Nothing Special   »   [go: up one dir, main page]
Next Article in Journal
Advancing Geomatics: Innovation, Inclusivity, and Global Perspectives
Previous Article in Journal
Monitoring the Net Primary Productivity of Togo’s Ecosystems in Relation to Changes in Precipitation and Temperature
You seem to have javascript disabled. Please note that many of the page functionalities won't work as expected without javascript enabled.
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Geomatics
Volume 4
Issue 4
10.3390/geomatics4040019
geomatics-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Spatiotemporal Analysis of Land Use and Land Cover Dynamics of Dinderesso and Peni Forests in Burkina Faso

by
Alphonse Maré David Millogo
1,2,*,
Boalidioa Tankoano
3,
Oblé Neya
4,
Fousseni Folega
2,
Kperkouma Wala
2,
Kwame Oppong Hackman
5,6,7,
Bernadin Namoano
8 and
Komlan Batawila
2
1
Graduate Research Program on Climate Change and Disaster Risks Management, West African Science Service Centre on Climate Change and Adapted Land Use (WASCAL), Faculty of Human and Social Sciences, University of Lomé, Lomé BP 1515, Togo
2
Laboratory of Botany and Plant Ecology, Department of Botany, Faculty of Sciences, University of Lomé, Lomé BP 1515, Togo
3
Laboratoire Bioressource, Agrosystèmes et Santé de l’Environnement (LaBASE), Institute for Rural Development, Nazi BONI University of Bobo-Dioulasso, Bobo-Dioulasso BP 1091, Burkina Faso
4
National Centre for Scientific and Technological Research, Ouagadougou BP 7047, Burkina Faso
5
Regional Coordinator-Land Use, Land Cover, Land Degradation, WASCAL Competence Center, Ouagadougou BP 9507, Burkina Faso
6
Department of Environmental Management, University of Energy and Natural Resources, Sunyani P.O. Box 214, Ghana
7
Building and Road Research Institute (BRRI), Council for Scientific and Industrial Research (CSIR), Kumasi P.O. Box UP 40, Ghana
8
Center for Digital Engineering and Manufacturing, Cranfield University, Cranfield MK43 0AL, UK
*
Author to whom correspondence should be addressed.
Geomatics 2024, 4(4), 362-381; https://doi.org/10.3390/geomatics4040019
Submission received: 20 September 2024 / Revised: 29 September 2024 / Accepted: 29 September 2024 / Published: 4 October 2024
Browse Figures
Figure 1
<p>Dinderesso and Peni classified forest location.</p> "> Figure 2
<p>Landsat land use land cover assessment and household heads survey flowchart.</p> "> Figure 3
<p>Land uses land cover classes in Dinderesso and Peni classified forests.</p> "> Figure 4
<p>Land use land cover map of Dinderesso classified forest in 1986, 2006, 2010, 2016, and 2022.</p> "> Figure 5
<p>Land use land cover map of Peni classified forest in 1986, 2006, 2010, 2016, and 2022.</p> "> Figure 6
<p>Land use land cover change in the classified forest of Dinderesso in 1986, 2006, 2010, 2016, and 2022.</p> "> Figure 7
<p>Land use land change in the classified forest of Peni in 1986, 2006, 2010, 2016, and 2022.</p> "> Figure 8
<p>Anthropogenic drivers of Dinderesso and Peni classified forests degradation and deforestation.</p> "> Figure 9
<p>Dinderesso classified forest degradation and deforestation drivers.</p> "> Figure 10
<p>Peni classified forest degradation and deforestation drivers.</p> ">
Review Reports Versions Notes

Abstract

:
The sustainable management of protected areas has increasingly become difficult due to the lack of updated information on land use and land cover transformations caused by anthropogenic pressures. This study investigates the spatiotemporal dynamics of the Dinderesso and Peni classified forests in Burkina Faso from 1986 to 2022. First, a data driven method was adopted to investigate these forests degradation dynamics. Hence, relevant Landsat images data were collected, segmented, and analyzed using QGIS SCP plugin Random Forest algorithm. Ninety percent of the overall adjusted classification accuracies were obtained. The analysis also showed significant degradation and deforestation with high wooded vegetation classes such as clear forest and wooded savannah (i.e., tree savannah) converging to lower vegetation classes like shrub savannah and agroforestry parks. A second investigation carried out through surveys and field trips revealed key anthropogenic drivers including agricultural expansion, demographic pressure, bad management, wood cutting abuse, overexploitation, overgrazing, charcoal production, and bushfires. These findings highlight the critical need for better management to improve these protected areas.

1. Introduction

Forest resources, particularly the land use and land cover changes in protected forest areas driven by anthropogenic factors, have long been a global concern, especially in West African countries [1,2]. These changes pose significant challenges to biodiversity conservation and the preservation of local populations’ livelihoods. In this context, Geographic Information Systems (GIS) combined with satellite imagery presents a valuable opportunity to analyze the dynamics of land use and land cover in protected areas. Such analyses can enhance monitoring efforts by managers and inform sustainable policy investments, ultimately leading to better outcomes in biodiversity conservation and the protection of community livelihoods [3,4].
Numerous studies utilizing satellite imagery have demonstrated its effectiveness in detecting land use and cover changes within protected areas [5,6]. These findings underscore the utility of this technology in monitoring and managing forest areas that are increasingly threatened by human activities, which in turn jeopardize both biodiversity and the well-being of neighboring communities.
Located close to many rural municipalities and the Bobo-Dioulasso urban municipality, Dinderesso and Peni classified forests are persistently subjected to many illegal resource exploitation and prohibited human activities [7,8]. All these human activities, out of the legal regulations frame of Dinderesso and Peni classified forests, contribute to land use and land cover change, preventing them from ensuring their first function, biodiversity conservation, and socio-economic development as a second function.
The sustainable management of Dinderesso and Peni classified forests is challenging due to several anthropogenic pressures. Since they have been recognized as classified forests, which confer them a protection status to control human aggressions, some studies related to forest degradation drivers, flora composition, and structure have been conducted using surveys and inventories [8,9]. Recognizing the relevance of these studies to Dinderesso and Peni’s classified forest sustainable management, their implementation requires many resources, and they cannot provide detailed information on land use and land cover patterns on a global scale. Unfortunately, no knowledge of the spatial-temporal land use land cover dynamic of Dinderesso and Peni classified forests is available, making sustainable management difficult. Some spatial-temporal land use and land cover dynamics conducted in the Sudanian climatic area have highlighted the key role played by anthropogenic drivers in those forests’ land use land cover dynamics. Indeed, the Toessin classified forest land use and land cover dynamic analysis from 1986 to 2010 has revealed vegetation classes decreasing due to human activities [10]. Performing the same analysis in Tiogo classified forest from 1986 to 2006, the findings in Ref. [5] highlighted that the forest lost 0.49% of its vegetation cover annually. Many other studies’ results have also mentioned anthropogenic drivers’ role in protected areas’ land use and land cover change [6,11].
Looking at the evidence of protected areas’ threats due to human activities in the study area, it is necessary to fill the knowledge gap on Dinderesso and Peni classified forests spatiotemporal land use land cover dynamic to provide consistent and large-scale information for their sustainable management.
This study aims to analyze Dinderesso and Peni classified forests’ land use land cover dynamic from 1986 to 2022 and identify anthropogenic drivers of land use land cover change.

2. Materials and Methods

2.1. Study Area

The study was carried out in the classified forests of Dinderesso and Peni, located in Burkina Faso’s Sudanian climatic zone, specifically in the province of Houet (Figure 1). The Dinderesso classified forest is located between longitudes 4°18′46″ and 4°26′40″ W and latitudes 11°11′05″ and 11°18′10″ N. Initially created by decree number 422/SE of 27 February 1936 with a surface area of 7000 hectares (ha), the Dinderesso classified forest area was increased to 8500 hectares (ha) following the Burkina Faso decree issued under the reference number 3006/SE of 26 August 1941.
Peni classified forest is located between longitudes 4°27′26.5″ and 4°29′37.5″ W and latitudes 10°55′2.5″ and 10°56′33″ N. This forest, with an area of 1131.68 hectares (ha), was classified following the decree 3389 SE/F issued in 24 September 1942.
The Dinderesso and Peni classified forests enjoy the climate of the Sudanian zone. This climatic zone has the lowest thermal amplitudes in the country, varying between 20 and 25 degrees Celsius. During the rainy season that lasts approximatively six months, the rainfall is often between 900 mm and 1100 mm.
The vegetation in the Sudanian climate zone is characterized by savannah formations interspersed with gallery forests along rivers. These include a variety of protected species such as Faidherbia albida, Lannea microcarpa, Parkia biglobosa, Tamarindus indica, Vitellaria paradoxa, and Isoberlinia doka [1]. The herbaceous plants in this climate zone are primarily from the genera Pennisetum, Cymbopogon, and Hyparrhenia. In addition, the classified forests of Dinderesso and Peni also house artificial plantations and agroforestry parks, some of which have been established legally, while others are unauthorized.
The relief in the Dinderesso classified forest is uneven, with highly heterogeneous soils varying between lithosols, ferruginous tropical soils, and soils that have smoothly evolved. Crossed by cliffs, the relief of the Peni classified forest is characterized by a succession of hills. The soil types encountered mainly composed of ferralitic soils, soils that have sally evolved, and soils with sesquioxides.

2.2. Methods

Figure 2 shows the study’s overall methodology. It comprises Landsat images, land use and land cover assessment, and a survey on anthropogenic drivers of forest degradation and deforestation.

2.2.1. Landsat Data Collection

For land use land cover dynamic assessment, a set of five multi-temporal Landsat images covering path 197/52 were downloaded from the United States Geological Survey (USGS) website (available at: [12]) for the years 1986, 2006, 2010, 2016, and 2022 (Table 1). The images were selected at the same year period with image availability and cloud cover percent criteria to minimize the biases linked to solar angle, vegetation phenological change, and soil humidity variation [13]. Twenty (20) ground truth GPS (UTM zone 30 North with 3 m of precision) data were collected from each land cover class during the field trips from 23 August 2023 to 27 August 2023 in Dinderesso classified forest and from 8 December 2023 to 9 December 2023 in Peni classified forest. The national topographic database (BNDT, website: [14]) and the national land cover database for 2012 from Burkinabe Geographic Institute (IGB) were also used.

2.2.2. Digital Preprocessing of Landsat Images

All the downloaded Landsat images were already georeferenced to Universal Transverse Mercator (UTM) zone 30 North within the Coordinate Reference System WGS84. Atmospheric correction using QGIS Semi-automatic Classification Plugin (SCP) was applied to convert images from Digital Numbers (DN) to reflectance. The reflectance images of each respective study year were stacked before extracting the study area using forest masks from the national topographic database.

2.2.3. Digital Image Processing

The different studied years’ panchromatic band sets obtained after the data pre-processing step were uploaded to the QGIS SCP for image classification. Auxiliaries data, including false color composition (5-4-3 and 4-3-2, respectively, for OLI_TIRS and TM band sets); biophysical indices which are Normalized Difference Vegetation Index (NDVI), Normalized Difference Built-up Index (NDBI), and Normalized Difference Water Index (NDWI); the national land use database; Google Earth high-resolution image; and the ground truth data were used, as shown in Figure 2, to identify land cover classes, respectively, for Dinderesso and Peni classified forests with their periphery. Based on the particularity of the colored composition image to highlight each land cover class spectral signature [15], 563 and 377 regions of interest (ROI) polygons were selected randomly and homogeneously from the reference 2022 multi-spectral image in false color composition. Each ROI selected was assigned the corresponding land class [15]. For each ROI random sample selected, 2/3 (60% for training and 40% for validation) was used for Random Forest model training and 1/3 for the testing [16]. A total of 377 and 256 training ROIs were used as input data into the Orfeo ToolBox (OTB) “TrainImagesClassifier” program to train and generate the classification model. Then, the Random Forest model was used as input with the reference multi-spectral image to perform the whole image classification into the OTB “ImageClassifier” program. The supervised Random Forest classification was also applied to 1986, 2006, 2010, and 2016 multi-spectral images based on ROIs selected from identified land cover class areas that stayed stable over the years [16].

2.2.4. Testing of Classified Images

The classified images were validated using visual and statistical approaches [16]. The visual comparison of the Random Forest classified image with the false color composition image revealed a high similarity between the two images. This visual testing, recognized as a subjective method, was used as a first step in the classification testing before applying the statistical testing, which is more accurate. The ROIs testing polygons were used to validate the Random Forest classified image into the QGIS SCP “Accuracy” program [16]. The statistical classification testing generated confusion matrices for each classified image in an Excel spreadsheet that was used to assess the classification accuracy by computing the overall accuracy, producer and user accuracy, and kappa coefficient [6,17].
The producer accuracy of land cover classes was measured by dividing the total number of perfectly classified pixels within the testing zone of the given class and the total number of pixels in the testing zones of said class. The user’s accuracy was calculated for each class by dividing the total number of pixels correctly classified in the given class in all testing areas of the said class by the total number of pixels classified in that class in the given class within all testing zones for all LULC classes. For more reliability to the Random Forest classification described as an outperformer classifier model [18], other metrics, including the accuracy, the recall, and the F1-score, were computed. They were computed based on the true positive (TP) for recognized objects, false negative (FN) for non-detected objects, and false positive (FP) for incorrectly identified objects. The following equations [19] were used to assess these metrics:
A c c u r a c y = T P T P + F P
R e c a l l = T P T P + F N
F 1 − s c o r e = A c c u r a c y   ×   R e c a l l   A c c u r a c y   +   R e c a l l
In addition to previous metrics computing, the accuracy assessment good practices recommended by [20] were considered to highlight the misclassification bias by adjusting producer, user accuracy, and estimated areas. This adjustment was achieved using Card’s confusion matrix correction [21] using a Microsoft Excel version 2408 spreadsheet [22].
For post-classification processing, the SCP “Classification sieve” algorithm was applied to the classified images to clean isolated pixels and thus improve their sharpness. Following this processing and the various analyses relating to the dynamics of LULC, the images were vectorized before their mapping in QGIS.

2.2.5. Evaluation of Land Use Land Cover Dynamics

For this analysis, we once again used the shapefiles of the two forests to extract the data relating to each of the forests. The images extracted after cleaning the isolated pixels were used as input data in the SCP “Land cover change” algorithm to generate the transition matrices. The SCP’s “Classification report” algorithm was subsequently used to generate, in Excel format, information relating to the areas of the different LULC classes for each study period. These data made it possible to calculate the average annual rate of land cover change between two periods using the following formula from [23]:
C = B 2   −   B 1 B 1 × 100 T 2 − T 1
where C is the average annual rate of change; B1 is the amount of land cover type in time 1 (T1); and B2 is the amount of land cover type in time 2 (T2).

2.2.6. Identification of Anthropogenic Factors of Degradation and Deforestation

During the field visits to identify land cover classes in the Dinderesso and Peni classified forests, the anthropogenic factors of degradation and deforestation were noted, and pictures were recorded. An individual interview with the two departmental heads in charge of Dinderesso and Peni classified forests management as well as the literature review contributed to confirming field observations. A survey was conducted to collect additional data related to degradation and deforestation causes of both forests. The survey involved household heads of eight neighboring communities.

Household Heads Sampling

A survey was undertaken in eight neighboring communities of Dinderesso and Peni classified forests to identify forest degradation deforestation drivers. The survey participants were mainly the household chief in the neighboring communities. The number of households considered in this study has been determined by using the Schwartz sampling equation:
N = z 2 × p × ( 1 − p ) / α 2
where N is the sampling size, z is the confidence level at 95%, z is equal to 1.96, p is the proportion of the population is 0.5, and α is the error equal to 5%.
Considering that the four neighboring communities of Dinderesso and Peni classified forests have different populations, the equation was applied based on each community household proportion determined through the fifth national census of communities in Burkina Faso [24]. A total of 850 heads of households from eight neighboring communities were interviewed (see Table 2).

Data Collection

The data were collected using the Kobocollect tool. A semi-structured questionnaire was addressed in an individual interview with household heads randomly chosen in each community. Prior to each interview, the interviewee approval form was filled and signed. A local guide was used in data collection to reduce bias related to local language adaptation.

Data Processing

The data obtained from the previous step was cleaned by removing empty data lines and updating some outlier data. After the data cleaning process, 454 answers, representing 53.4% of the initial sampling size, were analyzed using an Excel spreadsheet to identify Dinderesso and Peni classified forests anthropogenic drivers.

3. Results

3.1. Land Use Land Cover Classes Identified

Seven and six LULC classes were identified in Dinderesso and Peni’s classified forests, respectively (Figure 3). These classes are related to Water Body/Wetland Area (WB/WA), Gallery Forest (GF), Wooded Savannah/Tree Savannah (WS/TS), Shrub Savannah (SS), Agroforestry Park (AP), and Build Area/Bare Soil (BA/BS).

3.2. Classification Testing

The complementary data set highlights the details of both Dinderesso and Peni classified image overall accuracy, the kappa coefficient, producer user accuracy, and user accuracy.
The overall accuracy unadjusted of Dinderesso classified images ranges from 92.81% to 94.32% for 2006 and 2010. The kappa coefficient ranges from 0.90 to 0.92 for 2006, 1986, 2010, and 2022. The lowest producer accuracy unadjusted of Dinderesso classified forest land cover classes is 81.48% for WS/TS. The lowest user accuracy unadjusted is 80% for AP.
For the Peni classified forest, the overall accuracy unadjusted ranges from 94.85% to 99.76% for 2010 and 2006, respectively. The kappa coefficient ranges from 0.92 to 0.995 for 2010 and 2006, respectively. The lowest producer accuracy unadjusted of Peni classified forest land cover classes is 90.48% for BA/BS. The lowest user accuracy unadjusted is 90% for WB/WA.
Table 3 presents the metric values obtained from the RF confusion matrix. Dinderesso’s accuracy values range from 0.95 to 0.98 for 2016 and 2006, respectively. Peni’s classified forest accuracy values range from 0.96 to 0.99 for 1986 and 2006. All these high values suggest a high overall performance of image classification.
The recall values of Dinderesso classified forest range from 0.95 to 0.97, respectively, for 2016 and 2006. Peni classified forest recall values range from 0.97 to 0.99 for 2010, 1986, and 2016, respectively. The recall values in both Dinderesso and Peni classified forests are high, attesting to reliable performance in class identification.
Dinderesso classified forest F1-score values fluctuate from 0.95 to 0.97 for 2016 and 2010, respectively. Peni classified forest F1-score ranges from 0.97 to 0.99 for 2022, 2006, and 2016, respectively. The high F1-score values in both forests indicate a good balance between accuracy and recall, showing the performance of the RF classification model.
Table 4 and Table 5 present the 1986 and 2022 bias-adjusted confusion matrix by Card’s method derived from sample counts and the proportion of LULC class areas in the raw confusion matrix available in the complementary data. The adjusted metrics show similarity with unadjusted metrics concerning overall accuracy, which is over 90%, and user adjusted accuracy metrics, which are at least equal to 80% for all the different year map classifications.
Conversely to user accuracy, the producer accuracy adjusted results decreased considerably compared to unadjusted producer accuracy with low values under 50%, particularly for the Built Area/Bare Soil LULC class for Dinderesso classified forest 1986, 2016, and 2022 maps and for the Water Body/Wetland Area 2016 map. In the Peni classified forest, these low values of producer accuracy are only related to the Built Area/Bare Soil LULC class for the 2010, 2016, and 2022 maps.

3.3. Dynamics of Land Use in the Classified Forests of Dinderesso and Peni from 1986 to 2022

Figure 4 and Figure 5 show the distribution in LULC classes of the classified forests of Dinderesso and Peni over 1986, 2006, 2010, 2016, and 2022. The distribution of these LULC classes during the studied years in the Dinderesso and Peni classified forests show some changes. Figure 6 and Figure 7 show the LULC classes’ evolution over time. It can be seen that Dinderesso classified forest high-vegetation classes, including Clear Forest (CF) and Wooded Savannah/Tree Savannah (WS/TS), have been changing to low-vegetation classes, including Shrub Savannah (SS) and Agroforestry Park (AP) from 1986 to 2022. The same observation is made also in the Peni classified forest where Wooded Savannah/Tree Savannah (WS/TS) and Gallery Forest (GF) have been changing to Shrub Savannah (SS) and Agroforestry Park (AP). The annual growth area rate in complementary data for Shrub Savannah (SS) and Agroforestry Park (AP) in the Dinderesso classified forest is 3.86% and 3.09%, respectively. In the Peni classified forest, the Shrub Savannah (SS) growth area rate per annum is assessed at 358.54% and 0.95% for the Agroforestry Park (AP) growth rate per annum. The Shrub Savannah in the Peni classified forest changed from 1.49 ha in 1986 to 193.81 ha in 2022, which explains the too-large growth rate per annum observed.

3.4. Anthropogenic Factors of Degradation and Deforestation of the Classified Forests of Dinderesso and Peni

3.4.1. Field Observation

The field trip in both Dinderesso and Peni classified forests, and their periphery area, revealed that mainly the practices related to human activities majorly contribute to their degradation and deforestation. Figure 8 shows these anthropogenic drivers including clearing for installing crops and orchards, cutting trees for fuelwood and charcoal production, livestock overgrazing, improper harvesting of non-timber forest products, and bushfires.

3.4.2. Household Heads Profile

The heads of households interviewed were mostly men, who represent 84.58%. The household head’s age mean is 46.89 ± 14.77, ranging from 20 to 90 years old. Most household heads are natives, representing 75.11%. Household heads belong to 20 ethnic groups, with 59.03% of Bobo, the most represented ethnic group, and 0.22%, the least represented ethnic group, encompassing Karaboro, Gouin, and Gourounsi. Regarding their education level, 40.31% of household heads surveyed did not receive any education, and 1.54% said they had a university level education. Household heads are engaged in multiple professional occupations to improve their income. Still, considering the main professional occupation, most household heads surveyed are farmers, representing 76.21% of the sample, while the student is the least represented professional occupation with 0.44%.

3.4.3. Dinderesso and Peni Classified Forests Neighboring Communities’ Perception of Degradation and Deforestation Drivers

Neighboring Dinderesso classified forest communities mentioned sixteen forest degradation and deforestation drivers. Figure 9 shows the proportion of forest degradation and deforestation drivers concerning Dinderesso classified forest. The main forest degradation and deforestation drivers mentioned are related to human activities. The drivers mentioned are related to wood cutting abuse (WC), which represents 34.48%; demographic pressure (DP) with 18.68%; poverty (PV), representing 11.49%; bad management (BM), with 8.05%; bushfires (BF), with 6.61%; overexploitation (OE), representing 6.32%; overgrazing (OG), 3.74%; agriculture (AG), 2.87%; rainfall scarcity (RF), 2.30%; natural death (ND), 1.44%; silting (ST), 1.15%; wind (WD), 0.86%; 0.57% each for Poaching (BC), drought (DR), and soil fertility (FS); and 0.29% for pollution (PL).
The results of Peni classified forest degradation and deforestation drivers perceived by neighboring communities are globally similar to those identified in the classified forest of Dinderesso. Figure 10 illustrates the Peni classified forest degradation and deforestation drivers. Peni classified forest neighboring communities mentioned at least fifteen forest degradation deforestation drivers. Similar to the Dinderesso classified forest, most drivers are related to human activities. Forest degradation and deforestation drivers mentioned are wood cutting abuse (WC), 33.49%; demographic pressure (DP), 11.93%; agriculture (AG), 9.17%; 8.72% each for bad management (BM) and elephants (EP); charcoal production (CC), representing 7.34%; poverty (PV) with 6.88%; 3.67% each for rainfall scarcity (RF) and bushfires (BF); overexploitation (OE), with 1.83%; natural death (ND), 1.38%; 0.92% each for overgrazing (OG), drought (DR), and diseases (DS); and 0.46% for soil fertility (FS).

4. Discussion

4.1. Peni Land Use Land Cover Classes and Landsat Images Analysis

The Landsat images classification of Dinderesso and Peni classified forests allowed the identification of seven land use land cover classes. These classes, including Water Body/Water Area (WB/WA), Gallery Forest (GF), Clear Forest (CF), Wooded Savannah/Tree Savannah (WS/TS), Shrub Savannah (SS), Agroforestry Park (AP), and Build Area/Bare Soil (BA/BS), align with the national LULC database and are consistent with many authors’ LULC descriptions in the Sudanian climatic zone [6,11,25]. The supervised random forest classification approach provided high overall accuracy, kappa coefficient, producer accuracy, and user accuracy statistics. These statistic accuracies are consistent with other land use land cover studies achieved using Random Forest [15,17].
The high values of classification overall accuracy and kappa coefficient suggest a high agreement between the classified thematic maps and the reference image used to generate the classification [26]. According to [27], these high statistical results could be explained by training and testing pixels that are homogeneously chosen within LULC classes and well identified. These statistics values are also supported by the metrics, including the recall and the F1-score, which highlighted the Random Forest model’s high performance and reliability in classified images. Despite the high accuracy statistics obtained, some confusions between LULC classes are highlighted by the producer and user accuracy statistics, which are not all 100%. The confusion observed between LULC classes could be related to the difficulties in choosing training ROIs among classes with very close spectral signatures. These confusions are common in the Sahel area, where vegetation class heterogeneity associated with Landsat image spatial resolution may affect the identification of LULC classes correctly [28]. Research conducted by Aka et al. [29] in the Ivory Coast revealed the limits of Landsat-8 image spatial resolution by distinguishing some land cover classes where Sentinel-2 images with a better spatial resolution proved more successful. According to [30], the confusion between LULC classes could also belong to pixel sampling. According to [18], which compared the performance between Random Forest and the SVM models, the Random Forest is sensitive to the number of pixels. This case may concern the land cover class Water Area/Wetland Body (WA/WB) and Built Area/Bare Soil with low producer accuracy adjusted values in our study, where it was difficult to sample many training and testing polygons ROIs covering many pixels related to that particular class. This study, compared to previous LULC assessment studies conducted in Burkina Faso [10,31], highlighted the relevance of recall and F1-score metrics and the accuracy assessment by applying Olofsson’s [20] and Card’s [21] methodology in the classification testing approach.

4.2. Land Use and Land Cover Dynamic

The main change observed in land use and land cover dynamics in both Dinderesso and Peni classified forests from 1986 to 2022 is the decreasing area of high-vegetation classes, including clear forest, wooded savannah/tree savannah, and the increasing area of low-vegetation classes, including shrub savannah and agroforestry park. These results reflect the trend towards Dinderesso and Peni classified forests degradation and deforestation, characterized by high-vegetation classes replacement into low-vegetation classes. These results are consistent with those of [7], who highlighted Dinderesso and Peni classified forests degradation and deforestation. Similar research studies achieved in the Sudanian area have also highlighted protected forest degradation and deforestation. A similar study conducted in the classified Koulbi forest in the Sudanian area showed that clear forest and tree savannah areas are being replaced by shrub savannah and cropland [32]. Other research studies in the Sudanian area in Burkina Faso [11,27] and in West Africa [33,34,35] mentioned this transformation of forest high-vegetation classes into lower-vegetation classes. Many factors, including human activities and climate change, could explain the decrease in high-vegetation class areas observed in Dinderesso and Peni classified forests. The contribution of human activities in protected areas to degradation and deforestation has been recognized by many authors [36,37]. Dinderesso and Peni classified forests’ proximity to Bobo-Dioulasso town makes them ideal sources of fuelwood supply for the population. According to [27], the population exploits fuelwood in high-vegetation classes such as clear forest and wooded savannah/tree savannah, where the biggest trees are located. Tree cutting abuse has been identified among the most common causes of degradation and deforestation of Dinderesso classified forest [8]. The demographic pressure, land tenure, and rural population poverty could explain the increase in agroforestry park areas in both forests. A survey conducted in Zambia revealed that population growth, loss of soil fertility, land tenure management, and lack of laws are the leading causes of agricultural expansion into Zambia’s forest areas [38]. According to [39], agricultural land inaccessibility due to population poverty contributes to agricultural expansion in forest areas. Climate change also contributes to forest degradation and deforestation. Some studies have highlighted the influence of climate change through drought events on forest degradation and deforestation [2]. The findings of [40] revealed that the main challenge in the long term for sixteen food tree species in their natural habitat in Burkina Faso is related to climate change. Beyond the threats of human activities and climate change impact, this study’s findings raise crucial points related to biodiversity conservation in Dinderesso and Peni classified forests and the sustainability management system suitable for these protected areas. The conversion of high-vegetation classes into low-vegetation classes negatively affects plant species and their habitat, which can lead to the extinction of some plant species. According to [41], the overexploitation of medicine plant species, such as Securidaca longepedunculata, negatively affects the availability of adult plants in their natural habitat. The changes observed in Dinderesso and Peni classified forests with high threats of human activities despite their protection status highlight the inefficiency of the current management system. Some authors have suggested improving the management of protected forests for their sustainability [6,11]. A suitable management system for protected forest areas should be developed based on sustainable biodiversity conservation and the preservation of the livelihoods of the riparian population.

4.3. Anthropogenic Drivers of Land Use Land Cover Degradation

The findings highlighted that Dinderesso and Peni classified forests degradation and deforestation drivers are mainly related to human activities. These findings are consistent with many research works [1,8]. Deforestation anthropogenic drivers in both Dinderesso and Peni classified forests are related to wood cutting abuse, demographic pressure, bad management, charcoal production, agriculture, poverty, overgrazing, overexploitation, bushfires, and poaching. According to the interviewed household heads, they cut trees for many purposes, including energy needs, house construction, mortar production, and charcoal production. Fuelwood represents the first source of energy in Burkina Faso [42]. According to [43], more than 97% of household energy needs in Burkina Faso are met by woody forest resources. The proximity of both Dinderesso and Peni classified forests to Bobo-Dioulasso, the second urban town in the country, increases pressure on forest resources due to the demography. The research work of [27] highlighted that the proximity of Tiogo classified forest to Ouagadougou and Koudougou cities increases fuelwood exploitation, negatively affecting forest resources. The demographic pressure also negatively affects neighboring communities’ access to agricultural land. The current study revealed that 76.21% of surveyed household heads practice farming as their main occupation. As a result of the lack of agricultural land, some people, without any respect for classified forest regulations, establish their farms within the forest. Similar results have been highlighted by [38] in Zambia and Burkina Faso [10,44]. Since Dinderesso and Peni classified forests are not fenced, they became an open source of fodder and water for livestock during the dry and rainfall seasons. This uncontrolled and illegal practice of breeders contributes to the degradation of both forest resources. Many authors reported livestock’s negative impacts on forest resources [45,46]. The management system of both forests qualified as a bad one, which has been mentioned by neighboring communities who claimed that they are not taken into account in forest management and that foresters do not have sufficient means to monitor and control the whole forest. Classified forest management in Burkina Faso limits population access to forest resources, generating conflicts between neighboring communities and forest managers. For protected areas’ sustainable management and population livelihoods, many authors have suggested improving the management system in Burkina Faso [44,47]. According to surveyed household heads, poverty pushes people to non-timber forest products overexploitation for income. Previous studies have highlighted the overexploitation of non-timber forest products [41,48]. In addition to the surveyed anthropogenic drivers of forest degradation and deforestation, Dinderesso and Peni classified forests land cover change is also related to rainfall scarcity, drought, and wind. This information is consistent with other studies conducted in the Sudanian area, highlighting climate change variability’s negative effect on forest land cover [11,27]. Forest degradation and deforestation increase greenhouse gas emissions in the atmosphere, which has consequences such as rainfall variability, drought, heat waves, and temperature increases. Some studies’ results highlighted that climate change may negatively affect the natural distribution area of many plant tree species and forest land cover in Burkina Faso [40,49]. The findings on Dinderesso and Penic classified forests aligned with previous studies, highlighting that anthropogenic drivers were mainly responsible for forest LULC change in the study area. These findings are similar to the current situation of the country’s protected forest, which is marked by degradation and deforestation under anthropogenic drivers. Considering this situation, improving the relationship between protected areas and humans is crucial for a double benefit of biodiversity conservation and population livelihood preservation.

5. Conclusions

This study significantly contributes to bridging the knowledge gap concerning the LULC dynamics of the Dinderesso and Peni classified forests, as well as the anthropogenic factors driving these changes between 1986 and 2022. Utilizing a robust and reliable methodology that combines Landsat image classification via a Random Forest classifier model with field surveys conducted among household heads in eight neighboring communities, the study yielded important insights that can enable forest managers and the government to take appropriate actions.
The study firstly demonstrated a high accuracy and reliability of the Random Forest model for Landsat image classification, achieving an accuracy and F1-score exceeding 0.90. Secondly, the analysis of LULC dynamics revealed a trend toward degradation and deforestation in the Dinderesso and Peni classified forests. This trend is marked by the conversion of areas with high vegetation cover, such as clear forests and wooded savannahs/tree savannahs, into areas with low vegetation cover, including shrub savannahs and agroforestry parks. In the Dinderesso forest, shrub savannah and agroforestry parks have been expanding at annual rates of 3.86% and 3.09%, respectively. In the Peni forest, the annual growth rates for these categories are even more dramatic, at 358.54% and 0.95%, respectively. Furthermore, the study identified key anthropogenic drivers—such as excessive wood cutting, population pressure, poor management practices, agriculture, overexploitation, charcoal production, overgrazing, bushfires, and poverty—accounting for more than 70% of the causes behind the degradation and deforestation of these forests. The research also highlights systemic management deficiencies in both the Dinderesso and Peni forests, which are reflective of broader challenges faced by classified forests across the country, undermining efforts toward sustainable biodiversity conservation and the preservation of local communities’ livelihoods. The findings of this study are consistent with previous research that has raised alarms about the degradation and deforestation of protected forests, underscoring the need for the development of effective management systems for these areas and their neighboring communities.

Author Contributions

Conceptualization, A.M.D.M., B.T., O.N., F.F. and K.W.; methodology; A.M.D.M., B.T., O.N., F.F. and K.W; software, A.M.D.M. and B.T.; testing, A.M.D.M., B.T., O.N., F.F., B.N., K.W. and K.B.; formal analysis, A.M.D.M.; investigation, A.M.D.M.; resources, A.M.D.M. and B.T.; data curation, A.M.D.M.; writing—original draft preparation; A.M.D.M.; writing—review and editing, A.M.D.M., B.T., O.N., F.F., K.W., B.N., K.O.H. and K.B.; visualization, A.M.D.M., B.T., O.N., F.F. and K.W.; supervision, A.M.D.M., B.T., O.N., F.F. and K.W.; project administration; O.N., F.F. and K.W.; funding acquisition, A.M.D.M. All authors have read and agreed to the published version of the manuscript.

Funding

This study was funded by the German Federal Ministry of Education and Research (BMBF), which funded this research through the West African Science Service Center on Climate Change and Adapted Land Use (WASCAL) program.

Institutional Review Board Statement

The study was conducted in accordance with the Declaration of Helsinki, and approved by the Institutional Review Board (or Ethics Committee) of Burkina Faso National Assembly (protocol code 0011-2021/AN of 30 March 2021).

Informed Consent Statement

Informed consent was obtained from all subjects involved in the study.

Data Availability Statement

The corresponding author certifies that the data used to achieve land use land cover dynamics and findings in the current study will be available for publication.

Acknowledgments

The authors are grateful to the German Federal Ministry of Education and Research (BMBF), which funded this research through the West African Science Service Center on Climate Change and Adapted Land Use (WASCAL) program. Our thanks also go to the Dinderesso and Peni local populations, all the neighboring populations of Dinderesso and Peni classified forests, the resource persons, and all those who helped us carry out this study.

Conflicts of Interest

The authors declare no conflicts of interest. The funders had no role in the study’s design, data collection, analysis, interpretation, manuscript writing, or decision to publish the results.

References

  1. Thiombiano, A.; Kampmann, D. Atlas de la Biodiversité de l ’ Afrique de l ’ Ouest. Biodiversity Atlas of West Africa. Burkina Faso; BIOTA: Bonn, Germany, 2010; Volume 2, ISBN 9783981393309. [Google Scholar]
  2. FAO. Évaluation Des Ressources Forestières Mondiales: Forêt, Biodiversité Et Activité Humaine; FAO: Rome, Italy, 2020; ISBN 9789251307182. [Google Scholar]
  3. Falcucci, A.; Maiorano, L.; Boitani, L. Changes in land-use/land-cover patterns in Italy and their implications for biodiversity conservation. Landsc. Ecol. 2007, 22, 617–631. [Google Scholar] [CrossRef]
  4. Sánchez-Reyes, U.J.; Niño-Maldonado, S.; Barrientos-Lozano, L.; Treviño-Carreón, J. Assessment of land use-cover changes and successional stages of vegetation in the natural protected area altas cumbres, Northeastern Mexico, using landsat satellite imagery. Remote Sens. 2017, 9, 712. [Google Scholar] [CrossRef]
  5. Millogo, D.; Nikiema, A.; Koulibaly, B. Analyse de l ’ évolution de l ’ occupation des terres à partir de photographies aériennes de la localité de Loaga dans la province du Bam, Burkina Faso Analysis of the evolution of soil occupation using aerial photographs at Loaga area in Bam province. Int. J. Biol. Chem. Sci. 2017, 11, 2133–2143. [Google Scholar] [CrossRef]
  6. Dimobe, K.; Ouédraogo, A.; Soma, S.; Goetze, D.; Porembski, S.; Thiombiano, A. Identification of driving factors of land degradation and deforestation in the Wildlife Reserve of Bontioli (Burkina Faso, West Africa). Glob. Ecol. Conserv. 2015, 4, 559–571. [Google Scholar] [CrossRef]
  7. MECV. Situation des Forêts Classées du Burkina Faso et Plan de Rehabilitation; MECV: Ouagadougou, Burkina Faso, 2007. [Google Scholar]
  8. Djiguemde, F.; Ouoba, P.; Romaric, N.; Somda, I. Typologie des facteurs de dégradation de la Forêt Classée de Dindéresso (FCD), réserve naturelle péri urbaine au sud-ouest du Burkina Faso. Afrique Sci. 2023, 23, 119–129. [Google Scholar]
  9. Yaméogo, J.T.; Ouattara, R.Y.S.; Tankoano, B.; Hien, M.; Ouoba, P. Flore, structure et état sanitaire des peuplements ligneux des parcs agroforestiers des forêts de Dindéresso et de Kuinima à l’ouest du Burkina Faso. Eur. Sci. J. ESJ 2020, 16, 48–70. [Google Scholar] [CrossRef]
  10. Belem, M.; Zoungrana, M.; Nabaloum, M. Les effets combinés du climat et des pressions anthropiques sur la forêt classée de Toéssin, Burkina Faso. Int. J. Biol. Chem. Sci. 2019, 12, 2186–2201. [Google Scholar] [CrossRef]
  11. Savadogo, O. Fourniture des Services Écosystémiques par les Réserves Totale et Partielle de Faune de Bontioli dans le Sud-Ouest du Burkina Faso: Perception Paysanne et Contribution à la Résilience des Populations Riveraines Dans un Contexte de Changement Climatique; Université Senghor: Alexandrie, Egypte, 2021; p. 68. [Google Scholar]
  12. USGS Earthexplorer. Available online: http://earthexplorer.usgs.gov/ (accessed on 16 November 2023).
  13. Tankoano, B.; Hien, M.; Dibi, N.H.; Sanon, Z.; Yameogo, J.T.; Somda, I. Dynamique spatio-temporelle des savanes boisées de la forêt classée de Tiogo au Burkina Faso. Int. J. Biol. Chem. Sci. 2015, 9, 1983. [Google Scholar] [CrossRef]
  14. IGB. Available online: https://www.igb.bf/ (accessed on 22 November 2022).
  15. Tempa, K.; Aryal, K.R. Semi-automatic classification for rapid delineation of the geohazard-prone areas using Sentinel-2 satellite imagery. SN Appl. Sci. 2022, 141, 15. [Google Scholar] [CrossRef]
  16. Denis, A. Travaux Pratiques de Télédétection Spatiale II Avec données et logiciels; ULiège—Université de Liège [Faculté des Sciences]: Liège, Belgium, 2020; p. 429. Available online: https://hdl.handle.net/2268/240835 (accessed on 10 June 2022).
  17. Ghosh, A.; Sharma, R.; Joshi, P.K. Random forest classification of urban landscape using Landsat archive and ancillary data: Combining seasonal maps with decision level fusion. Appl. Geogr. 2014, 48, 31–41. [Google Scholar] [CrossRef]
  18. Traoré, F.; Palé, S.; Zaré, A.; Traoré, M.K.; Ouédraogo, B.; Bonkoungou, J. A Comparative Analysis of Random Forest and Support Vector Machines for Classifying Irrigated Cropping Areas in The Upper-Comoé Basin, Burkina Faso. Indian J. Sci. Technol. 2024, 17, 713–722. [Google Scholar] [CrossRef]
  19. Thanh Tam, N.; Thanh Dat, H.; Minh Tam, P.; Tuyet Trinh, V.; Thanh Hung, N.; Quyet-Thang, H.; Jun, J. Monitoring agriculture areas with satellite images and deep learning. Appl. Soft Comput. J. 2020, 95, 16. [Google Scholar] [CrossRef]
  20. Olofsson, P.; Foody, G.M.; Herold, M.; Stehman, S.V.; Woodcock, C.E.; Wulder, M.A. Good practices for estimating area and assessing accuracy of land change. Remote Sens. Environ. 2014, 148, 42–57. [Google Scholar] [CrossRef]
  21. Card, D.H. Using Known Map Category Marginal Frequencies to Improve Estimates of Thematic Map Accuracy. Photogramm. Eng. Remote Sensing 1982, 48, 431–439. [Google Scholar]
  22. Obodai, J.; Adjei, K.A.; Odai, S.N.; Lumor, M. Land use/land cover dynamics using landsat data in a gold mining basin-the Ankobra, Ghana. Remote Sens. Appl. Soc. Environ. 2019, 13, 247–256. [Google Scholar] [CrossRef]
  23. Long, H.; Tang, G.; Li, X.; Heilig, G.K. Socio-economic driving forces of land-use change in Kunshan, the Yangtze River Delta economic area of China. J. Environ. Manag. 2007, 83, 351–364. [Google Scholar] [CrossRef]
  24. INSD. Fichier des Localités du 5 e RGPH; INSD: Ouagadougou, Burkina Faso, 2022; p. 400. [Google Scholar]
  25. SE/CCNUCC. Niveau d ’ Émissions de Référence pour les Forêts du Burkina Faso; SE/CCNUCC: Ouagadougou, Burkina Faso, 2020; p. 85. Available online: https://redd.unfccc.int/media/ (accessed on 19 September 2024).
  26. Mohd Hasmadi, I.; Pakhriazad, H.Z.; Shahrin, M.F. Evaluating supervised and unsupervised techniques for land cover mapping using remote sensing data. Malays. J. Soc. Sp. 2009, 5, 1–10. [Google Scholar]
  27. Tankoano, B.; Sanon, Z.; Hien, M.; Dibi, N.H.; Yameogo, J.T.; Somda, I. Anthropogenic pressure and vegetation dynamics in the Classified Forest of Tiogo in Burkina Faso: Contribution of remote sensing|Pression anthropique et dynamique végétale dans la Forêt Classée de Tiogo au Burkina Faso: Apport de la Télédétection. Tropicultura 2016, 34, 193–207. [Google Scholar]
  28. Dash, P.; Sanders, S.L.; Parajuli, P.; Ouyang, Y. Improving the Accuracy of Land Use and Land Cover Classification of Landsat Data in an Agricultural Watershed. Remote Sens. 2023, 15, 4020. [Google Scholar] [CrossRef]
  29. Saint, K.; Aka, R.; Da Dibi, H.N.; Koffi, J.N. Étude comparative de Sentinel-2 et Landsat-8 Oli à l’ évaluation de l’occupation du sol du site R amsar d’Azagny, Sud de la Côte d’Ivoire Résumé. Afrique Sci. 2022, 20, 48–63. [Google Scholar]
  30. Khatami, R.; Mountrakis, G.; Stehman, S.V. Mapping per-pixel predicted accuracy of classified remote sensing images. Remote Sens. Environ. 2017, 191, 156–167. [Google Scholar] [CrossRef]
  31. Tankoano, B.; Hien, M.; Dibi, N.; Sanon, Z.; L Akpa, Y.; Sokeng, V.; Somda, I. Cartographie de la dynamique du couvert végétal du Parc National des Deux Balé à l’Ouest du Burkina Faso [Mapping land cover dynamics of Deux Balé National Park, Western Burkina Faso]. Int. J. Innov. Appl. Stud. 2016, 16, 837–846. [Google Scholar]
  32. Sanon, Z.; Hien, M.; Tankoano, B.; Karlson, M.; Somda, I. Impact des villages et du réseau routier sur la structure spatiale de la forêt classée de Koulbi (Sud-Ouest du Burkina Faso). Sci. Tech. Rev. Burkinabè la Rech. Sci. Nat. Appl. Sci. Tech. 2019, 38, 59–72. [Google Scholar]
  33. N’Guessan, A.E.; N’dja, J.K.; Yao, O.N.; Amani, B.H.K.; Gouli, R.G.Z.; Piponiot, C.; Zo-Bi, I.C.; Hérault, B. Drivers of biomass recovery in a secondary forested landscape of West Africa. For. Ecol. Manag. 2019, 433, 325–331. [Google Scholar] [CrossRef]
  34. Boukeng, D.; Elvire, J. Dynamique et Services des Écosystèmes Agroforestiers Autour d’une Aire Protégée du Cameroun. Ph.D. Thesis, Univeriste LAVAL, Quebec, QC, Canada, 2021; p. 132. [Google Scholar]
  35. Atakpama, W.; Joseph, K.; Klevor, A.; Kanda, M.; Dourma, M.; Batawila, K.; Akpagana, K.; Kanda, M.; Kanda, M.; Batawila, K.; et al. Wildfires degrade and deforest Fazao-Malfakassa National Park ( FMNP ) in Togo. Afr. J. Land Policy Geospat. Sci. 2024, 7, 2657–2664. [Google Scholar] [CrossRef]
  36. Sanou, L. Perceptions locales des perturbations écologiques sur la dynamique de la végétation de la réserve de biosphère transfrontalière Parc National du W, Afrique de l’Ouest. Rev. Ecosystèmes Paysages 2023, 3, 1–14. [Google Scholar] [CrossRef]
  37. Tindano, E.; Ganame, M.; Bongoungou, S.; Bayen, P. Effect of artisanal gold mining on woody plant diversity in Western Burkina Faso. Sci. Afr. 2024, 24, 1–12. [Google Scholar] [CrossRef]
  38. Phiri, D.; Mwitwa, J.; Ng’andwe, P.; Kanja, K.; Munyaka, J.; Chileshe, F.; Hamazakaza, P.; Kapembwa, S.; Kwenye, J.M. Agricultural expansion into forest reserves in Zambia: A remote sensing approach. Geocarto Int. 2023, 38, 221320. [Google Scholar] [CrossRef]
  39. Miyamoto, M. Poverty reduction saves forests sustainably: Lessons for deforestation policies. World Dev. 2020, 127, 104746. [Google Scholar] [CrossRef]
  40. Gaisberger, H.; Kindt, R.; Loo, J.; Schmidt, M.; Bognounou, F.; Da, S.S.; Diallo, O.B.; Ganaba, S.; Gnoumou, A.; Lompo, D.; et al. Spatially explicit multi-threat assessment of food tree species in Burkina Faso: A fine-scale approach. PLoS ONE 2017, 12, e0184457. [Google Scholar] [CrossRef]
  41. Savadogo, S.; Ouoba, P.A.; Da, D.E.C. Extraction des PFNL et Menace sur les especes pourvoyeuses au Burkina Faso: Cas de Kokologho et Tenado dans la region du Centre-Ouest. Rev. Ivoirienne Sci. Technol. 2019, 33, 240–258. [Google Scholar]
  42. Kabore, A.M.; Somda, J.; Savadogo, M.; Nianogo, A.J. Bois-Energie au Burkina Faso: Consolidation des Moyens D’existences Durable (2006–2009); IUCN: Ouagadougou, Burkina Faso, 2009. [Google Scholar]
  43. Mundler, P.; Audras, A.; Ler, I.L. Inra Sfer Cirad. 2010, pp. 1–19. Available online: https://agritrop.cirad.fr/554636/1/document_554636.pdf (accessed on 19 September 2024).
  44. Sama, A.; Ramdé, Z.; Ouédraogo, P.C.; Sikuzani, Y.U.; Dipama, J.; Bogaert, J. Pression anthropique et dynamique de l ’ occupation du sol autour du corridor n ° 1 du complexe écologique Pô-Nazinga- Sissili au Burkina Faso. Tropicultura 2023, 41, 2200. [Google Scholar] [CrossRef]
  45. Crovo, O.; Aburto, F.; Montecino, F.; Rodríguez, R. Effects of livestock grazing on soil health and recovery of a degraded Andean Araucaria forest. L. Degrad. Dev. 2021, 32, 4907–4919. [Google Scholar] [CrossRef]
  46. Papoutsa, C.; Kouhartsiouk, D.; Themistocleous, K.; Christoforou, M.; Hadjimitsis, D.G. Monitoring of land degradation from overgrazing using space-borne radar and optical imagery: A case study in Randi Forest, Cyprus. In Proceedings of the SPIE, Edinburg, UK, 26–29 September 2016; Volume 10005, p. 10. [Google Scholar] [CrossRef]
  47. MEEVCC. Plan D’Investissement Forestier Version Préliminaire: 15 Mai 2011 Plan d’Investissement Forestier (PIF/Burkina Faso); MEEVCC: Ouagadougou, Burkina Faso, 2020. [Google Scholar]
  48. Ministery of Environment. Faso Volume 1: Les Facteurs de Deforestation et de Degradation des Forets au Burkina Faso. Rapport d’Etude; Volume 1: Tendances Actuelles; Ministery of Environment: Ouagadougou, Burkina Faso, 2019; Volume 1.
  49. Alphonse Mare David, M.; Jean-Louis, D.; Felicien, T. Analyse de la Dynamique de la Population de Detarium Microcarpum Guill et Perr. au Burkina Faso. Master’s Thesis, Universite de Liege, Liege, Belgium, 2019; p. 65. [Google Scholar]
Geomatics 04 00019 g001
Figure 1. Dinderesso and Peni classified forest location.
Figure 1. Dinderesso and Peni classified forest location.
Geomatics 04 00019 g001
Geomatics 04 00019 g002
Figure 2. Landsat land use land cover assessment and household heads survey flowchart.
Figure 2. Landsat land use land cover assessment and household heads survey flowchart.
Geomatics 04 00019 g002
Geomatics 04 00019 g003
Figure 3. Land uses land cover classes in Dinderesso and Peni classified forests.
Figure 3. Land uses land cover classes in Dinderesso and Peni classified forests.
Geomatics 04 00019 g003
Geomatics 04 00019 g004
Figure 4. Land use land cover map of Dinderesso classified forest in 1986, 2006, 2010, 2016, and 2022.
Figure 4. Land use land cover map of Dinderesso classified forest in 1986, 2006, 2010, 2016, and 2022.
Geomatics 04 00019 g004
Geomatics 04 00019 g005
Figure 5. Land use land cover map of Peni classified forest in 1986, 2006, 2010, 2016, and 2022.
Figure 5. Land use land cover map of Peni classified forest in 1986, 2006, 2010, 2016, and 2022.
Geomatics 04 00019 g005
Geomatics 04 00019 g006
Figure 6. Land use land cover change in the classified forest of Dinderesso in 1986, 2006, 2010, 2016, and 2022.
Figure 6. Land use land cover change in the classified forest of Dinderesso in 1986, 2006, 2010, 2016, and 2022.
Geomatics 04 00019 g006
Geomatics 04 00019 g007
Figure 7. Land use land change in the classified forest of Peni in 1986, 2006, 2010, 2016, and 2022.
Figure 7. Land use land change in the classified forest of Peni in 1986, 2006, 2010, 2016, and 2022.
Geomatics 04 00019 g007
Geomatics 04 00019 g008
Figure 8. Anthropogenic drivers of Dinderesso and Peni classified forests degradation and deforestation.
Figure 8. Anthropogenic drivers of Dinderesso and Peni classified forests degradation and deforestation.
Geomatics 04 00019 g008
Geomatics 04 00019 g009
Figure 9. Dinderesso classified forest degradation and deforestation drivers.
Figure 9. Dinderesso classified forest degradation and deforestation drivers.
Geomatics 04 00019 g009
Geomatics 04 00019 g010
Figure 10. Peni classified forest degradation and deforestation drivers.
Figure 10. Peni classified forest degradation and deforestation drivers.
Geomatics 04 00019 g010
Table 1. Landsat image characteristics.
Table 1. Landsat image characteristics.
ForestLandsat NDate AcquiredSensor_IDCloud CoverPathRow
Dinderesso516 November 1986TM019752
57 November 2006TM019752
518 November 2010TM019752
817 October 2016OLI_TIRS0.1719752
818 October 2022OLI_TIRS019752
Peni516 November 1986TM019752
57 November 2006TM019752
518 November 2010TM219752
817 October 2016OLI_TIRS0.1719752
818 October 2022OLI_TIRS019752
Table 2. Number of household heads surveyed per community.
Table 2. Number of household heads surveyed per community.
ForestCommunityHousehold Heads Number
DinderessoBanakeledaga136
Dinderesso41
Nasso113
Ouolonkoto133
PeniPeni190
Sokouranie30
Taga37
Gnafongo170
Total850
Table 3. Accuracy, recall, F1-score, and kappa Landsat classification testing metrics for Dinderesso and Peni classified forests.
Table 3. Accuracy, recall, F1-score, and kappa Landsat classification testing metrics for Dinderesso and Peni classified forests.
ForestYearAccuracyRecallF1-ScoreKappa
Dinderesso19860.970.960.960.92
20060.980.970.970.9
20100.970.960.970.92
20160.950.950.950.91
20220.960.960.960.92
Peni19860.960.990.980.97
20060.990.980.990.99
20100.970.970.980.92
20160.990.990.990.97
20220.980.980.970.96
Table 4. Dinderesso classified forest Card’s confusion matrix with producer accuracy (PA), user accuracy (UA), and areas adjusted with a 95% confidence interval.
Table 4. Dinderesso classified forest Card’s confusion matrix with producer accuracy (PA), user accuracy (UA), and areas adjusted with a 95% confidence interval.
1986/Dinderesso classified forest
WB/WAGFCFWS/TSSSAPBA/BSTotalUA (%)
WB/WA0.060.000.000.000.000.000.000.06100.00 ± 0.00
GF0.000.020.000.000.000.000.000.0295.52 ± 4.99
CF0.000.000.410.000.000.000.000.4299.12 ± 1.00
WS/TS0.000.000.000.260.000.010.000.2796.34 ± 4.09
SS0.000.000.000.000.060.000.000.06100.00 ± 0.00
AP0.000.000.000.030.000.140.010.1880.00 ± 10.67
BA/BS0.000.000.000.000.000.000.000.00100.00 ± 0.00
Total0.060.020.420.290.060.150.011.00
PA (%)97.93 ± 3.9897.63 ± 13.3899.82 ± 0.1989.77 ± 5.5597.89 ± 4.0493.62 ± 6.7339.75 ± 32.88
Adjusted area (ha) 524.69 ± 21147.66 ± 223662.42 ± 382526.02 ± 183516.86 ± 211349.64 ± 19495.32 ± 79
OA (%)95 ± 2.24
2022/Dinderesso classified forest
WB/WAGFCFWS/TSSSAPBA/BSTotalUA (%)
WB/WA0.030.000.000.000.000.000.000.03100.00 ± 0.00
GF0.000.020.000.000.000.000.000.0298.46 ± 3.02
CF0.010.010.330.000.000.030.000.3887.93 ± 8.46
WS/TS0.000.000.000.120.000.010.000.1486.89 ± 8.54
SS0.000.000.000.000.140.000.000.14100.00 ± 0.00
AP0.000.000.000.000.000.280.000.2998.86 ± 2.23
BA/BS0.000.000.000.000.000.000.000.0095.89 ± 3.48
Total0.040.030.330.120.140.320.011.00
PA (%)70.77 ± 28.4276.07 ± 35.6899.90 ± 0.1997.20 ± 5.09100.00 ± 0.0087.73 ± 7.3446.41 ± 34.20
Adjusted area (ha)395.25 ± 159241.37 ± 1132949.78 ± 2831078.59 ± 1171234.71 ± 0.002850.09 ± 24573.82 ± 54
OA (%)93.25 ± 3.48
Legend: = Water Body/Wetland Area (WB/WA), (WB/WA), Gallery Forest (GF), Clear Forest (CF), Wooded Savannah/Tree Savannah (WS/TS), Shrub Savannah (SS), Agroforestry Park (AP), and Build Area/Bare Soil (BA/BS).
Table 5. Peni classified forest Card’s confusion matrix with producer accuracy (PA), user accuracy (UA), and areas adjusted with a 95% confidence interval.
Table 5. Peni classified forest Card’s confusion matrix with producer accuracy (PA), user accuracy (UA), and areas adjusted with a 95% confidence interval.
1986/Peni classified forest
WB/WAGFWS/TSSSAPBA/BSTotalUA (%)
WB/WA0.270.000.000.000.030.000.2991.30 ± 11.77
GF0.000.030.000.000.000.000.03100.00 ± 0.00
WS/TS0.000.000.230.000.000.000.2398.53 ± 2.88
SS0.000.000.000.000.000.000.0097.18 ± 3.88
AP0.000.000.000.000.430.000.43100.00 ± 0.00
BA/BS0.000.000.000.000.000.010.01100.00 ± 0.00
Total0.270.030.230.000.460.011.00
PA (%)100.00 ± 0.00100.00 ± 0.0099.98 ± 0.02100.00 ± 0.0093.76 ± 7.11100.00 ± 0.00
Adjusted area (ha) 294.26 ± 3834.92 ± 0.00252.33 ± 71.49 ± 0.00509.78 ± 3911.43 ±0.00
OA (%)97.12 ± 3.50
2022/Peni classified forest
WB/WAGFWS/TSSSAPBA/BSTotalUA (%)
WB/WA0.060.000.000.000.000.000.0690.00 ± 13.49
GF0.000.010.000.000.000.000.0194.74 ± 7.20
WS/TS0.000.000.130.000.000.000.13100.00 ± 0.00
SS0.000.000.000.170.000.010.1897.01 ± 4.11
AP0.000.000.000.000.620.000.62100.00 ± 0.00
BA/BS0.000.000.000.000.000.000.0099.23 ± 1.51
Total0.060.010.130.180.620.011.00
PA (%)100.00 ± 0.0073.89 ±37.8499.62 ± 0.5198.21 ± 3.44100.00 ± 0.0014.36 ± 16.92
Adjusted area (ha)62.37 ± 913.27 ± 7144.27 ± 1193.81 ± 11683.65 ± 0.006.84 ± 8
OA (%)98.79 ± 1.12
Legend: = Water Body/Wetland Area (WB/WA), (WB/WA), Gallery Forest (GF), Wooded Savannah/Tree Savannah (WS/TS), Shrub Savannah (SS), Agroforestry Park (AP), and Build Area/Bare Soil (BA/BS).
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Millogo, A.M.D.; Tankoano, B.; Neya, O.; Folega, F.; Wala, K.; Hackman, K.O.; Namoano, B.; Batawila, K. Spatiotemporal Analysis of Land Use and Land Cover Dynamics of Dinderesso and Peni Forests in Burkina Faso. Geomatics 2024, 4, 362-381. https://doi.org/10.3390/geomatics4040019

AMA Style

Millogo AMD, Tankoano B, Neya O, Folega F, Wala K, Hackman KO, Namoano B, Batawila K. Spatiotemporal Analysis of Land Use and Land Cover Dynamics of Dinderesso and Peni Forests in Burkina Faso. Geomatics. 2024; 4(4):362-381. https://doi.org/10.3390/geomatics4040019

Chicago/Turabian Style

Millogo, Alphonse Maré David, Boalidioa Tankoano, Oblé Neya, Fousseni Folega, Kperkouma Wala, Kwame Oppong Hackman, Bernadin Namoano, and Komlan Batawila. 2024. "Spatiotemporal Analysis of Land Use and Land Cover Dynamics of Dinderesso and Peni Forests in Burkina Faso" Geomatics 4, no. 4: 362-381. https://doi.org/10.3390/geomatics4040019

APA Style

Millogo, A. M. D., Tankoano, B., Neya, O., Folega, F., Wala, K., Hackman, K. O., Namoano, B., & Batawila, K. (2024). Spatiotemporal Analysis of Land Use and Land Cover Dynamics of Dinderesso and Peni Forests in Burkina Faso. Geomatics, 4(4), 362-381. https://doi.org/10.3390/geomatics4040019

Article Metrics

Back to TopTop