


































115 

GeoPlanning 
Journal of Geomatics and Planning                                                                                                                   Vol. 8, No. 2, 2021     

 

Original Research 

Prospective Mapping of Land Cover and Land 

Use in The Classified Forest of The Upper 

Alibori Based on Satellite Imagery 

Dramane Issiako 1,2*, Ousséni Arouna 1,2, Karimou Soufiyanou 1, Ismaila 

T. Imorou 1, Brice Tente 3 

1. Laboratory of Cartography (LaCarto), University of Abomey-Calavi (UAC), Cotonou, Benin 

2. Geosciences, Environment and Applications Laboratory (LaGEA), National University of 

Sciences, Technologies, Engineering and Mathematics (UNSTIM), Abomey, Benin 

3. Laboratory of Biogeography and Environmental Expertise (LaBEE), University of Abomey-

Calavi, Benin 

DOI: 10.14710/geoplanning.8.2.115-126  

Abstract 

The dynamics of land cover and land use in the classified forest of the upper Alibori (FCAS) in relation to the disturbance 

of agro-pastoral activities is a major issue in the rational management of forest resources. The objective of this research is 

to simulate the evolutionary trend of land cover and land use in the FCAS by 2069 based on satellite images. Landsat images 

from 2009, 2014 and 2019 obtained from the earthexplorer-usgs archive were used. The methods used are diachronic 

mapping and spatial forecasting based on senarii. The MOLUSCE module available under QGIS remote sensing 2.18.2 is 

used to simulate the future evolution of land cover and land use in the FCAS. The land cover and use in the year 2069 is 

simulated using cellular automata based on the scenarios. The results show that natural land cover units have decreased 

while anthropogenic formations have increased between 2009 and 2014 and between 2014 and 2019. Under the "absence 

multi-criteria zoning (MZM)" scenario over a 50-year interval, land cover and use will be dominated by crop-fallow mosaics 

(88%). On the other hand, the scenario "implementation of a multicriteria zoning (MZE)", was issued with the aim of 

reversing the regressive trend of vegetation types by making a rational and sustainable management of resources. 

Copyright © 2021 GJGP-Undip 

This open access article is distributed under a  

Creative Commons Attribution (CC-BY-NC-SA) 4.0 International license 

1. Introduction 

Africa had the highest net loss of forest area over the period 2010-2020, at 3.94 million hectares per year 

(FAO et PNUE, 2020). Forests managed with the support of forestry projects are not spared (Gbedahi et al., 

2019). Sustainable forest management requires the ability to characterise and spatialise the resource within a 

forest massif (Munoz et al., 2015). Land cover and its evolution over time is a good indicator of these interactions, 

as it reflects the impacts of land cover and climate change on natural environments (Monier, 2010). Furthermore, 

modelling and projecting land cover changes is becoming a relevant tool for decision support (Thierry et al., 

2018). It allows territorial planning policies to be analysed in order to assess and anticipate their environmental 

impacts (Samie et al., 2017). Exploring the future in a quantitative way is a scientific challenge. Taking into 

account the spatial dimension (in the quantitative sense) in foresight is relatively recent and remains delicate, 

calling on various skills in geomatics and remote sensing for the reconstruction of past trajectories, but also in 

modelling (Houet, 2015) 

 

e-ISSN: 2355-6544 
 
Received: 10 August 2021;  
Accepted: 10 December 2021;  
Published: 30 December 2021. 
 
Keywords:  
Land use, land cover, spatial 
prospective, trend, Classified 
Forest, Upper Alibori. 
 
*Corresponding author(s) email: 
dramaneissiako@gmail.com 

 
 

 

https://doi.org/10.14710/geoplanning.8.2.115-126
mailto:dramaneissiako@gmail.com


Issiako et al. / Geoplanning: Journal of Geomatics and Planning, Vol 8, No 2, 2021, 115-126 
DOI: 10.14710/geoplanning.8.2.115-126 

 

116 

In Benin, forests are undergoing deforestation or degradation processes of varying severity, with negative 

impacts on ecosystems and the livelihoods of local populations in particular (Moumouni et al., 2019). The human 

activities directly responsible for forest destruction are commercial timber exploitation, the establishment of 

crops and plantations, the use of firewood and intensive livestock farming (overgrazing) (Sinsin et al., 2010). 

Nevertheless, deforestation, which was estimated at 150,000 ha/year between 1960 and 1980, fell from 70,000 

ha/year between 1990 and 2000 to 50,000 ha/year from 2000. This regression is a testimony to the efforts of 

the Beninese state to curb the degradation of vegetation cover (FAO, 2015).  

In relation to the national forest cover, the north of Benin concentrates nearly 92.5% of natural resources 

thanks to the presence of a series of protected areas (classified forests, hunting zones and national parks) (Mama 

et al., 2020). But with high population growth correlated with unsustainable land cover and land use, these 

reforestation efforts are being challenged. Furthermore, the increasing lack of fertile land, the inadequacy of 

grazing areas and the search for water in the lands bordering the FCAS are invoked to justify the unsustainable 

exploitation of the natural resources available in this area by the indigenous population (Mama et al., 2020). 

Agropastoral exploitation observed in the FCAS in 2000 by (Akindélé, 2000), in 2002 has continued and been 

reinforced (Issiako & Arouna, 2018; Mama et al., 2020; Seidou et al., 2017). Deforestation and forest degradation 

continue in the FACS despite the management plan developed with the participation of various stakeholders. 

The practice of agro-pastoral activities in the FCAS and forest dynamics are intimately linked, leading 

objectively to a restructuring of forest areas. The understanding and monitoring of land cover dynamics as well 

as the representation of changes affecting the territory are thus political, economic and social issues in the FCAS. 

What is the evolutionary trend of land cover in the FCAS in the current context of intense agropastoral 

practices? The objective of this paper is to simulate the evolutionary trend of land cover and land use in the 

FCAS by 2069 on the basis of scenarios. This research is based on the hypothesis that the evolutionary trends of 

the vegetation cover by 2069 will vary according to the scenarios put in place. 

 

2. Material and Methods 

2.1. Study Area 

The FCAS straddles the Provinces of Atacora, Borgou and Alibori. It is located between 10°14 and 11°40 

North latitude and between 1°54' and 2°55' East longitude. It is located in ecological zones 1 and 2 which include 

the Districts of Gogounou, Kandi, Banikoara, Kèrou, Ouassa-Péhunco and Sinendé, which are known to be major 

producers of cotton, maize and yams. This state forest estate is globally subject to degradation factors including 

agriculture, hunting, livestock, logging and various forms of encroachment related to the installation of housing 

and other infrastructure (Issiako & Arouna, 2018). Figure 1 shows the geographical location of the FCAS. 

 

2.2. Rationale for the choice of dates  

The participatory management plan for the FCAS was developed for the period 2010 - 2019. Thus, the 

use of spatial imagery before the plan (2009), during the implementation of the plan (2014) and after the 

implementation of the plan (2019).  

 

https://doi.org/10.14710/geoplanning.8.2.115-126


Issiako et al. / Geoplanning: Journal of Geomatics and Planning, Vol 8, No 2, 2021, 115-126 
DOI: 10.14710/geoplanning.8.2.115-126 

 

117 

 

  Data source: Benin general map, National Geographic Institute (IGN) of Benin, 2018 

Figure 1. Geographical location of the Upper Alibori Classified Forest 

 

2.3. Methodological flow chart  

This research used methodology that can be seen at Figure 2. 

 

Legend: AZM: Absence of multi-criteria zoning; MZE: Implementation of zoning with effectiveness 

Source: inspired by (Hakim et al., 2019) 

Figure 2. Flow chart of the research method 

https://doi.org/10.14710/geoplanning.8.2.115-126


Issiako et al. / Geoplanning: Journal of Geomatics and Planning, Vol 8, No 2, 2021, 115-126 
DOI: 10.14710/geoplanning.8.2.115-126 

 

118 

2.4. Planimetric data used 

The planimetric data used are topographic maps, at 1:50,000 scale, sheets of Bagou, Alibori Forest, 

Goumori, Kérou, Péhunco, Sinendé and Sonsoro produced by the Institut Géographique National (IGN) of Benin 

in 2018; Landsat 7 ETM+ multi-spectral images in Geotiff format from 22 december 2009: Path192 and Row 

52; with a spatial resolution of 30 m; Landsat OLI-TIRS (Landsat 8) images in Geotiff format from 25 december 

2014: Path192 and Row 52; with a spatial resolution of 30 m; and Landsat OLI-TIRS (Landsat 8) images in 

Geotiff format, of 20 december 2019; Path192 and Row 52;with a spatial resolution of 30 m. These images have 

been downloaded from www.earthexplorer-usgs.gov/usa. These images were radiometrically corrected before 

digital processing. 

2.5. Digital processing of Landsat images 

The mapping of spatio-temporal land cover and land use changes started with the digital processing of 

satellite images using QGIS2.18.2 software, in particular the Train Radom Forest Image Classifier module 

contained in the Orfeo toolkit. The "RandomForest" algorithm has already been used in previous studies on 

satellite image classification (Rodriguez-Galiano et al., 2012; Shao et al., 2016). This digital processing includes: 

importing Landsat images into QGIS software, clipping the area of interest, calculating the image pyramid, 

colour composition, choosing training areas and supervised classification by maximum likelihood. 

Importing the images into QGIS. The different image scenes were imported into the QGIS software. 

Mosaicing 

A mosaic of two (02) scenes was made to cover the entire FCAS (Figure 3). 

 

 

 

Data source: Landsat 7 ETM+, December 2009, P.192 and R.52, 30 m 

Figure 3. Mosaic of two (02) scenes covering the FCAS 

 

 

a. Image from 2009 before mosaic b. Image from 2009 after mosaic 

https://doi.org/10.14710/geoplanning.8.2.115-126


Issiako et al. / Geoplanning: Journal of Geomatics and Planning, Vol 8, No 2, 2021, 115-126 
DOI: 10.14710/geoplanning.8.2.115-126 

 

119 

Correction of the 2009 image 

The 2009 image has been corrected (Fill Gab); the strips have been filled in to allow further work. Figure 

4 shows the 2009 image in the study area before and after treatment. 

   

 

Data source: Landsat 7 ETM+, December 2009, P.192 and R.52, 30 m 

Figure 4. Landsat 7 ETM+ image from 2009 before and after correction 

 

Creation of ROIs (Region Of interest) 

After a colour composite, the land-use units were identified and coded on the different scenes. For each 

land-use unit, training areas (ROIs) were delineated away from the transition zones to avoid including mixed 

pixels that could be classified in two distinct classes. 

Creation of the classification model 

In order to classify under Random Forest, a model was created to run the classification using the ROIs 

created previously. The Train Radom Forest Image Classifier module of the Orfeo toolbox was used to create 

the model. Once the model was validated through the value "Global performance", for each image (Olouloi et al., 

2006; Toko Mouhamadou, 2014), the classification was done using the Image classifier module contained in the 

Orfeo toolbox. 

Create Image Classification 

This application performs a classification of the input image, based on the Model file created with the 

Train Random Image Image Classifier algorithm. The supervised classification was then performed. The 

training plots were used to establish a key numerical feature that could best describe the spectral attributes for 

each class type. In this case, the parametric algorithm chosen is maximum likelihood (Toko Mouhamadou, 2014). 

In supervised classification, the image analyst supervises the pixel categorisation process by specifying to 

the computer algorithm numerical descriptors of various land cover types present in the scene. Thus, 

representative samples of known land cover sites (training plots) were used.  

a. 2009 image before 

correction 

 

b. 2009 image after 

correction 

https://doi.org/10.14710/geoplanning.8.2.115-126


Issiako et al. / Geoplanning: Journal of Geomatics and Planning, Vol 8, No 2, 2021, 115-126 
DOI: 10.14710/geoplanning.8.2.115-126 

 

120 

Classification evaluation 

The control points were geolocalised using a Garmin 62s GPS receiver at all homogeneous units across 

the FCAS. Two validation visits were carried out to confirm and reclassify the results of the satellite image 

interpretation. Of two hundred (195) sampled control points, 190 were found to be correctly classified, i.e. a 

proportion of 97 %. Table I shows the distribution of the ground control points by land cover unit. 

 

Table I. Distribution of fieldwork points by land cover/landuse unit 

Units of land cover  Traded points Validated points Percentage (%) 

Field mosaic and fallow 45 44 23 
Woodlands 14 13 7 
Gallery forest / riparian formation 40 38 19 
Plantation 1 1 1 

Tree and shrub savannahs 95 94 48 

 

Vectorisation and development of the transition matrix 

The classified images were transformed into a shapefile in order to determine the areas of each land cover 

unit and to establish the transition matrix. The transition matrix is in the form of a square matrix and consists 

of X rows and Y columns. The number of rows in the matrix indicates the number of land-use units at time t 0; 

the number Y of columns in the matrix is the number of converted units at time t 1 and the diagonal contains 

the areas of the units that remain unchanged. The transformations are done from rows to columns. 

 

2.6. Detecting changes  

Average annual rates of spatial expansion (T) 

The annual average rate of spatial expansion expresses the proportion of each land cover unit that changes 

annually (Zakari et al., 2018).  

T = 
(𝑙𝑛𝑆2−𝑙𝑛𝑆1)

(𝑡2−𝑡1)
 x 100 

with T: average annual rate of spatial expansion; S1 and S2: area of a land-use unit at dates t1 and t2 

respectively; t2 - t1: number of years of evolution; ln: natural logarithm; e: base of natural logarithm; (e = 2.71828 

invariant coefficient). 

Conversion rate of land cover units  

The conversion rate of a land cover class is the degree to which the land cover class has changed by 

converting to other classes. 

Tc = 
𝑆𝑖𝑡−𝑆𝑖𝑠

𝑆𝑖𝑡
 x 100 

with Tc: conversion rate; Sit : Area of unit i at initial date t; Sis: Area of the same unit remaining stable at 

date t1. 

 

 

 

[Eq :1] 

[Eq :2] 

https://doi.org/10.14710/geoplanning.8.2.115-126


Issiako et al. / Geoplanning: Journal of Geomatics and Planning, Vol 8, No 2, 2021, 115-126 
DOI: 10.14710/geoplanning.8.2.115-126 

 

121 

2.7. Forward-looking land cover mapping method  

The 2014 and 2019 land cover maps were used to simulate the 2069 land cover. To estimate the reliability 

and predictive capacity of the simulation to 2069, the 2019 land cover was simulated from the transition of the 

land cover dynamics observed between 2009 and 2014 as a test. The reference map of 2019 and the one simulated 

in the same year were compared. Therefore, if the validated result reaches an acceptable accuracy (50%), then 

the simulation for 2069 will be valid. However, if the result is less accurate, the simulation will not be valid. The 

simulated map in 2019 was produced using QGIS remote sensing 2.18.2 software including the MOLUSCE 

(Model for Land cover Change Evaluation) module (Mienmany, 2018). 

Several factors influencing land cover change were incorporated into this model. These are distance to 

settlements, distance to fields, distance to roads and population density. Cramer's V index was calculated for each 

explanatory factor and used to select those that best contribute to land cover dynamics. This is the Cramer's V 

coefficient, which is a correlation coefficient that varies from 0.0 (no correlation) to 1.0 (perfect correlation). 

Two scenarios were developed to project the future of land cover in order to facilitate decision-making. 

These are: Scenario 1 Absence of multi-criteria zoning (AZM), trend (2014-2019) maintained, the AZM scenario 

is a trend scenario that assumes no new forest management policies; and Scenario 2 Implementation of zoning 

with effectiveness Sustainable Management 2069, the main objective of the "Sustainable Management" scenario 

is to manage the remaining forest resource through the implementation of an integrated forest management 

plan.  

 

3. Results 

3.1. Land cover dynamics of the FCAS  

Land cover in 2009, 2014, 2019 in the FCAS can be seen in Figure 5. 

 

Data source : Landsat 7 ETM+, Landsat 8 OLI-TIRS, P.192 and R.52; resolution : 30 m ; method supervised classicification : Train 

Radom Forest of the Orfeo 

Figure 5. Land cover in 2009, 2014 and 2019 in the FCAS 

https://doi.org/10.14710/geoplanning.8.2.115-126


Issiako et al. / Geoplanning: Journal of Geomatics and Planning, Vol 8, No 2, 2021, 115-126 
DOI: 10.14710/geoplanning.8.2.115-126 

 

122 

These three units have respectively -28%, -9% and -7% spatial expansion rates between 2009 and 2014. 

Between 2014 and 2019, units such as dense dry forest (98%), tree and shrub savannahs (64%), gallery forest 

(49%) and open forest and wooded savannahs (38%) experienced a high conversion rate with a negative average 

annual spatial expansion rate.  

3.2. Probability of change of land-use units in scenario 1 

The transition probabilities provide information on the likelihood of conversion of units to other land-

use units between 2019 and 2069 (scenario 1). 

Scenario 1 transition probability matrix (AZM) 

Table II. Transition probability matrix of land cover and land use units 

Class name 2019 

Class name 2069 

TOTAL 
Gallery 

forest / 

riparian 

formation 

Woodlands 

Tree and 

shrub 

savannahs 

Field 

mosaic and 

fallow 

Plantation 

Rocky / 

uncovered 

areas 

Water body Habitation 

Gallery forest / 

riparian formation 
0.7209 0 0 0.2061 0.0691 0 0 0 1 

Woodlands 0 0.2646 0.1503 0.5766 0 0 0 0.01 1 

Tree and shrub 

savannahs 
0 0 0.12 0.8796 0 0 0 0 1 

Field mosaic and 

fallow 
0 0 0 0.9981 0 0 0 0 1 

Plantation 0 0 0 0 0.8185 0 0 0.18 1 

Rocky / uncovered 

areas 
0 0 0 0 0 0.0111 0 0.99 1 

Water body 0 0 0 0 0 0 1 0 1 

Habitation 0 0 0 0 0 0 0 1 1 

Data source : Landsat 8 OLI-TIRS, LCLU, 2019, cell transmission rules, Simulation sotfware MOLUSCE (QGIS Remote Sensing) 

Based on Table II, it can be deduced that by 2069, open forests and wooded savannahs and tree and shrub 

savannahs will no longer exist and the other vegetation formations if the evolutionary trends observed between 

2014 and 2019 are maintained. 

Scenario 2 transition probability matrix (MZE) 

Table III. Transition probability matrix from anthropogenic to natural formations 

Class name 2019 

Class name 2069 

Total 
Gallery 

forest / 

riparian 

formation 

Dense 

forest 
Woodlands 

Tree and 

shrub 

savannahs 

Field 

mosaic and 

fallow 

Plantation 

Rocky / 

uncovered 

areas 

Water 

body 
Habitation 

Gallery forest / 

riparian formation 
0.9959 0 0 0 0 0.0041 0 0 0 1 

Dense forest 0 1 0 0 0 0 0 0 0 1 

Woodlands 0 0.0402 0.9551 0 0 0.0048 0 0 0 1 

Tree and shrub 

savannahs 
0 0.1326 0.3106 0.5531 0.0014 0.0022 0 0 0 1 

Field mosaic and 

fallow 
0.0267 0.0227 0.0791 0.8671 0.0037 0.0008 0 0 0 1 

Plantation 0.0073 0.0045 0.0139 0.106 0.0071 0.8607 0 0 0.0005 1 

Rocky / uncovered 

areas 
0 0 0.3939 0.6061 0 0 0 0 0 1 

Water body 0.8718 0 0 0 0 0 0 0.1282 0 1 

Habitation 0 0.001 0.0036 0.9894 0 0 0 0 0.0059 1 

Data source: Landsat 8 OLI-TIRS, LCLU, 2019, cell transmission rules, Simulation sotfware MOLUSCE (QGIS Remote Sensing) 

https://doi.org/10.14710/geoplanning.8.2.115-126


Issiako et al. / Geoplanning: Journal of Geomatics and Planning, Vol 8, No 2, 2021, 115-126 
DOI: 10.14710/geoplanning.8.2.115-126 

 

123 

Based on Table III with the implementation of multi-criteria zoning, there will be more anarchic 

installation of fields; the probability of reconstitution and stability of natural vegetation by 2069 will be high. 

3.3. Prospective states of land cover units by 2069  

Land use by 2069 based on the "Absence Multi-Criteria Zoning (AZM)" and "Implementation Multi-Criteria 

Zoning" scenarios can be seen in Figure 6. 

 

Data source : Landsat 8 OLI-TIRS, LCLU, 2019, cell transmission rules, Simulation sotfware MOLUSCE (QGIS Remote Sensing) 

Figure 6. Land use by 2069 based on the "Absence Multi-Criteria Zoning (AZM)" and "Implementation 

Multi-Criteria Zoning" scenarios 

 

Changes observed with"Absence Multi-Criteria Zoning (AZM)" scenario 

 

Data source: Landsat 8 OLI-TIRS, LCLU, 2019, cell transmission rules, Simulation sotfware MOLUSCE (QGIS Remote Sensing) 

Figure 7. Changes observed between 2019 and 2069 in the Absence Multi-criteria Zoning scenario 

Indeed, in this scenario, gallery forests, dense dry forests, woodlands, tree and shrub savannahs and water 

bodies will lose 4%, 31% and 63% respectively by 2069 compared to the base year 2019 (Figure 7). The area of 

fields and fallow land will increase by 88%.  

-100 -50 0 50 100

Gallery forest / riparian formation

Dense forest

Woodlands

Tree and shrub savannahs

Field mosaic and fallow

Plantation

Rocky / uncovered areas

Water body

Habitation

Percentage 

C
la

ss
 n

am
e 

a) Loss and gain in 2019 and 2069 (AZM)

Loss

-80 -60 -40 -20 0 20 40 60 80 100 120

Gallery forest / riparian formation

Dense forest

Woodlands

Tree and shrub savannahs

Field mosaic and fallow

Plantation

Rocky / uncovered areas

Water body

Habitation

Percentage 

C
la

ss
 n

am
e 

b) Net change in 2019 and 2069 (AZM)

Loss Gain

https://doi.org/10.14710/geoplanning.8.2.115-126


Issiako et al. / Geoplanning: Journal of Geomatics and Planning, Vol 8, No 2, 2021, 115-126 
DOI: 10.14710/geoplanning.8.2.115-126 

 

124 

Changes observed with "Implementing Multi-criteria Zoning with Efficiency (MZE)" Scenario 

 

Data source: Landsat 8 OLI-TIRS, LCLU, 2019, cell transmission rules, Simulation sotfware MOLUSCE (QGIS Remote Sensing) 

Figure 8. Observed changes between 2019 and Implementation Multi-Criteria Zoning" scenarios 

The exam of the Figure 8 shows that this loss is likely to be in favour of gallery, dense forest, woodlands 

and plantations, which will gain 2%, 5%, 13% and 79% respectively in area, or 207,359 ha (Figure 8a). This 

indicates that the establishment of fields in the FCAS will be regulated by a new forest management policy in 

the period 2019-2069. In contrast, gallery, dense forest, woodlands, plantations will experience a net positive 

change of 23.25% (Figure 8b). 

 

4. Discussion 

The diachronic analysis with the combination of the transition matrix allowed to highlight the different 

forms of conversion that the land cover units in the Forêt Classée de l'Alibori Supérieur underwent between 

2009, 2014 and 2019. This concerns the regression of natural vegetation formations in favour of anthropogenic 

formations. Anthropogenic pressures on natural resources through these activities have favoured the 

transformation of natural formations (Issiako & Arouna, 2018; Mama et al., 2020). Knowledge of recent dynamics 

is essential to understand future evolution and its modelling (Paegelow et al., 2004).  

The MOLUSCE method is used for the spatial survey in 2069. This method is implemented on the 

Quantum GIS software in which there is a Modules for land cover Change Simulations (MOLUSCE) plugin 

(Hakim et al., 2019). The land cover change was predicted using a MOLUSCE analysis method based on the 

Cellular Automata method by Mirici et al. (2018) and Subiyanto & Suprayogi (2019). The Multi-Layer 

Perceptron (MLP) Cellular Automata Simulator tool simulates the land cover data for the period 2019 and the 

actual referenced 2019 land cover map obtained from the satellite image in 2019 was used to validate the model 

and the performance of the model. The overall accuracy indices of the land cover survey in 2069 are 87% (AZM) 

and 85% (MZE). This shows that the land cover survey based on the scenarios can be continued as the overall 

precision value is high. The results of the trend scenario1 (AZM), show that the most likely evolutionary trend 

will be the conversion of gallery forest, dense dry forest, woodlands and shrub and tree savannah into field and 

fallow mosaic. Thus, the areas of crop-fallow mosaics, settlements and plantations will increase by 2069. This 

increase could be explained by population growth and the lack of arable land in village areas. Deforestation due 

to selective logging will lead to the loss of plant biodiversity. Faced with these anthropogenic disturbances in 

these protected natural landscapes, which impart a change in composition and spatial configuration, biodiversity 

is in permanent danger in Benin and Africa (Mama et al., 2020). This is in line with the work of Thierry et al. 

(2018) and Seko et al. (2018) who conducted their research in the North-West and North-East of Benin 

respectively and who explain the high demand for cultivated land by the increase in population.  

In the framework of the implementation of the Multicriteria Efficiency Zoning (MZE) which is nothing 

else than the implementation of an integrated forest management plan that can combine both environmental 

conservation and agropastoral activities. The results obtained show that with the implementation of a new 

management policy, there will be no more anarchic installation of fields inside the forest. The probability of 

-100 -50 0 50 100

Gallery forest / riparian formation

Dense forest

Woodlands

Tree and shrub savannahs

Field mosaic and fallow

Plantation

Rocky / uncovered areas

Water body

Habitation

Percentage 

C
la

ss
 n

am
e 

a) Loss and gain in 2019 and 2069 (MZE)

Loss Gain

-100 -50 0 50 100

Gallery forest / riparian formation

Dense forest

Woodlands

Tree and shrub savannahs

Field mosaic and fallow

Plantation

Rocky / uncovered areas

Water body

Habitation

Percentage 

C
la

ss
 n

am
e 

b) Net change in 2019 and 2069 (MZE)

Loss Gain

https://doi.org/10.14710/geoplanning.8.2.115-126


Issiako et al. / Geoplanning: Journal of Geomatics and Planning, Vol 8, No 2, 2021, 115-126 
DOI: 10.14710/geoplanning.8.2.115-126 

 

125 

reconstitution and stability of natural vegetation formations by 2069 will be very high. The hypothesis that the 

evolutionary trends of the vegetation cover by 2069 vary according to the scenarios implemented is verified. 

 

5. Conclusion 

At the end of the spatial and temporal evaluation of the FCAS land use and land cover units, it appears 

that natural formations have regressed in favour of anthropogenic formations between 2009 and 2014 and 

between 2014 and 2019, despite the status of protected area with a management plan for this geographical space. 

Thus, the ecological balance of natural vegetation types has been severely disrupted. The projection of current 

land cover trends using cellular automata has made it possible to assess the dynamics of land cover and land use 

for the period 2019 - 2069. During this period, the cultivated area will increase significantly by almost 88% 

compared to the total forest area if the current trend is maintained with little implementation of the management 

plan. On the other hand, the implementation of an effective zoning system will allow the control of anthropic 

pressures, which will in turn allow the restoration of degraded areas. The maps translate into forward-looking 

trend scenarios and will allow the identification of degraded management areas on the one hand and favourable 

areas for the conversion of forest resources on the other. This research can provide decision-makers with the 

necessary data for the elaboration of a future spatial management plan for the conservation and rational 

exploitation of forest resources. Therefore, the prospective study of potential areas for plant biodiversity 

conservation deserves to be done to further inform decision makers. 

 

6. Acknowledgements 

Thank you to all those who have helped in this research. 

 

7. References 

Akindélé, G. (2000). Possibilités d’aménagement durable de la Forêt Classée de l’Alibori Supérieur: structure et 
dynamique des principaux groupements végétaux et périodicité d’exploitation. Ecole Polytechnique d’Abomey-
Calavi, Université d’Abomey-Calavi, Bénin. 

FAO. (2015). Evaluation des ressources forestières mondiales : rapport national du Bénin. 
FAO et PNUE. (2020). La situation des forêts du monde 2020. Forêts, biodiversité et activité humaine. [Crossref] 
Gbedahi, O. L. C., Biaou, S. S. H., Mama, A., Gouwakinnou, G. N., & Yorou, N. S. (2019). Dynamique du couvert 

végétal à Bassila au nord Bénin pendant et après la mise en œuvre d’un projet d’aménagement forestier. 
International Journal of Biological and Chemical Sciences, 13(1), 311–324. 

Hakim, A. M. Y., Baja, S., Rampisela, D. A., & Arif, S. (2019). Spatial dynamic prediction of landuse/landcover 
change (case study: tamalanrea sub-district, makassar city). IOP Conference Series: Earth and Environmental 
Science, 280(1), 12023. 

Houet, T. (2015). Usages des modèles spatiaux pour la prospective. Revue Internationale de Géomatique, 25(1), 123–
143. 

Issiako, D., & Arouna, O. (2018). Dynamique de l’occupation des terres avant et après l’élaboration du plan 
d’aménagement participatif de la forêt classée de l’Alibori Supérieur au Nord-Benin. Actes de La Conférence 

Scientifique Internationale OSFACO : Images Satellitaires Pour Un Meilleur Aménagement Des Territoires et Une 
Gestion Durable de La Biodiversité, 193–2010. 

Mama, A., Djaouga, M., Mazo, I., & Abiala, S. (2020). Cartographie De La Dynamique Spatio-Temporel De 
L’occupation Des Terres Dans L’arrondissement D’ifangni Au Sud-Benin. Annales De La Fashs, 1, 111–125. 

Mienmany, B. (2018). Analysis land use and land cover changes and the driving forces: A case study in Kaysone 
Phomvihan District, Laos. University of Porto, Porto. 

 

https://doi.org/10.14710/geoplanning.8.2.115-126
https://doi.org/https:/doi.org/10.4060/ca8642fr


Issiako et al. / Geoplanning: Journal of Geomatics and Planning, Vol 8, No 2, 2021, 115-126 
DOI: 10.14710/geoplanning.8.2.115-126 

 

126 

Mirici, M. E., Berberoglu, S., Akin, A., & Satir, O. (2018). Land use/cover change modelling in a mediterranean 
rural landscape using multi-layer perceptron and markov chain (mlp-mc). Appl. Ecol. Environ. Res, 16, 467–
486. 

Monier, A. (2010). Analyse des dynamiques d’occupation du sol de la vallée du Vicdessos (Pyrénées ariégeoises) entre 1953 
et 1983. 39. 

Moumouni, Y. I., Imorou, I. T., & Moussa, A. D. (2019). Caractérisation de la Dégradation à Travers la Diversité 
Floristique et la Structure de la Végétation dans le Bassin Moyen de la Sota au Nord-Bénin. International 
Journal of Progressive Sciences and Technologies, 18(1), 53–70. 

Munoz, A., Bock, J., Monnet, J.-M., Renaud, J. P., Jolly, A., & Riond, C. (2015). Évaluation par validation 
indépendante des prédictions des paramètres forestiers réalisées à partir de données LiDAR aéroporté. 
Revue Française de Photogrammétrie et de Télédétection, 211, 81–92. 

Olouloi, J., Mama, V. J., & Agbo, F. B. (2006). Modélisation de la dynamique de l’occupation des terres dans le 
Département des Collines au Bénin. Télédétection, 6(4), 305–323. 

Paegelow, M., Villa, N., Cornez, L., Ferraty, F., Ferré, L., & Sarda, P. (2004). Modélisations prospectives de 
l’occupation du sol. Le cas d’une montagne méditerranéenne. Cybergeo: European Journal of Geography. 

Rodriguez-Galiano, V. F., Ghimire, B., Rogan, J., Chica-Olmo, M., & Rigol-Sanchez, J. P. (2012). An assessment 
of the effectiveness of a random forest classifier for land-cover classification. ISPRS Journal of 
Photogrammetry and Remote Sensing, 67, 93–104. 

Samie, A., Deng, X., Jia, S., & Chen, D. (2017). Scenario-based simulation on dynamics of land-use-land-cover 
change in Punjab Province, Pakistan. Sustainability, 9(8), 1285. 

Seidou, A. A., Agbayigbo, A.-A., Traore, I. A., Houinato, M., & others. (2017). Spatio-temporal dynamics of 
natural rangelands exploited by transhumance cattle herds in the Classified Forest of Upper Alibori, 
northern Benin. American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS), 
33(1), 111–123. 

Seko, S. S. M. O., Ogouwale, E., Tente, B., & others. (2018). Modeling of Environmental Dynamics Using the 
SpaCelle Software in the Department of Alibori in Benin (West Africa). American Scientific Research Journal 
for Engineering, Technology, and Sciences (ASRJETS), 43(1), 211–226. 

Shao, Z., Zhang, Y., Zhang, L., Song, Y., & Peng, M. (2016). Combining spectral and texture features using 
random forest algorithm: Extracting impervious surface area in Wuhan. International Archives of the 
Photogrammetry, Remote Sensing and Spatial Information Sciences, 41, 351–358. 

Sinsin, B., Kampmann, D., Thiombiano, A., & Konaté, S. (2010). Atlas de la Biodiversité de l’Afrique de l’Ouest. 
Cotonou & Frankfurt/Main, 1. 

Subiyanto, S., & Suprayogi, A. (2019). Modeling And Spatial Analysis Of Change Settlement And Fair Market 
Land Price Using Markov Chain Model In Banyumanik District. KnE Engineering, 278–290. 

Thierry, A., Martin, P., Ismaila, T. I., & Brice, T. (2018). Modélisation des changements d’occupation des terres 
en région soudanienne au nord-ouest du Bénin. European Scientific Journal, ESJ, 14(12), 248. 

Toko Mouhamadou, I. (2014). Facteurs de la fragmentation des écosystèmes forestiers: cas des {\^\i}lots de forêts denses 
séches de la forêt classée des Monts Kouffé et de sa périphérie au Bénin. Thèse de Doctorat, Université d’Abomey-
Calavi, Bénin. 

Zakari, S., Imorou, I. T., Thomas, O. A. B., Djaouga, M., & Arouna, O. (2018). Application De La Télédétection 
Et Du Sig Au Suivi Des Formations Végétales De La Forêt Classée Des Trois Rivières Au Nord-Est Du 
Bénin. European Scientific Journal, ESJ, 14(15), 450. [Crossref] 

https://doi.org/10.14710/geoplanning.8.2.115-126
https://doi.org/10.19044/esj.2018.v14n15p450

