be implemented to harmonise the demand and diminish
the impacts that arise from it.

6. Conclusions
The review of 25 articles very recent articles on LULCCs in Ethiopia pointed out that
the predominant methods to classify lands are unsupervised and maximum likelihood
supervised classification, generally performed via GIS and ERDAS Imagine software.
To validate the information retrieved from satellite images, the majority of the studies used
key informant interviews, focal group discussions, and ground truth data. Among the
classes analysed, there is a large variability, but the most common ones are agricultural
land, forest, grazing land, water body, swamp area, shrubland, barren land, etc.
A comparison between the articles indicates that, in most of the studied basins, agri-
cultural land, water body, commercial farm, built-up/settlement, and bare/rock outcrop
increased during the last decades in a dramatic manner, while the area covered by forest,
grazing land, and shrubland decreased. Such changes are mostly connected with increas-
ing human pressure on the Ethiopian environment, driven by the need of improving the
socioeconomic situation of the local population.
As pointed out in discussing the single articles, the monitoring on LULCCs can be
performed with a number of techniques and software, eventually driving to dissimilar o
very site-specific results. For the future, therefore, there is a need for agreeing on a common
methodology, aiming to obtain consistent results worldwide.

Author Contributions: Conceptualisation, M.S.R. and M.N.; writing—original draft preparation,

M.S.R. and M.N.; literature review, M.S.R. and D.A.; manuscript revision, M.S.R., M.N. and D.A.;
supervision, M.N.; project administration, M.N.; funding acquisition, M.N. All authors have read
and agreed to the published version of the manuscript.
Funding: This research was funded by NCN National Science Centre Poland–call PRELUDIUM
BIS-1, Grant Number 2019/35/O/ST10/00167. Project website: https://sites.google.com/view/lulc-
fincha/home.
Data Availability Statement: The data presented in this study are available on request from the
corresponding author.
Conflicts of Interest: The authors declare no conflict of interest. The funder had no role in the design
of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; in
the decision to publish the results.
Land 2021, 10, 585 17 of 18

References
1. IPCC. Land Use, Land-Use Change, and Forestry; Watson, R.T., Noble, I.R., Bolin, B., Ravindranath, N.H., Verardo, D.J., Dokken, D.J.,
Eds.; Cambridge University Press: Cambridge, UK, 2000; p. 375.
2. Mariye, M.; Mariyo, M.; Changming, Y.; Teffera, Z.L.; Weldegebrial, B. Effects of land use and land cover change on soil erosion
potential in Berhe district: A case study of Legedadi watershed, Ethiopia. Int. J. River Basin Manag. 2020, 1–13. [CrossRef]
3. Maitima, J.M.; Mugatha, S.M.; Reid, R.S.; Gachimbi, L.N.; Majule, A.; Lyaruu, H.; Mugisha, S. The linkages between land use
change, land degradation and biodiversity across East Africa. Afr. J. Environ. Sci. Technol. 2009, 3, 310–325.
4. Haregeweyn, N.; Tesfaye, S.; Tsunekawa, A.; Tsubo, M.; Meshesha, D.T.; Adgo, E.; Elias, A. Dynamics of land use and land cover
and its effects on hydrologic responses: Case study of the Gilgel Tekeze catchment in the highlands of Northern Ethiopia. Environ.
Monit. Assess. 2014, 187, 1–14. [CrossRef]
5. Goldewijk, K.K.; Ramankutty, N. Land cover change over the last three centuries due to human activities: The availability of new
global data sets. GeoJournal 2004, 61, 335–344. [CrossRef]
6. Chang, Y.; Hou, K.; Li, X.; Zhang, Y.; Chen, P. Review of Land Use and Land Cover Change research progress. IOP Conf. Ser.
Earth Environ. Sci. 2018, 113, 012087. [CrossRef]
7. Alemayehu, F.; Tolera, M.; Tesfaye, G. Land use land cover change trend and its drivers in Somodo watershed south western,
Ethiopia. Afr. J. Agric. Res. 2019, 14, 102–117.
8. Lambin, E.F.; Turner, B.L.; Geist, H.J.; Agbola, S.B.; Angelsen, A.; Bruce, J.W.; Coomes, O.T.; Dirzo, R.; Fischer, G.; Folke, C.; et al.
The causes of land-use and land-cover change: Moving beyond the myths. Glob. Environ. Chang. 2001, 11, 261–269. [CrossRef]
9. Tolessa, T.; Dechassa, C.; Simane, B.; Alamerew, B.; Kidane, M. Land use/land cover dynamics in response to various driving
forces in Didessa sub-basin, Ethiopia. GeoJournal 2020, 85, 747–760. [CrossRef]
10. Berihun, M.L.; Tsunekawa, A.; Haregeweyn, N.; Meshesha, D.T.; Adgo, E.; Tsubo, M.; Masunaga, T.; Fenta, A.A.; Sultan, D.;
Yibeltal, M. Exploring land use/land cover changes, drivers and their implications in contrasting agro-ecological environments of
Ethiopia. Land Use Policy 2019, 87, 104052. [CrossRef]
11. Guzha, A.; Rufino, M.; Okoth, S.; Jacobs, S.; Nobrega, R. Impacts of land use and land cover change on surface runoff, discharge
and low flows: Evidence from East Africa. J. Hydrol. Reg. Stud. 2018, 15, 49–67. [CrossRef]
12. Marchant, R.; Richer, S.; Boles, O.; Capitani, C.; Courtney-Mustaphi, C.J.; Lane, P.; Prendergast, M.E.; Stump, D.; De Cort, G.;
Kaplan, J.O.; et al. Drivers and trajectories of land cover change in East Africa: Human and environmental interactions from 6000
years ago to present. Earth Sci. Rev. 2018, 178, 322–378. [CrossRef]
13. Gashaw, T.; Tulu, T.; Argaw, M.; Worqlul, A.W. Evaluation and prediction of land use/land cover changes in the Andassa
watershed, Blue Nile Basin, Ethiopia. Environ. Syst. Res. 2017, 6, 17. [CrossRef]
14. Tefera, M.M. Land-use/land-cover dynamics in Nonno district, central Ethiopia. J. Sustain. Dev. Afr. 2011, 13, 123–141.
15. Demeke, M.; Guta, F.; Ferede, T. Agricultural Development in Ethiopia: Are There Alternatives to Food Aid? Department of
Economics, Addis Ababa University 2004. Available online: sarpn.org.za/documents/d0001583/FAO2005_Casestudies_Ethiopia.
pdf (accessed on 15 May 2021).
16. Gebreslassie, H. Land Use-Land Cover dynamics of Huluka watershed, Central Rift Valley, Ethiopia. Int. Soil Water Conserv. Res.
2014, 2, 25–33. [CrossRef]
17. Dadi, D.; Azadi, H.; Senbeta, F.; Abebe, K.; Taheri, F.; Stellmacher, T. Urban sprawl and its impacts on land use change in Central
Ethiopia. Urban For. Urban Green. 2016, 16, 132–141. [CrossRef]
18. Dibaba, W.T.; Demissie, T.A.; Miegel, K. Drivers and Implications of Land Use/Land Cover Dynamics in Finchaa Catchment,
Northwestern Ethiopia. Land 2020, 9, 113. [CrossRef]
19. Romanowicz, R.J. The impacts of changes in climate and land use on hydrological processes. Acta Geophys. 2017, 65, 785–787.
[CrossRef]
20. Dinka, M.O.; Klik, A. Effect of land use–land cover change on the regimes of surface runoff—the case of Lake Basaka catchment
(Ethiopia). Environ. Monit. Assess. 2019, 191, 278. [CrossRef]
21. Othow, O.O.; Gebre, S.L.; Gemeda, D.O. Analyzing the Rate of Land Use and Land Cover Change and Determining the Causes of
Forest Cover Change in Gog District, Gambella Regional State, Ethiopia. J. Remote Sens. GIS 2017, 6, 1–13. [CrossRef]
22. Grover, D.K.; Temesgen, A. Enhancing Land-Use-Efficiency through Appropriate Land Policies in Ethiopia. 2006. Available
online: https://ideas.repec.org/p/ags/iaae06/25615.html (accessed on 16 May 2021).
23. Bekele, T.; Tsegaye, B. Effect of Land Use and Land Cover Changes on Soil Erosion in Ethiopia. Int. J. Agric. Sci. Food Technol.
2019, 5, 026–034. [CrossRef]
24. Reusing, M.; Schneider, T.; Ammer, U. Modelling soil loss rates in the Ethiopian Highlands by integration of high resolution
MOMS-02/D2-stereo-data in a GIS. Int. J. Remote Sens. 2000, 21, 1885–1896. [CrossRef]
25. Ebabu, K.; Tsunekawa, A.; Haregeweyn, N.; Adgo, E.; Meshesha, D.T.; Aklog, D.; Masunaga, T.; Tsubo, M.; Sultan, D.; Fenta,
A.A.; et al. Effects of land use and sustainable land management practices on runoff and soil loss in the Upper Blue Nile basin,
Ethiopia. Sci. Total Environ. 2019, 648, 1462–1475. [CrossRef] [PubMed]
26. Plant Genetic Resources Center. In Proceedings of the Country Report to the FAO International Technical Conference on Plant
Genetic Resources, Rome, Italy, 19–30 June 1995.
27. Suryabhagavan, K.V. GIS-based climate variability and drought characterization in Ethiopia over three decades. Weather Clim.
Extrem. 2017, 15, 11–23. [CrossRef]
Land 2021, 10, 585 18 of 18

28. Betru, T.; Tolera, M.; Sahle, K.; Kassa, H. Trends and drivers of land use/land cover change in Western Ethiopia. Appl. Geogr.
2019, 104, 83–93. [CrossRef]
29. Biazin, B.; Sterk, G. Drought vulnerability drives land-use and land cover changes in the Rift Valley dry lands of Ethiopia. Agric.
Ecosyst. Environ. 2013, 164, 100–113. [CrossRef]
30. Ayana, A.B.; Kositsakulchai, E. Land use change analysis using remote sensing and Markov Modeling in Fincha watershed,
Ethiopia. Agric. Nat. Resour. 2012, 46, 135–149.
31. Ayana, A.B.; Edossa, D.; Kositsakulchai, E. Modeling the effects of land use change and management practices on runoff and
sediment yields in Fincha watershed, Blue Nile. OIDA Int. J. Sustain. Dev. 2014, 7, 75–88.
32. Hailu, A.; Mammo, S.; Kidane, M. Dynamics of land use, land cover change trend and its drivers in Jimma Geneti District,
Western Ethiopia. Land Use Policy 2020, 99, 105011. [CrossRef]
33. Tolessa, T.; Senbeta, F.; Kidane, M. The impact of land use/land cover change on ecosystem services in the central highlands of
Ethiopia. Ecosyst. Serv. 2017, 23, 47–54. [CrossRef]
34. Alemu, B.; Garedew, E.; Eshetu, Z.; Kassa, H. Land use and land cover changes and associated driving forces in north western
lowlands of Ethiopia. Int. Res. J. Agric. Sci. Soil Sci. 2015, 5, 28–44.
35. Fisseha, G.; Gebrekidan, H.; Kibret, K.; Yitaferu, B.; Bedadi, B. Analysis of land use/land cover changes in the Debre-Mewi
watershed at the upper catchment of the Blue Nile Basin, North West Ethiopia. J. Biodivers. Environ. Sci. 2011, 1, 184–198.
36. Mussa, M.; Teka, H.; Mesfin, Y. Land use/cover change analysis and local community perception towards land cover change in
the lowland of Bale rangelands, Southeast Ethiopia. Int. J. Biodivers. Conserv. 2017, 9, 363–372.
37. Andualem, T.G.; Gebremariam, B. Impact of land use land cover change on stream flow and sediment yield: A case study of
Gilgel Abay watershed, Lake Tana sub-basin, Ethiopia. Int. J. Technol. Enhanc. Merg. Eng. Res. 2015, 3, 28–42.
38. Getachew, H.E.; Melesse, A.M. The impact of land use change on the hydrology of the Angereb Watershed, Ethiopia. Int. J.
Water Sci. 2012, 1, 6.
39. Meshesha, T.W.; Tripathi, S.K.; Khare, D. Analyses of land use and land cover change dynamics using GIS and remote sensing
during 1984 and 2015 in the Beressa Watershed Northern Central Highland of Ethiopia. Model Earth Syst. Environ. 2016, 2, 1–12.
[CrossRef]
40. Miheretu, B.A.; Yimer, A.A. Land use/land cover changes and their environmental implications in the Gelana sub-watershed of
Northern highlands of Ethiopia. Environ. Syst. Res. 2017, 6, 7. [CrossRef]
41. Molla, M.B. Land use/land cover dynamics in the central rift valley region of Ethiopia: Case of Arsi Negele District. Afr. J.
Agric. Res. 2015, 10, 434–449.
42. Sewnet, A.; Abebe, G. Land use and land cover change and implication to watershed degradation by using GIS and remote
sensing in the Koga watershed, North Western Ethiopia. Earth Sci. Inform. 2017, 11, 99–108. [CrossRef]
43. Azadi, H.; Ho, P.; Hasfiati, L. Agricultural land conversion drivers: A comparison between less developed, developing and
developed countries. Land Degrad. Dev. 2010, 22, 596–604. [CrossRef]
44. Paül, V.; McKenzie, F.H. Peri-urban farmland conservation and development of alternative food networks: Insights from a
case-study area in metropolitan Barcelona (Catalonia, Spain). Land Use Policy 2013, 30, 94–105. [CrossRef]
45. Agarwal, C.; Green, G.M.; Grove, J.M.; Evans, T.P.; Schweik, G.M. A Review and Assessment of Land-Use Change Models.
Dynamics of Space, Time and Human Choice. Bloomington and South Burlington, Center for the Study of Institutions, Population
and Environmental Change, Indiana University and USDA Forest Service. CIPEC Collab. Rep. Ser. 2001, 1. [CrossRef]
46. Mondal, M.S.; Sharma, N.; Kappas, M.; Garg, P.K. Modeling of spatio-temporal dynamics of LULC-A review and assessment.
J. Geomat. 2012, 6, 93–103.
47. Veldkamp, A.; Lambin, E. Predicting land-use change. Agric. Ecosyst. Environ. 2001, 85, 1–6. [CrossRef]
48. Nones, M. On the main components of landscape evolution modelling of river systems. Acta Geophys. 2020, 68, 459–475.
[CrossRef]
49. Midekisa, A.; Holl, F.; Savory, D.J.; Andrade-Pacheco, R.; Gething, P.W.; Bennett, A.; Sturrock, H. Mapping land cover change over
continental Africa using Landsat and Google Earth Engine cloud computing. PLoS ONE 2017, 12, e0184926. [CrossRef] [PubMed]
50. Gidey, E.; Dikinya, O.; Sebego, R.; Segosebe, E.; Zenebe, A. Cellular automata and Markov Chain (CA_Markov) model-based
predictions of future land use and land cover scenarios (2015–2033) in Raya, northern Ethiopia. Model Earth Syst. Environ. 2017, 3,
1245–1262. [CrossRef]
51. Hishe, S.; Bewket, W.; Nyssen, J.; Lyimo, J. Analysing past land use land cover change and CA-Markov-based future modelling in
the Middle Suluh Valley, Northern Ethiopia. Geocarto Int. 2020, 35, 225–255. [CrossRef]
52. Mohamed, A.; Worku, H. Simulating urban land use and cover dynamics using cellular automata and Markov chain approach in
Addis Ababa and the surrounding. Urban Clim. 2020, 31, 100545. [CrossRef]