Sikuzani, Et Al. Anthropogenic Pressures and Spatiotemporal Dynamics of Forest Ecosystems in The Rural and Border Municipality of Kasenga

Landscape and Ecological Engineering (2024) 20:195–212
https://doi.org/10.1007/s11355-023-00589-z
ORIGINAL PAPER
Anthropogenic pressures and spatio‑temporal dynamics of forest

ecosystems in the rural and border municipality of Kasenga (DRC)
Yannick Useni Sikuzani1,2 · Ildephonse Kipili Mwenya3 · Héritier Khoji Muteya1,4 · François Malaisse4 ·
Sylvestre Cabala Kaleba1 · Jan Bogaert4,5
Received: 11 December 2022 / Revised: 21 October 2023 / Accepted: 5 December 2023 / Published online: 11 January 2024
© The Author(s) under exclusive licence to International Consortium of Landscape and Ecological Engineering 2024
Abstract
Migration and the dependence of rural communities on forest resources for subsistence have profoundly altered the com-
position and spatial structure of the landscapes of the border municipality of Kasenga in the southeast of DR Congo. The
spatio-temporal dynamics of anthropogenic effects on forest ecosystems were mapped and quantified in the municipality of
Kasenga using Landsat image classification from 1989 to 2022, combined with landscape ecology metrics to analyze spatial
patterns. Our results show that the landscape has undergone profound disturbances. The area of large patches of forest that
used to dominate the landscape has been reduced by a factor of 4 from 1989 to 2022, thus indicating the anthropogenic
impact on the fragmentation of forest ecosystems. If in 33 years (from 1989 to 2022) forest has lost more than a third of their
coverage through the dissection, fragmentation and attrition of patches, agriculture, grassland and wetland, and built-up and
bare land have recorded a progressive dynamic resulting from the creation and aggregation of patches. These anthropogenic
transformations, coupled with a lack of land management planning, will compromise the future of forest ecosystems since
the level of landscape disturbance has quintupled from 1.1 to 5.5 in 33 years. There is then an urgent need to develop an
integrated and participatory land management strategy to preserve forest resources and guarantee their resilience.
Keywords Forests · Ecosystem services · Anthropization of landscapes · Deforestation · Migration flows
Introduction
In sub-Saharan Africa, forests are a survival capital for

human populations, especially as they provide various
edible forest products (Akinnifesi et al. 2008; Mng’omba
* Yannick Useni Sikuzani
sikuzaniu@unilu.ac.cd et al. 2015), medicinal plants (Moyo et al. 2015; Moura
et al. 2018), fuelwood (Dewees et al. 2010), construction
1
Unité Ecologie, Restauration Ecologique et Paysage, Faculté materials (Assale et al. 2016), and fodder (Chinomona et al.
des Sciences Agronomiques, Université de Lubumbashi, BP 2018; Handavu et al. 2019). They also contribute to cli-
1825, Lubumbashi, Democratic Republic of the Congo
mate and water flow regulation (Vinya et al. 2019), carbon
2
Centre de Compétences en Planification et Gestion sequestration (Lusambo et al. 2016), biodiversity conserva-
Urbaine, Université de Lubumbashi, BP 1825, Lubumbashi,
Democratic Republic of the Congo tion (Arroyo-Rodriguez et al. 2020), and soil conservation
3 (Xiao et al. 2017). However, the dependence of sub-Saharan
Département de Gestion des Ressources Naturelles
Renouvelables, Faculté des Sciences Agronomiques, rural communities on forest resources for their livelihoods
Université Technologique Katumba Mwanke, BP 74, is increasingly causing deforestation and the degradation of
Lubumbashi, Democratic Republic of the Congo forest ecosystems (Yaovi et al. 2021). As a result, a loss of
4
Unité Biodiversité et Paysage, Gembloux Agro‑Bio Tech, 129 million hectares of forest was reported between 1990
Université de Liège, 5030 Gembloux, Belgium and 2015 up to almost 151 million hectares in 2020, rep-
5
Ecole Régionale Post-Universitaire d’Aménagement resenting an annual rate of − 0.13% of the forest in sub-
et de Gestion Intégrés des Forêts et Territoires Saharan Africa (Keenan et al. 2015; FAO 2020). According
Tropicaux (ERAIFT), BP 15373, Kinshasa, to data from Global Forest Watch, while in the early 2000s
Democratic Republic of the Congo
Vol.:(0123456789)
196 Landscape and Ecological Engineering (2024) 20:195–212
gross primary forest loss averaged 150,000 ha per year, it stability. Although quantifying the extent of forest cover and
has been around 450,000 ha per year since 2014 (Karsenty its change is crucial to highlight the ecological processes
2020). In Central Africa, the Democratic Republic of the taking place to guide planners, such information is non-
Congo, a country with high diversity and large forest cover, existent for the rural and border municipality of Kasenga.
is becoming an illustration of the negative evolution of forest Yet, by analyzing the spatial pattern of landscapes and their
cover (Potapov et al. 2012; Tyukavina et al. 2018), especially dynamics through remote sensing, geographic information
since from 1990 to 2010 the deforestation rate remained con- systems, and landscape ecology analysis tools, timely con-
stant at more than 0.20%, which is equivalent to the loss clusions regarding fundamental ecological processes can be
of 311,000 ha per year (FAO 2010). However, during the defined and vice versa (Bogaert and André 2013). Remote
decade 2010–2020, the deforestation rate has quadrupled to sensing provides accurate and precise data, enabling detailed
0.83%, or 1,101,000 ha of forest lost annually (FAO 2020). and reliable monitoring of land cover changes over time
As for woodland, particularly of the miombo wood- (Hemati et al. 2021). In addition, it provides a consistent
land type, they are experiencing a significant decline in and reproducible methodology for monitoring land cover
the Katanga region (Khoji et al. 2022), particularly in the dynamics, which is essential for long-term multi-scale stud-
Katangese copper belt area (KCA), a region of intense min- ies (Amarnath et al. 2017; Li et al. 2023). As a result, it can
ing activities (Mwitwa et al. 2012; Potapov et al. 2012; be more cost-effective than traditional ground-based meth-
Cabala et al. 2017). It should be noted that the miombo ods, particularly for monitoring large areas (Shapiro et al.
woodland is a dominant vegetation type in the Zambe- 2015). On the other hand, it is a non-disturbing method for
zian region, characterized by the preponderance of species the areas to be studied (Bakó et al. 2014).
belonging to the genera Brachystegia, Julbernardia and Lisa et al. (2021) used images from the Landsat 5 sensor
Isoberlinia of the Caesalpionioideae (Malaisse 2010). The to estimate the urban extent, count the number of buildings
causes of the regression of miombo woodland cover, sup- within the border City of Goma (DR Congo) and calculate
ported by rapid population growth, are mainly agricultural housing densities, by manually digitizing urban spatial
development, charcoal production (peri-)urbanization and expansion. Yet, several Landsat sensors have been launched,
the development of mining activities (Useni et al. 2017). and each new generation brings technological improvements
The extent of anthropogenic activities is such that the KCA and innovations that make them invaluable tools for monitor-
lost 15% of the miombo woodland cover between 2002 ing the environment or managing natural resources (Wulder
and 2015 (Cabala et al. 2017). The rural areas of the KCA et al. 2022). Since images from recent sensors (e.g. Landsat
deserve special attention as they are constantly experiencing 8 and Landsat 9) are compatible with those from earlier sen-
unplanned local changes in land cover (Mpanda et al. 2022), sors, it is generally recommended to use multi-date images
while the spatial pattern of the landscape is timidly evolving which enable continuous analysis and long-term monitoring
at the scale of the KCA (Mwitwa et al. 2012). In turn, these of land cover change (Li and Chen 2020).
changes in the spatial pattern of landscape (composition However, since several forest cover products are avail-
and configuration) are likely to lead to local climatic distur- able (e.g., Hansen et al. 2013), in landscape ecology,
bances, contributing to global change and affect biodiversity the impact of spatial scale in the analysis of landscape
(Bogaert and André 2013; Biaou et al. 2022). dynamics is well documented (Bogaert et al. 2005). The
The rural municipality of Kasenga to the northeast of reduction of the spatial scale can thus bring to light phe-
Lubumbashi City is no exception to this trend. Its location nomena that are not observable at large spatial scales
on the border with Zambia means that it currently has a (Turner et al. 1989; Wu et al. 1997; Barima et al. 2009).
custom post open to border traffic, making it attractive to Thus, the analysis of the local dynamics of forest cover
neighboring populations because of the economic opportu- would be relevant to understand and highlight local driv-
nities (CAID 2017). Once known as a fishing area due to its ers, which potentially contribute to national and global
proximity to the Luapula River separating it from Zambia changes (Sabah et al. 2022). Also, in the current con-
(Malaisse 2010), this area has collapsed due to overexploita- text of decentralization in DRC, the study of landscape
tion of fisheries resources, leading to a rush of people into dynamics at the municipality level seems very timely, as
slash-and-burn agriculture (Mpanda et al. 2022) as well as the issues of land cover/use allocation, which have major
charcoal production to meet local needs and those of the City impacts on the landscape, are the responsibility of tradi-
of Lubumbashi (Kabulu et al. 2018). tional chiefs around the city (Mpanda et al. 2022). The
In the rural and border municipality of Kasenga, the for- municipality scale is therefore relevant for any study that
est ecosystems, which are otherwise scarce, have never been could lead to the implementation of a responsible natural
the subject of rational management. Constantly subjected to resource management strategy. Also, areas characterized
anthropogenic actions, these ecosystems are in an advanced by slash-and-burn agriculture deserve special attention
state of degradation, which compromises their ecological as they correspond to places where the direction of local
Landscape and Ecological Engineering (2024) 20:195–212 197
changes is unplanned and frequent, while the spatial Materials and methods
structures of the landscape evolve timidly at the regional
scale. Yet, these changes in landscape spatial pattern Study area
lead to forest fragmentation, local climatic disturbances
(Assani 1999) which in turn contribute to global change. The present study covers the rural and border municipality
For this reason, the present study maps and quantifies of Kasenga, including the City of Kasenga and its periphery.
the spatio-temporal pattern dynamics of forest ecosys- This area corresponds administratively to part of the Kisa-
tems in the rural and border municipality of Kasenga in mamba sector and covers a total area of 783 km2 (Fig. 1).
response to anthropogenic pressures. It tests the hypoth- With an altitude ranging from 925 to 960 m, the municipal-
esis that, due to a lack of land management planning, ity of Kasenga has an Aw5 type climate according to the
overpopulation coupled with unsustainable use in the Köppen classification system (Kasongo 2008), with an alter-
context of climate change, the landscape dynamics are nating rainy season (October to April) and dry season (May
characterized by a fragmentation of forest ecosystems, to September). The dry season is currently being extended
materialized by a decrease in the patch area in parallel at the expense of the rainy season, with temperatures often
with an increase in their number, to the benefit of agri- being harsh. Average annual temperatures vary between
cultural areas and savannahs, which extend their hold on 16 and 33 °C, while annual rainfall of around 1260 mm is
the landscape and amplify its anthropization level. recorded (CAID 2017). In contrast to previous years, the
frequency and volume of rain is currently disturbed, the dry
Fig. 1 Location of the rural and border municipality of Kasenga in Congo and Zambia. The two countries are separated by the Luapula
the southeast of the Democratic Republic of the Congo (A), in the River, on which a border post (port), controlling the movement of
territory of Kasenga (B), Upper Katanga province. The municipal- people and their goods, is located
ity is located on the border between the Democratic Republic of the
season being extended at the expense of the rainy season required for change analysis (Mas 2000). In addition, atmos-
(Kalombo 2016). In the first half of the last century, the pheric corrections, using the fast line-of-sight atmospheric
municipality was dominated by miombo woodland grow- analysis of spectral hypercubes method based on radiative
ing on ferralsols and cambisols (Engelen et al. 2006). This transfer models (Kruse 2004), were performed to improve
forest is currently fragmented due to extensive anthropo- sharpness and facilitate visual interpretation. False com-
genic activities (Cabala et al. 2017). The precious trees of posite colors combining the green, red and near-infrared
the miombo woodland, mainly the Pterocarpus tinctorius, bands were created (Byomkesh et al. 2012). The infrared
are coveted by economic operators who export them to Asia, and red bands distinguish vegetation from other land use
where the market value is very high (Mpanda et al. 2022). units (Barima et al. 2009). Next, a series of unsupervised
The municipality, surrounded by savanna, is crossed from classifications were carried out to provide a rough indica-
north to south by a wetland area in which rice cultivation is tion of the main land cover types in the study area and to
developed: the Lubi. Agriculture, charcoal production, resi- help identify the training plots based on 296 points collected
dential livestock farming, NTFP collection and small-scale randomly with a GPS 64st Garmin (accuracy of about 3 m)
trade are the main activities of the population, whose size during the fieldwork from May 21, 2022 to June 30, 2022,
has been estimated at around 75,000 people in 2021 (Rap- with reference to the oldest image (1989). Indeed, a set of
port semestriel de la Commune de Kasenga 2022). Agri- 468 points representative of the selected land cover types
culture is the main activity of the inhabitants of Kasenga was used for the supervised classification, with 120 for for-
territory, practiced by farmers with hoes. The main agricul- est, 105 for agriculture, 70 for savanna, 67 for grassland
tural products are maize, cassava, rice, groundnuts, beans, and wetland, 46 for built-up and bare land and 60 for water.
and sweet potatoes. The periphery of the municipality also However, the classification data include different polygons
produces sugarcane in Makungu in the Kisamamba group of land cover types, which have been located by geographic
(CAID 2017). coordinate points. These training plots were used to train the
classification of all years, considering only the stable areas
Data between the old images and recent image for 2022 taken as
reference. For this reason, discussions with resource persons
To understand the spatio-temporal changes in the rural and in the villages around Kasenga (e.g., traditional chiefs) were
border municipality of Kasenga and the resulting land- decisive to retain only the stable/unchanged polygons of the
scape structure, six Landsat images (collection 2, level 1) land cover types over the period of our study. Consequently,
with a spatial resolution of 30 m from the Thematic Map- the same training zones defined and collected in 2022 were
per (02/06/1989; 26/05/1998; 10/05/2004; 24/05/2009) superimposed on all the remaining Landsat images (between
and OLI (07/06/2014 and 13/06/2022) sensors were used. 1989–2014).
These images, which were downloaded from the United However, supervised classifications based on the maxi-
States Geographical Survey website (https://earthexplo mum likelihood algorithm were performed. This algorithm
rer.usgs.gov), were chosen based on their availability and uses the statistics of the training areas to calculate the maxi-
low cloud cover. To better discriminate the different land- mum probability that the pixels belong to a predefined land
scape features, all dates of six images correspond to the cover (Caloz and Collet 2001). Indeed, by assigning each
dry season in the region. In addition, these dates cover the pixel to a specific class with a probability estimate, the maxi-
period before the liberalization of the mining sector and the mum likelihood algorithm provides easy-to-interpret results,
promulgation of the forestry code (1989 and 1998) and the especially when the occupancy types are well defined and
period after (2004–2022). The period after 2014 also cor- have distinct spectral signatures (Hassan et al. 2016; Jiménez
responds to when the City of Kasenga was elevated as a et al. 2018). Also, it can be applied efficiently to large sets of
rural municipality. images considering its adaptation to multispectral imaging
(Payne et al. 2018). According to the objective of the study,
Classification the following land cover classes were selected: built-up and
bare land (bare land, soil background, residential land mini-
The Landsat images used in this study are not annual com- mally vegetated with impervious surface), water (water bod-
posites, but single images that have received essential ies), grassland and wetlands (floodplain grassland and grassy
pre-processing, including georeferencing. Indeed, the pre- vegetation), agriculture (parcels cultivated or put in rest to be
processing consisted first in georeferencing the 1989 Land- cleared after a few years in a crop rotation system), savanna
sat image from the 2022 Landsat image taken as reference (wooded and shrubby savanna) and forest (mosaic of multi-
for the orthorectification. This operation was carried out ple forests types: dry and riparian evergreen forest, but with
using the geographic coordinates of the fixed points, with a miombo woodland remaining dominant). For the validation
geometric accuracy of the setting of 1 pixel, the minimum of classifications obtained, in accordance with good practice
recommendations, unbiazed area estimators and estimation as the ratio of the cumulative area of anthropogenic land
of uncertainty were determined from a sample of reference cover (agriculture, built-up and bare land, savanna, grassland
observations of land cover and land change (Olofsson et al. and wetland) in the landscape and the cumulative area of
2014). This implies the use of a statistical approach which natural classes (forest), was calculated. It should be noted
accounts for rare classes and provides “misclassification cor- that the term “anthropogenic land cover” used in this study
rected” estimates of classes area (Olofsson et al. 2013). The designates any entity of the landscape where direct human
stratified random sampling technique was used to generate alteration is remarkable and where ecological patterns and
error matrix using reference points distributed within each processes are significantly disturbed. This index quantified
thematic class (Churches et al. 2014). Eleven strata were the level of anthropization of the landscape, a generic term
selected, of which 6 were for the stable classes (“built and describing the influence of human activities on the environ-
bare soil”, “water”, “savanna”, “grassland and wetlands”, ment (Useni 2019). Furthermore, spatio-temporal changes
“forest”, “agriculture”) and 5 for the most relevant classes in the landscape pattern were quantified using transition
of change, except for the “water” class due to its relative matrices from which a stability index, defined as the ratio
stability over time. The number of points allocated to each between stable areas and the sum of areas lost and gained
stratum depended on its proportion in the landscape. Thus, by a land cover, was calculated (Bogaert et al. 2014). The
400–500 points were allocated to strata occupying a propor- ecological processes underlying the observed landscape
tion between 21 and 30%, 300–400 points to those occupy- dynamics were qualified using the decision tree of Bogaert
ing between 11 and 20%, 200–300 points to strata occu- et al. (2004) (Fig. 2). To determine the spatial transformation
pying between 1 and 10%, and 100–200 points for strata process, the model uses comparisons between the number of
occupying less than 1% of the landscape. A total of 2173 patches, area and perimeter of patch type before and after the
points were sampled for the period 1989–1998, 2201 points landscape transformation (Bogaert et al. 2004; Barima et al.
for 1998–2004, 2200 points for 2004–2009, 2199 points for 2009; Diallo et al. 2011). Ten spatial transformation pro-
2009–2014 and 2199 points for 2014–2022. Estimates of cesses were defined, namely aggregation, attrition, creation,
area and area change were adjusted to account for biases in deformation, dissection, enlargement, fragmentation, perfo-
the change map, resulting in more accurate estimates. The ration, displacement and shrinkage (Bogaert et al. 2004). In
precision of the estimates of area and changes in area was the case of anthropogenic landscape dynamics, natural patch
then quantified by calculating confidence intervals (Olofs- types will be characterized by attrition, deformation, dis-
son et al. 2014). The Eqs. (1)–(3) of Olofsson et al. (2014) section, fragmentation, perforation and/or shrinkage. Patch
were used to calculate the precision measurements, which types reflecting anthropogenic activities display aggregation,
represent, respectively, the overall precision, the user preci- creation, deformation, enlargement and/or displacement
sion and the producer precision. An “error-adjusted” area (Bogaert et al. 2011). The value of t = 0.75 was used to dis-
estimate of each class and the standard error of the error- sociate the process of fragmentation from dissection, with
adjusted area estimate (calculated using a 95% confidence values above 0.75 suggesting dissection, while those below
interval which is obtained by multiplying the standard error or equal to 0.75 indicated the prevalence of fragmentation
by 1.96) using Eqs. (10) and (11) of Olofsson et al. (2014) (De Haulleville et al. 2018).
(Table 1). The image processing operations were carried out
using ENVI 5.3 software, while ArcGIS 10.7 software was
used for vectorization and map dressing. Results
Quantifying the spatio‑temporal pattern change Land cover mapping and accuracy
within the rural and boarder municipality
of Kasenga The results of accuracy assessment and area estimate for
land cover and land cover change maps from 1989 to 2022
The human impact on landscape morphology was quanti- are presented in the confusion matrix analysis (Table 1).
fied through landscape metrics, namely the patch number, Based on overall accuracy, theses matrices revealed overall
the class area and the largest patch index (LPI) defined as accuracy values ranging from 92 to 93%. Furthermore, the
the ratio of the area of the largest patch to the class area matrices show also user’s and producer’s accuracies for sta-
(McGarigal and Cushman 2002). Patch number and class ble and changing land cover types.
area provided information on the land cover fragmentation
between two periods. Indeed, an increase in patch number Landscape composition dynamics of the rural
of a land cover class while the area decreases may be due to and border municipality of Kasenga
its fragmentation (Bogaert and Mahamane 2005). In addi-
tion, a disturbance index (U) (O’Neill et al. 1988), defined • Landscape composition dynamics
Table 1 Accuracy assessment and area estimate for land cover and land cover change maps from 1989 to 2022
1989–1998 BBS Water Savanna GW Forest Agriculture BBS loss Savanna gain GW gain Forest loss Agriculture loss
Accuracy measure
Prod. acc. (%) 100 100 90 100 100 97 95 92 80 99 61
User acc. (%) 100 100 100 100 99 99 100 73 95 62 66
Overall acc. (%) 92
Stratified estimators of area ± CI [% of total map area]
Area (%) 0.14 0.89 8.39 9.71 26.26 10.60 6.84 6.74 20.04 4.58 5.82
95% CI 0 0 0.5 0 0.19 0.22 0.27 0.75 0.78 0.55 0.75
1998–2004 BBS Water Savanna GW Forest Agriculture BBS gain Savanna loss GW gain Forest loss Agriculture loss
Accuracy measure
Prod. acc. (%) 100 100 99 100 100 98 81 87 72 77 84
User acc. (%) 100 100 98 99 97 100 100 99 54 54 100
Overall acc. (%) 93
Area (%) 0.94 0.74 8.64 16.96 18.92 14.98 0.30 15.02 6.77 2.69 14.05
95% CI 0 0 0.2 0 0.39 0.21 0.11 0.53 0.68 0.46 0.51
2004–2009 BBS Water Savanna GW Forest Agriculture BBS gain Savanna loss GW loss Forest loss Agriculture gain
Accuracy measure
Prod. acc. (%) 100 100 100 99 99 100 100 99 74 90 88
User acc. (%) 96 100 96 100 100 98 88 89 100 100 83
Overall acc. (%) 92
Area (%) 0.31 0.73 4.74 23.99 15.68 7.64 6.30 11.72 6.81 10.88 11.21
95% CI 0 0 0.1 0 0.14 0.17 0.33 0.52 0.34 0.43 0.75
2009–2014 BBS Water Savanna GW Forest Agriculture BBS gain Savanna gain GW loss Forest loss Agriculture gain
Accuracy measure
Prod. acc. (%) 100 100 100 99 100 100 100 71 89 74 93
User acc. (%) 100 100 94 100 100 81 92 93 100 90 100
Overall acc. (%) 93
Area (%) 0.51 1.27 13.57 16.02 15.86 17.68 1.95 5.95 5.26 13.81 8.11
95% CI 0 0 0.4 0 0.00 0.82 0.10 0.63 0.31 0.93 0.34
2014–2022 BBS Water Savanna GW Forest Agriculture BBS gain Savanna loss GW loss Forest loss Agriculture gain
Accuracy measure
Prod. acc. (%) 100 97 94 99 98 100 100 79 79 88 70
User acc. (%) 99 100 85 100 100 94 98 82 78 98 78
Overall acc. (%) 92
Area (%) 0.55 1.01 5.75 13.68 9.83 28.83 2.35 12.54 8.15 10.81 6.50
95% CI 0 0 0.4 0 0.21 0.68 0.06 0.95 0.94 0.44 0.65
GW grassland and wetland, BBS built-up and bare soil
In the rural and border municipality of Kasenga, the land- significantly, rising from 23.75% in 1989 to 52.16% (2022)
scape matrix once dominated by forests has been replaced at the expense of forest cover. Indeed, forest cover has also
by an agricultural matrix, occupying more than half of the declined to the benefit of grassland and savanna (but with a
landscape. Forest, having covered almost half the landscape very low percentage), rising from 18.34 to 19.49% and from
in 1989 (48.22%), has been divided by almost 3 in 33 years 8.45 to 11.16%, respectively, between 1989 and 2022. As for
(15.36%). On the other hand, agriculture land has increased built-up and bare land, although only slightly represented in
Fig. 2 Decision tree to identify

transformation processes that
alter the spatial pattern of
landscapes. The parameters a0,
p0, and n0 refer to the habitat
area, perimeter and number of
patches before transformation,
while a1, p1 and n1 are the
reciprocal values after pattern
change (Bogaert et al. 2004)
the landscape, its proportion has almost doubled, rising from occupied by grassland and wetland was converted to savanna
0.34 to 0.82% (Table 2). between 1989 and 1998, compared to 5. 86% between
2014and 2022.
• Land cover transfer within the rural and border munici- Landscape dynamics are also characterized by the conver-
pality of Kasenga from 1989 to 2022 sion of 4.96% occupied by agriculture to savanna between
1989 and 1998. This transfer rate from agriculture to
savanna has fallen by half, moving to 2.22% between 2014
The regression of the forest is explained by the transfer of and 2022. Furthermore, while 7.11% of the landscape occu-
its area to savanna, at varying rates depending on the period pied by agriculture evolved toward grassland and wetland
(5.95% between 1989–1998; 5.53% between 2014–2022). between 1989 and 1998, this rate had fallen significantly
Conversely, a relative reconstitution of forest cover has been to 1.63% between 2014 and 2022. Finally, the transition
observed, mainly at the expense of the savanna. Moreover, matrices reveal that the expansion of the built-up and bare
the spatial dynamics of the savanna is characterized by an land occurs at the expense of the agriculture located on the
evolution toward agriculture. The regression of savanna city’s outskirts. Indeed, 0.15% and 0. 34% of the landscape
cover in the landscape is also explained by their conversion occupied by agriculture were invaded by built-up and bare
to grassland and wetland (Table 2). In addition, Table 2 also land, respectively, between 1989 and 1998 and 2014 and
indicates a change in landscape composition in the range 2022 (Table 2).
of 4.37 to 18.89%, about the shift from grassland and wet- The analysis of the stability of the land cover shows that
land to agriculture. Furthermore, 1.87% of the landscape forest was the most stable land cover between 1989 and
Table 2 Land cover area transition matrices in the rural and border defined as the ratio between stable areas and the sum of areas lost and
municipality of Kasenga (%) between 1989 and 1998, 1998 and 2004, gained by a land cover (Bogaert et al. 2014)
2004 and 2009, 2009 and 2014, and 2014 and 2022. Stability index is
1989–1998
BBS Savanna GW Forest Agriculture Total 1989
BBS 0.10 0.00 0.16 0.00 0.09 0.34

Savanna 0.00 2.99 2.49 1.82 1.15 8.45
GW 0.08 1.87 9.62 1.61 4.37 18.34
Forest 0.00 5.95 8.36 32.14 1.75 48.22
Agriculture 0.15 4.96 7.11 0.71 10.42 23.75
Total 1998 0.33 15.78 27.79 36.28 17.78
Stability index 0.20 0.16 0.36 1.59 0.50
1998–2004
BBS 0.19 0.00 0.01 0.00 0.11 0.32

Savanna 0.00 4.43 7.25 3.06 0.96 15.70
GW 0.08 2.27 19.37 1.25 4.81 27.79
Forest 0.01 2.14 10.38 23.36 0.42 36.31
Agriculture 0.05 2.78 9.79 0.33 4.88 17.84
Total 2004 0.34 11.62 47.52 27.99 11.80
Stability index 0.72 0.24 0.53 1.33 0.25
2004–2009
BBS 0.26 0.00 0.00 0.00 0.08 0.35

Savanna 0.01 1.07 5.81 2.79 1.95 11.63
GW 0.05 1.76 30.00 4.44 10.93 47.48
Forest 0.00 4.71 3.59 19.19 0.52 28.00
Agriculture 0.21 0.04 4.66 0.23 6.40 11.80
Total 2009 0.53 7.58 44.07 26.65 19.88
Stability index 0.75 0.06 0.95 1.18 0.34
2009–2014
BBS 0.36 0.00 0.00 0.00 0.17 0.53

Savanna 0.00 1.11 2.54 3.46 0.43 7.55
GW 0.02 5.86 15.41 3.75 18.89 44.04
Forest 0.00 3.35 5.47 15.65 2.24 26.71
Agriculture 0.24 0.72 1.04 1.38 16.50 19.89
Total 2014 0.62 11.05 24.67 24.26 38.25
Stability index 0.83 0.07 0.41 0.80 0.66
2014–2022
BBS 0.46 0.00 0.00 0.00 0.17 0.64

Savanna 0.00 3.10 1.02 1.94 5.00 11.07
GW 0.01 0.31 13.60 2.58 8.07 24.62
Forest 0.00 5.53 3.03 9.90 5.79 24.26
Agriculture 0.34 2.22 1.63 0.93 33.12 38.26
Total 2022 0.82 11.16 19.49 15.36 52.16
Stability index 0.89 0.19 0.80 0.50 1.37
Values in bold refer to the proportion of land cover that remained unchanged over the period. In addition, the total of the year in column cor-
responds to the sum of the values of classes in rows. However, the total of the year in row relates to the sum of the values of classes in columns.
For a total area of 783 km2, 1% corresponds to 7.83 km2. The totals do not add up to 100% as the water land cover type was excluded from the
analyses due to its relative stability in the landscape
Table 2 (continued)
GW grassland and wetland, BBS built-up and bare soil
2009. A first transition was noted and concerns the appear- In addition, the reduction in the patch number of savanna
ance of the built-up and bare land as the most stable land and agriculture, in parallel with the increase in class area,
cover between 2009 and 2014, to the detriment of savanna. attests to the aggregation of patches as a spatial transforma-
Between 2014 and 2022, a new transition, materialized by tion process. As for the built-up and bare soil, a simultane-
the appearance of agriculture as the most stable land cover, ous increase in the class area and patch number was noted,
to the detriment of savanna, was recorded (Table 2). suggesting the creation of patches as a spatial transformation
process.
Configurational dynamics of land cover in the rural For the period between 2014 and 2022, the attrition of
and border municipality of Kasenga patches persisted as a spatial transformation process of the
forest, as well as savanna, was characterized by the decrease
The results in Fig. 3 indicate that between 1989 and 1998, in the class area associated with the reduction in the patch
the forest recorded the fragmentation of patches (t = 0.75) number. In addition, grassland and wetland recorded dissec-
because the decrease in the class area was accompanied tion of patches as a process of spatial transformation, illus-
by an increase in patch number. During the same period, it trated by the decrease of the class area in parallel with the
was noted, on the one hand, the aggregation of the savanna increase of the patch number (t = 0.79). Finally, agriculture
patches materialized by an increase in the class area paral- and built-up and bare land recorded an increase in the class
lel to the decrease in the patch number. On the other hand, area in parallel with the decrease in the patch number, thus
the process of spatial transformation characteristic of the suggesting aggregation as a spatial transformation process.
agriculture as well as of the built-up and bare land was the
attrition of patches, especially since a decrease in the patch Quantification of landscape anthropization
number and class area was noted. For grassland and wetland,
the creation process was observed, as the increase class area There was a regression in dominance values for forest over
was accompanied by an increase in the number of patches. all time periods studied, from 92.59% in 1989 to 21.42% in
In addition, the period of 1998–2004 was characterized 2022 (Fig. 4). This suggests that the area of large patches
by the dissection of forest patches (t = 0.77) against the frag- of forest that dominate the landscape has been reduced by a
mentation for agriculture (t = 0.66) and savanna (t = 0.74), factor of 4 in 33 years (from 1989 to 2022), indicating the
because the decrease in the class area was accompanied by anthropogenic impact on the fragmentation of forest ecosys-
an increase in patch number. Grassland and wetland recorded tems. Overall, our results confirm a trend toward an increase
an increase in class area in parallel with the decrease in patch in the level of anthropization of natural landscapes, espe-
number, suggesting aggregation as a spatial transformation cially since the level of landscape disturbance has quintupled
process. Finally, the creation was the process of spatial trans- from 1.1 to 5.5 in 33 years (Fig. 4).
formation characteristic of the of the built-up and bare land,
especially since an increase in the patch number and class Land cover mapping
area was noted.
On the other hand, between 2004 and 2009, dissection The land cover mapping in 1989, 1998, 2004, 2009, 2014
of patches was the process of spatial transformation for the and 2022 is presented in Fig. 5. The visual analysis shows a
forest and grassland and wetland, as the decrease in class regression of forest cover in parallel with an increase in agri-
area is accompanied by an increase in patch number (t = 0.95 culture, as well as in built-up and bare land. The water land
for the forest and 0.93 for the grassland and wetland). In cover was excluded from the rest of the analyses because of
addition, the attrition of patches of savanna was the spatial its relative stability in the landscape.
transformation process, since a decrease in the patch number
and class area was identified. Finally, agriculture, as well as
built-up and bare land, recorded creation as a spatial trans- Discussion
formation process, especially since the increase in class area
was due to the patch number. Methodological approach
Between 2009 and 2014, the attrition of patches was
identified as a spatial transformation process of the forest The use of Landsat images allowed us to map and quan-
and grassland and wetland because the decrease in the class tify the landscape dynamics of Kasenga municipality
area was associated with the reduction in the patch number. over 33 years (1989–2022). Indeed, among the tools for
Fig. 3 Identification of the land a

covers spatial transformation
1.0
Relative value of spatial configuration indices

processes between 1989 and
1998, 1998 and 2004, 2004 0.8
and 2009, 2009 and 2014, and
2014 and 2022 by the decision 0.6
tree algorithm of Bogaert et al.
0.4
(2004) based on the evolution of
the class area and patch number 0.2
of forest (a), savanna (b),
grassland and wetland (c), fields 0.0
1989 1994 1999 2004 2009 2014 2019 2024
and fallows (d) and built-up and
bare land (e) in the rural and Year
border municipality of Kasenga
for the years 1989, 1998, 2004, b
2009, 2014 and 2022. The rela- 1.0
tive values (of class area and 0.8
patch number) were obtained
from the ratio of the index value 0.6
for each land cover type to the 0.4
maximum value recorded. Over- 0.2
all, the forest is characterized by
a simultaneous decrease in class 0.0
1989 1994 1999 2004 2009 2014 2019 2024
area and patch number
Year
c
1.0
0.8
0.6
0.4
0.2
0.0
1989 1994 1999 2004 2009 2014 2019 2024
Year
d
1.0
0.8
0.6
0.4
0.2
0.0
1989 1994 1999 2004 2009 2014 2019 2024
Year
e
1.0
0.8
0.6
0.4
0.2
0.0
1989 1994 1999 2004 2009 2014 2019 2024
Year
Class area Patch number
Fig. 4 Evolution of the domi- 100 1989

nance of forest patches (%) as
Largest pacth index of the forest (%)

a function of the landscape
disturbance index in the rural 80
and border municipality of 1998
Kasenga, southeast of the City
of Lubumbashi, based on Land- 60
sat image processing data from
1989, 1998, 2004, 2009, 2014
and 2022
40
2004
2022
2009 2014
20
0
1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6
Landscape disturbance index (U)
Fig. 5 Land cover maps of the rural and border municipality of Kasenga in 1989, 1998, 2004, 2009, 2014 and 2022 based on supervised clas-
sification of Landsat images using the maximum likelihood algorithm. A regression of the forest is visually remarkable between 1989 and 2022
monitoring the dynamics of spatio-temporal phenomena, accuracy, user’s and producer’s accuracies (Lasko et al.
remote sensing is the most suitable (Bamba et al. 2008; 2005; Liu et al. 2007; Rwanga and Ndambuki 2017) show
Mama et al. 2013). Furthermore, values from confusion good agreement between classifications and the reference
matrices for supervised classification through overall data collected during the field missions. Therefore, many
time steps between our Landsat images did not prevent the the residual patches are presented in small, very isolated
study to highlights the main trends of landscape evolution fragments (Sadda et al. 2016). These landscape dynamics
in the study area, and our results are very similar to those are mainly due to the intensification of collection activi-
of previous studies in the region, which used the same ties of (non-)wood forest products to meet food, medicine
approach (Useni et al. 2018a, b; Munyemba et al. 2008) or and energy needs (Useni et al. 2017; Kabulu et al. 2018;
local population perception (Amisi et al. 2018; N’Tambwe Useni et al. 2018a, b; Khoji et al. 2022; Mpanda et al.
Nghonda et al. 2023). Furthermore, the rapid transitions 2022; Mutombo et al. 2022). Construction materials are
of the observed land cover types are related to the intense also collected from forests where population growth and
anthropogenic activities in the study area, most often shift- poverty amplify dependence on forest ecosystems (Ndenge
ing cultivation, charcoal production and human-induced and Perfet-Mrema 2022) and anthropogenic pressure on
vegetation burning. These activities are recurrent and make them. In addition, studies have shown that high deforesta-
the different land cover types very unstable. Based on the tion rates in the miombo ecoregion are primarily related
overall accuracies, values obtained show the success of the to slash-and-burn agriculture (Jew et al. 2017; Rannestad
supervised classification, which would be mainly attribut- and Gessesse 2020) in a context of fragile soil fertility
able to the knowledge of the study area and quality of images due to high rainfall and temperature (De Hipt et al. 2018),
after several pre-processing, which made it possible to define thus justifying the expansion of agriculture in our study. It
the reliable training plots for all classes. should be noted that research conducted by Mpanda et al.
Furthermore, the calculation of landscape metrics, par- (2021) in the same area has shown that the local popula-
ticularly those related to the area and number of patches, tion lives mainly from agriculture, an agriculture that does
is reliable for appreciating the spatial pattern changes of not always comply with ecological standards of sustain-
landscape (Bogaert et al. 2014) within the municipality of ability. Our results corroborate the findings of Handavu
Kasenga, resulting from various anthropogenic pressures et al. (2019) that household size significantly influences
(Mpanda et al. 2022). Indeed, the anthropization of a natu- agricultural expansion in rural Zambia and, consequently,
ral environment is generally characterized by fragmentation land cover changes. In addition, another cause of deforest-
and a progressive dynamic of anthropogenic land cover at ation within the rural and border municipality of Kasenga
the expense of natural land cover types (Bogaert et al. 2011), is the extraction of wood fuel, generally leading to a
which is confirmed by our results. The reduced number of savanna expansion process as observed around Lubum-
landscape metrics calculated diminished redundancy, as sev- bashi City (Useni et al. 2017). It should be noted that
eral landscape metrics are correlated (Mama et al. 2013). when forest resources are available, large-diameter trees
Although the time interval between the images used may are cut for charcoal production (Mama et al. 2014), since
seem relatively short, the transition matrix was based on charcoal production is the most important strategy used to
the change/stability of intensity over time of the same pixel cope with food shortages (Kalaba et al. 2013). In addition,
(Mama et al. 2013). Moreover, the same training areas were changes in logging practices are being highlighted through
used, regardless of the date of the Landsat image, supported the shift from selective to non-selective logging due to
by the with resource persons (Mpanda et al. 2022). the unavailability of large timber used for charcoal pro-
duction within the Lufira Biosphere Reserve near Lubum-
Agricultural development, savanna expansion, bashi City (Useni et al. 2020b). As a result, frequent tree
(peri)urbanization and deforestation in the rural cutting results in a landscape dominated by scattered,
and border municipality of Kasenga small-diameter trees, giving the initial forest patches the
appearance of savannas (Useni et al. 2020a, b). Malaisse
In the rural and border municipality of Kasenga, except (2010) points out that savanna is anthropogenic in origin
potential natural drivers of landscape dynamics, the in the region, which corroborates our results showing an
dependence of local communities on forest resources has expansion of savanna around Kasenga municipality with
led to remarkable changes in their properties between 1989 increasing levels of landscape disturbance. Furthermore,
and 2022. Indeed, repeated human intrusion into the for- in Central Africa, bushfire, recognized as a widespread
est around Kasenga municipality has led to a continuous technique for clearing dense vegetation cover and prepar-
decline in their cover. In the absence of high spatial reso- ing land for cultivation, potentially leads to savanna expan-
lution data to characterize the land cover change in the sion, as the case around Kasenga municipality (Useni et al.
municipality of Kasenga, the use of Landsat data made it 2023), through changes in species composition, biomass
possible to understand and interpret the landscape dynam- and nutrient levels (Bundschuh et al. 2010; Mganga 2022).
ics over a period of 33 years. In addition, over time, for- Indeed, to prepare the fields and hunting activities in the
est ecosystems are converted, in the absence of a sustain- village around Kasenga City, vegetation is burned almost
able land use policy, into anthropogenic land cover, and twice in the dry season (Malaisse 2010).
In addition, deforestation rates reported in this study far anthropogenic activities. Savannahs with significant wood-
exceed those estimated annually at the national level (0.2%; land, tree and shrub cover when abandoned show an increase
Kyale et al. 2019) and in the Katanga clear forest (0.6–1.8%) in the density of woodland pockets. The continuous reduc-
(Defourny et al. 2011; Kabulu et al. 2018). This is a result of tion in the abundance of the tree stratum allows light pene-
the proximity of Kasenga municipality to Lubumbashi City, tration, which is favorable to the development of herbaceous
where demand for agricultural products, energy, and phar- species (Rakotondrasoa et al. 2013) and most of which result
macopeia is constantly increasing (Useni et al. 2017). Ideas in the formation of grassy savanna, thus justifying the trend
agree that uncontrolled exploitation of forest resources and toward the progression of grasslands in the landscapes of
resulting deforestation remain largely associated with the Kasenga municipality. In addition, the built-up and bare land
expansion of subsistence activities (agriculture and energy) records the merging of patches, resulting on the one hand
and are indeed more evident around major cities (Bamba from the process of building expansion within residential
et al. 2010; Mwitwa et al. 2012). However, miombo plant plots to absorb new housing demands (Useni et al. 2018a,
species used for their fruits or their wood are not maintained b) from students and civil servants posted to Kasenga. On
in agro-ecosystems since there is therefore a separation the other hand, the merging of built and bare soil can also
between agriculture and forest. As a result, there is a lack be explained by the parceling of plots by indigenous popula-
of management of tree in agriculture system in the rural tions. In the context of poverty, the indigenous populations
area adjacent to Lubumbashi City (Hick et al. 2018). Yet, sell part of their land to the new arrivals to finalize the con-
the depletion of forest patches could have many negative struction of their own houses (Groupe Huit 2009). Finally,
socio-ecological consequences, among them the gradual in the Kasenga housing estate, local land registrars relegate
disappearance of many non-timber resources, gathering excess housing demands to the outskirts of the municipal-
products such as honey, mushrooms, but also caterpillars ity on agricultural land. Indeed, in these agricultural areas
due to selective cutting of host plants, etc. (Bogaert et al. located to the west of the municipality, the prices of plots,
2011; Maseko et al. 2017). This deforestation also reduces which are otherwise large (at least 900 m2), are relatively
the number of rainy days (Leite-Filho et al. 2021). Also, the affordable and easy to develop (Useni et al. 2020a, b). Bren
consequences related to deforestation are experienced with d’Amour et al. (2017) point out that the planned expansion
severity in rural areas where the population survives mainly of built-up will take place on some of the world's most pro-
on forest resources but where paradoxically management ductive cropland, particularly in Asia and Africa. As in our
measures are still precarious (Mpanda et al. 2022). Conse- study, Lasisi et al. (2017) recorded a loss of agricultural
quently, our results suggest a trend toward the fragmentation land of about 331 ha/year between 1986 and 2014 due to
and isolation of forest patches through the spatial transfor- unplanned urban expansion in the peri-urban areas of Olo-
mation processes of fragmentation and dissection, which runda and Osogbo in Ghana. However, the transition from
may limit the exchange of propagules between species and built-up areas to fields, illustrated by the transition matrices,
negatively impact their viability (Cristofoli and Mahy 2010). is explained by the influence of (urban) agriculture, which
Indeed, in the mountains of the Eastern Arc of Tanzania, results in the cultivation of spaces that could previously be
it has been noted that the abundance of dominant species identified as bare soil. Nevertheless, the similarity of the
(such as Maytenus undata Thunb, Zenkerella capparidacea spectral signatures of bare soils and harvested fields would
(Taub.) J. Leon, and Oxyanthus specious DC.) decreases be a second reason to explain this observed transition (Con-
with a reduction in patch area (Ojoyi et al. 2015). gedo and Munafo 2012).
In addition, savanna-covered formations remained
prevalent, but were nevertheless characterized over time Implications for land use planning
by regression–progression sequences. This recovery of the
savanna would be largely attributable to their strong capacity Unplanned urban growth and extensive agricultural devel-
for reconstitution. This assertion is supported by Syampun- opment are perpetuating the rate of deforestation and caus-
gani et al. (2016), who report rapid development of miombo ing deep concern, especially in an area of high population
recruit stands upon cessation of anthropogenic disturbance density where the shortening of the fallows period no longer
(cultivation and charcoal production). Their regression is allows the forest to recover (Cabala et al. 2017; Khoji et al.
generally explained by the fact that after the disappearance 2022). Indeed, the current pattern of agricultural develop-
of large-diameter species, the medium-diameter species, pre- ment at the expense of forests contributes to their degrada-
sent in savanna are cut and charred (Mama et al. 2014). Fur- tion and reduced abundance in the landscape. As for defor-
thermore, it is noted that wooded, tree and shrub savannahs estation related to agricultural development, the production
are rarely natural in the region and generally result from of maize, a crop that requires more soil fertility (Useni et al.
forest degradation following various anthropogenic activities 2013), is thought to be one of the leading causes. Further-
and therefore their importance increases with the increase of more, the ease of penetration into these different forest
environments by farmers in the region may explain their a holistic approach to the planning process. It is therefore
vulnerability to anthropization. In this context, the public important to produce a concerted master plan for the devel-
technical services should promote models based on agro- opment of the rural and border municipality of Kasenga,
forestry with strong accompaniment of Kasenga farmers, with zoning scrupulously applied as suggested by Angel
to limit the expansion of extensive agriculture identified by et al. (2011). However, there are still several gaps for future
Mpanda et al. (2021) as the agriculture model of this region. research to enable further development of Kasenga border
The results of Amadu et al. (2020) show that agroforestry city. Combining data from multiple sources, including satel-
improves maize yields among smallholder farmers in the lite imagery, ground-based data and socio-economic data,
face of climate change, a crucial aspect of sustainable devel- could provide a more holistic understanding of land cover
opment goals. However, Etshekape et al. (2018) identified change drivers and impacts within Kasenga municipality.
farmers’ education level among the important factors that Indeed, understanding how policy, population growth and
positively influence farmers’ decision to adopt agroforestry. economic development impact land cover changes is a line
Furthermore, our results suggest a shift from selective of research to be explored. However, extending the analysis
logging, particularly in terms of tree diameter, to non- of land cover change over longer time periods, including
selective logging. However, not only are forests with large- historical data and predictive modeling into the future, is
diameter individuals becoming scarcer, but also savanna essential for understanding trends and anticipating potential
that are supposed to replenish forests are being attacked by environmental consequences, especially in the context of
tree-cutting activities for carbonization, confirming the trend climate change and unplanned urbanization. More research
of results obtained by Useni et al. (2020b) in the context of is needed to understand the ecological and environmen-
Lubumbashi City. tal consequences of land cover changes, such as impacts
Agroforestry with tree plantations of species native to on biodiversity, water resources, and carbon sequestration.
the miombo woodland, needed for producing good quality Addressing these gaps will contribute to a more comprehen-
charcoal, should be established. This would allow for the sive understanding of land cover changes, their drivers and
production of maize and charcoal on the same land. On the their impacts and can help inform better land use planning
Batéké plateau in Kinshasa (the Democratic Republic of and environmental management strategies within Kasenga
the Congo), 8000 ha of Acacia auriculiformis plantations municipality.
provide a total charcoal production of 8000 to 12,000 tons
per year, to which must be added 10,000 tons of cassava,
1200 tons of maize and 6 tons of honey (Bisiaux et al. 2009). Conclusion
Finally, the current pattern of one-way expansion of the
city (to the west) due to the presence of the Luapula River This study highlights the relevance of remote sensing, geo-
as a barrier dangerously compromises the persistence of graphic information systems and landscape ecology analysis
the Lubi Valley, which is crucial for flooded rice produc- tools to map and quantify the spatio-temporal dynamics of
tion. This situation is amplified by the development of the landscape anthropization in the rural and border munici-
Kasenga customs post, leading to an influx of civil serv- pality of Kasenga. Our results confirm that forest ecosys-
ants and their families. This is accompanied by an over- tems undergo surface losses, notably through dissection and
riding need for decent housing, justifying the expansion of fragmentation that removes residual fragments. Indeed, the
housing on marshy land, which is also used for rice and proportion of forest in the landscape has been reduced by
horticulture. This reduces the soil’s water storage capacity almost 3 in 33 years, especially since its area, which covered
(Oiro et al. 2020) and compromises the future of rice culti- 48.22% of the landscape in 1989, has dropped to 15.36% in
vation. In addition, although Kasenga municipality offices 2022. This is the result of the development of extensive agri-
may have urban planning and development technicians, they cultural activities, but also savanization. Savanization and
are outdated due to lack of practical tools and training like agricultural development occur mainly in the city’s vicin-
elsewhere in the ecoregion (Chitonge and Mfune 2015). As ity, which continues to expand its spatial reach. Indeed, our
a result, they become more involved in creating subdivi- results show that the surface area of buildings has doubled
sions and parceling out residential plots, without coherence in the landscape in 33 years, notably to the detriment of rice
or overall plan, to increase their own financial availability and horticultural areas. The progressive dynamics of anthro-
like in Butembo in the eastern DR Congo (Kasereka and pogenic land cover, the proportion of which has doubled
Mate 2018). Kaswamila and Songorwa (2009) findings in 33 years through the creation and merging of patches,
indicate that the main causes of failure of sustainable land has amplified the level of disturbance of the landscape. One
use planning are insufficient stakeholder participation in the of the main causes of the observed landscape dynamics is
planning process, lack of sound, transparent and accountable slash-and-burn agriculture and anarchic urbanization in
implementation strategies, lack of qualified staff and lack of the context of galloping population growth. It is therefore
important to reconcile these results of landscape anthropiza- Barima YSS, Barbier N, Bamba I, Traoré D, Lejoly J, Bogaert J (2009)
tion with the municipality’s agricultural and urban planning Dynamique paysagère en milieu de transition forêt-savane ivoir-
ienne. Bois For Trop. https://doi.org/10.19182/bft2009.299.
policies for sustainable management of the meager forest a20419
resources that remain. Biaou S, Gouwakinnou GN, Houessou GL, Ahononga CF, Idjigberou
SE, Houeto OF (2022) Dynamique spatio-temporelle de la forêt
Acknowledgements We would like to thank the local authorities of classée de l’Alibori-Supérieur et impact sur la biodiversité. In:
the rural municipality of Kasenga and the outlying villages for their Profizi J, Ardila-Chauvet S, Billot C, Couteron P, Delmas M,
support during data collection in the field. The research was funded by Hanh Diep TM, Sonke B (eds) Biodiversité des écosystèmes
the CHARLU project (ARES-CCD, Belgium). intertropicaux: Connaissance, gestion durable et valorisation.
IRD Editions, Marseille, pp 553–564. https://doi.org/10.4000/
Funding This work funded by Académie de recherche et books.irdeditions.40598
d’enseignement supérieur to Héritier Khoji Muteya with PRD Bisiaux F, Peltier R, Muliele JC (2009) Plantations industrielles et
CHARLU. agroforesterie au service des populations des plateaux Batéké,
Mampu, en République Démocratique du Congo. Bois Forets
Trop. https://doi.org/10.19182/bft2009.301.a20404
Bogaert J, André M (2013) L’écologie du paysage : une discipline
unificatrice. Tropicultura 31(1):1–2
References Bogaert J, Mahamane A (2005) Ecologie du paysage : cibler la con-
figuration et l’échelle spatiale. Ann Sci Agron Bénin 7:39–68
Akinnifesi FK, Sileshi G, Ajayi OC, Chirwa PW, Kwesiga F, Harawa R Bogaert J, Ceulemans R, Salvador-Van Eysenrode D (2004) Decision
(2008) Contributions of agroforestry research and development tree algorithm for detection of spatial processes in landscape
to livelihood of smallholder farmers in southern Africa: 2. Fruit, transformation. Environ Manag 33(1):62–73. https://doi.org/10.
medicinal, fuelwood and fodder tree systems. Agric J 3:76–88 1007/s00267-003-0027-0
Amadu FO, Miller DC, McNamara PE (2020) Agroforestry as a Bogaert J, Farina A, Ceulemans R (2005) Entropy increase of frag-
pathway to agricultural yield impacts in climate-smart agricul- mented habitats: a sign of human impact? Ecol Indic 5(3):207–
ture investments: evidence from Southern Malawi. Ecol Econ 212. https://doi.org/10.1016/j.ecolind.2005.02.002
167:106443. https://doi.org/10.1016/j.ecolecon.2019.106443 Bogaert J, Barima YSS, Iyongo WML, Bamba I, Mama A, Toyi M,
Amarnath G, Babar S, Murthy MSR (2017) Evaluating MODIS-vege- Lafortezza R (2011) Forest fragmentation: causes, ecological
tation continuous field products to assess tree cover change and impacts and implications for landscape management. In: Li C,
forest fragmentation in India: a multi-scale satellite remote sens- Lafortezza R, Chen J (eds) Landscape ecology and forest man-
ing approach. Egypt J Remote Sens Space Sci 20(2):157–168 agement: challenges and solutions in a changing globe. Higher
Amisi MY, Vranken I, Nkulu J, Lubala Toto RF, Kyanika D, Tshibang Education Press/Springer, Beijing, pp 273–296
Nawej S, Mastaki Upite J, Bulambo Mwema JP, Bogaert J (2018) Bogaert J, Vranken I, André M (2014) Anthropogenic effects in land-
L’activité minière au Katanga et la perception de ses impacts à scapes: historical context and spatial pattern. In: Hong SK,
Lubumbashi, Kolwezi, Likasi et Kipushi. In: Bogaert J, Colinet Bogaert J, Min Q (eds) Biocultural landscapes diversity, func-
G, Mahy G (eds) Anthropization des paysages Katangais. Les tions and values. Springer, Dordrecht, pp 89–112
presses universitaires de Liège, Liège, pp 267–279 Bren d’Amour C, Reitsma F, Baiocchi G, Barthel S, Güneralp B, Erb
Angel S, Parent J, Civco DL, Blei AM (2011) Making room for a small KH, Haberl H, Creutzig F, Seto KC (2017) Future urban land
planet of cities. Lincoln Institute of Land Policy, Cambridge expansion and implications for global croplands. Proc Natl Acad
Arroyo-Rodriguez V, Fahrig L, Tabarelli M, Watling JI, Tischendorf L, Sci USA 114(34):8939–8944. https://doi.org/10.1073/pnas.
Benchimol M, Cazetta E, Faria D, Leal IR, Melo FPL, Morante- 1606036114
Filho JC, Santos BA, Arasa-Gisbert R, Arce-Peña N, Cervantes- Bundschuh TV, Wittig R, Hahn K (2010) Effects of human impact
López MJ, Cudney-Valenzuela S, Galán-Acedo C, San-José M, on miombo woodland in Northern Malawi. Flora et Vegetation
Vieira ICG, Slik JWK, Nowakowski AJ, Tscharntke T (2020) Sudano-Sambesica 13:22–34
Designing optimal human-modified landscapes for forest biodi- Byomkesh T, Nakagoshi N, Dewan AM (2012) Urbanization and green
versity conservation. Ecol Lett 23(9):1404–1420 space dynamics in greater Dhaka, Bangladesh. Landsc Ecol Eng
Assale AAY, Barima YSS, Kouakou KA, Kouakou ATM, Bogaert 8(1):45–58. https://doi.org/10.1007/s11355-010-0147-7
J (2016) Agents de dégradation d’une aire protégée après une Cabala KS, Useni SY, Sambieni KR, Bogaert J, Munyemba KF (2017)
décennie de conflits en Côte d’Ivoire: cas de la forêt classée du Dynamique des écosystèmes forestiers de l’Arc Cuprifère Katan-
Haut-Sassandra. Int J Innov Sci Res 22(1):123–133 gais en République Démocratique du Congo. I. Causes, transfor-
Assani AA (1999) Variabilité temporelle et persistance des épisodes mations spatiales et ampleur. Tropicultura 35:192–202
secs en saison des pluies à Lubumbashi (Congo-Kinshasa). CAID (2017) Données par province administrative. Kinshasa, Gombe.
Sécheresse 10(1):45–53 https://www.caid.cd. Consulté le. Accessed 06 Oct 2022
Bakó G, Tolnai M, Taká Á (2014) Introduction and testing of a moni- Caloz R, Collet C (2001) Précis de télédétection: traitements numé-
toring and colony-mapping method for waterbird populations that riques d’images de télédétection, vol 3. Presses de l’Université
uses high-speed and ultra-detailed aerial remote sensing. Sensors du Québec, Québec
14(7):12828–12846 Chinomona RR, Gandiwa E, Muposhi VK, Muboko N, Moyo MS
Bamba I, Mama A, Neuba DFR, Koffi KJ, Traore D, Visser M, Sinsin (2018) Forage preference of the greater kudu (Tragelaphus
B, Lejoly J, Bogaert J (2008) Influence des actions anthropiques strepsiceros) in a miombo woodland adjacent to Umfurudzi Park,
sur la dynamique spatio-temporelle de l’occupation du sol dans Zimbabwe. Cogent Environ Sci 4(1):1559461. https://d oi.o rg/1 0.
la province du Bas Congo (République Démocratique du Congo). 1080/23311843.2018.1559461
Sci Nat 5:49–60 Chitonge H, Mfune O (2015) The urban land question in Africa: the
Bamba I, Bogaert J, Yedmel MSC (2010) Effets des routes et des villes case of urban land conflicts in the city of Lusaka, 100 years after
sur la forêt dense dans la province orientale de la République its founding. Habitat Int 48:209–218. https://doi.org/10.1016/j.
Démocratique du Congo. Eur J Sci Res 43(3):417–429 habitatint.2015.03.012
Churches CE, Wampler PJ, Sun W, Smith AJ (2014) Evaluation of (LULCC) using geospatial techniques: a case study of Islama-
forest cover estimates for Haiti using supervised classification bad Pakistan. Springerplus 5:1–11. https://doi.org/10.1186/
of Landsat data. Int J Appl Earth Obs Geoinf. https://doi.org/ s40064-016-2414-z
10.1016/j.jag.2014.01.020 Hemati M, Hasanlou M, Mahdianpari M, Mohammadimanesh F
Congedo L, Munafò M (2012) Development of a methodology (2021) A systematic review of Landsat data for change detec-
for land cover classification in Dar es Salaam using Landsat tion applications: 50 years of monitoring the earth. Remote
imagery. Sapienza University, Rome. http://www.planning4a Sens 13(15):2869
daptation.eu. Accessed 17 May 2023 Hick A, Hallin M, Tshibungu NA, Mahy G (2018) La place de l’arbre
Cristofoli S, Mahy G (2010) Restauration écologique : contexte, con- dans les systèmes agricoles de la région de Lubumbashi. In:
traintes et indicateurs de suivi. Biotechnol Agron Soc Environ Bogaert J, Colinet G, Mahy G (eds) Anthropization des Pay-
14(1):203–211 sages Katangais. Les Presses Universitaires de Liège, Liège,
De Haulleville T, Rakotondrasoa OL, Rakoto Ratsimba H, Bastin JF, pp 111–126
Brostaux Y, Verheggen FJ, Rajoelison GL, Malaisse F, Ponce- Jew EK, Dougill AJ, Sallu SM (2017) Tobacco cultivation as a driver
let M, Haubruge E, Beeckman H, Bogaert J (2018) Fourteen of land use change and degradation in the miombo woodlands
years of anthropization dynamics in the Uapaca bojeri Baill. of South-West Tanzania. Land Degrad Dev 28(8):2636–2645.
forest of Madagascar. Landsc Ecol Eng 14(1):135–146. https:// https://doi.org/10.1002/ldr.2827
doi.org/10.1007/s11355-017-0340-z Jiménez A, Vilchez F, González O, Flores S (2018) Analysis of the
De Hipt FO, Diekkrueger B, Steup G, Yira Y, Hoffmann T, Rode land use and cover changes in the metropolitan area of Tepic-
M (2018) Modeling the impact of climate change on water Xalisco (1973–2015) through landsat images. Sustainability
resources and soil erosion in a tropical catchment in Burkina 10(6):1860. https://doi.org/10.3390/su10061860
Faso, West Africa. CATENA 163:63–77. https://doi.org/10. Kabulu DJ-P, Vranken I, Bastin JF, Malaisse F, Nyembwe S, Useni
1016/j.catena.2017.11.023 SY, Ngongo LM, Bogaert J (2018) Approvisionnement en
Defourny J-P, Delhage C, Kibambe JP (2011) Analyse quantitative charbon de bois des ménages lushois: quantités, alternatives
des causes de la déforestation et de la dégradation des forêts et conséquences. In: Bogaert J, Colinet G, Mahy G (eds)
en République démocratique du Congo. Rapport, Food and Anthropization des Paysages Katangais. Les Presses Univer-
Agriculture Organisation, Kinshasa, République Démocratique sitaires de Liège, Liège, pp 297–311
du Congo Kalaba FK, Quinn CH, Dougill AJ (2013) The role of forest provi-
Dewees PA, Campbell BM, Katerere Y, Sitoe A, Cunningham AB, sioning ecosystem services in coping with household stresses
Angelsen A, Wunder S (2010) Managing the miombo wood- and shocks in miombo woodlands, Zambia. Ecosyst Serv
lands of Southern Africa: policies, incentives and options for 5:143–148. https://doi.org/10.1016/j.ecoser.2013.07.008
the rural poor. J Nat Resour Policy Res 2:57–73. https://doi. Kalombo KD (2016) Evaluation des éléments du climat en R.D.C.
org/10.1080/19390450903350846 Editions Universitaires Européennes, Saarbrücken (Allemagne)
Diallo H, Bamba I, Barima YSS, Visser M, Ballo A, Mama A, Karsenty A (2020) Géopolitique des forêts d’Afrique centrale. Héro-
Bogaert J (2011) Effets combinés du climat et des pressions dote 172(4):108–129
anthropiques sur la dynamique évolutive de la végétation d’une Kasereka KK, Mate JP (2018) Gouvernance environnementale de
zone protégée du Mali (Réserve de Fina, Boucle du Baoulé). la ville de Butembo par les services publics urbains (Nord-
Secheresse 22(2):97–107. https://doi.org/10.1684/sec.2011. Kivu, République Démocratique du Congo). Tropicultura
0306 36(3):578–592
Engelen V, Verdoodt A, Dijkshoorn K, Van Ranst E (2006) Soil and Kasongo LM (2008) Système d’évaluation des terres à multiples
terrain database of Central Africa-DR Congo, Burundi and échelles pour la détermination de l’impact de la gestion agri-
Rwanda. SOTERCAF, Version 1.0.ISRIC-UGent-FAO cole sur la sécurité alimentaire au Katanga, R. D. Congo. PhD
Etshekape PG, Atangana AR, Khasa DP (2018) Tree planting in urban Thesis. Université de Ghent (Belgique)
and peri-urban of Kinshasa: survey of factors facilitating agro- Kaswamila AL, Songorwa AN (2009) Participatory land-use plan-
forestry adoption. Urban for Urban Green 30:12–23. https://doi. ning and conservation in northern Tanzania rangelands. Afr J
org/10.1016/j.ufug.2017.12.015 Ecol 47:128–134. https://doi.org/10.1111/j.1365-2028.2008.
FAO (2010) Evaluation des Ressources forestières mondiales 2010: 01059.x
Rapport principal. Food and Agriculture Organization of the Keenan RJ, Reams GA, Achard F, de Freitas JV, Grainger A, Lindquist
United States, Rome, Italy. https://w ww.f ao.o rg/3/i 1757e/i 1757e. E (2015) Dynamics of global forest area: results from the FAO
pdf Global Forest Ressources Assessment 2015. For Ecol Manag
FAO (2020) Global Forest Ressources Assessment 2020: main report. 3652:9–20. https://doi.org/10.1016/j.foreco.2015.06.014
Rome. https://doi.org/10.4060/ca9825en Khoji MH, N’Tambwe NDD, Malaisse F, Waselin S, Sambiéni KR,
Groupe Huit (2009) Elaboration du plan urbain de référence de Lubum- Cabala KS, Useni SY (2022) Quantification and simulation of
bashi. Rapport final, Groupe Huit, Lubumbashi, République landscape anthropization around the mining agglomerations of
Démocratique du Congo Southeastern Katanga (DR Congo) between 1979 and 2090.
Handavu F, Chirwa PW, Syampungani S (2019) Socio-economic fac- Land 11:850. https://doi.org/10.3390/land11060850
tors influencing land-use and land-cover changes in the miombo Kruse FA (2004) Comparison of ATREM, ACORN, and FLAASH
woodlands of the Copperbelt province in Zambia. For Pol Econ atmospheric corrections using low-altitude AVIRIS data of Boul-
100:75–94. https://doi.org/10.1016/j.forpol.2018.10.010 der, Colorado. In: Proceedings of 13th JPL airborne geoscience
Hansen MC, Potapov PV, Moore R, Hancher M, Turubanova SA, workshop (Pasadena, CA, USA, 31 March–2 April 2004), pp
Tyukavina A, Thau D, Stehman SV, Goetz SJ, Loveland TR, 1–10
Komareddy A, Egorov A, Chini L, Justice CO, Townshend J Kyale KJ, Wardell DA, Mikwa JF, Kabuanga JM, Monga Ngonga AM,
(2013) High-resolution global maps of 21st-century forest cover Oszwald J, Doumenge C (2019) Dynamique de la déforestation
change. Science 342(6160):850–853. https://doi.org/10.1126/ dans la réserve de biosphère de Yangambi (République démocra-
science.1244693 tique du Congo) variabilité spatiale et temporelle au cours des 30
Hassan Z, Shabbir R, Ahmad SS, Malik AH, Aziz N, Butt A, Erum dernières années. Bois Forêts Trop 341:15–28. https://d oi.o rg/1 0.
S (2016) Dynamics of Land Use and Land Cover Change 19182/bft2019.341.a31752
Lasisi M, Popoola A, Adediji A, Adedeji O, Babalola K (2017) City Mpanda MM, Kipili MI, Tshomba KJ, Kitabala MA, Cabala KS,
expansion and agricultural land loss within the peri-urban area Useni SY (2021) Perception de la dégradation de la fertilité
of Osun state, Nigeria. Ghana J Geogr 9(3):132–163 des sols et de sa gestion par les agriculteurs de la cité de
Lasko TA, Bhagwat JG, Zou KH, Ohno-Machado L (2005) The use of Kasenga en République Démocratique du Congo. Geo-Eco-
receiver operating characteristic curves in biomedical informat- Trop 45(2):211–220
ics. J Biomed Inform 38:404–415 Mpanda MM, Khoji MH, N’Tambwe NDD, Sambieni RK, Malaisse F,
Leite-Filho AT, Soares-Filho BS, Davis JL, Abrahão GM, Börner J Cabala KS, Bogaert J, Useni SY (2022) Uncontrolled exploita-
(2021) Deforestation reduces rainfall and agricultural revenues tion of Pterocarpus tinctorius Welw. and associated landscape
in the Brazilian amazon. Nat Commun 12(1):1–7. https://d oi.o rg/ dynamics in the Katanga territory: case of the rural area of Kas-
10.1038/s41467-021-22840-7 omeno (DR Congo). Land 11:15–41. https://doi.org/10.3390/
Li J, Chen B (2020) Global revisit interval analysis of Landsat 8-9 land11091541
and Sentinel-2A-2B data for terrestrial monitoring. Sensors Munyemba F, Bamba I, Kabulu J-P, Amisi MY, Veroustraete F,
20(22):6631. https://doi.org/10.3390/s20226631 Ngongo M, Bogaert J (2008) Occupation des sols dans le cône
Li R, Wang L, Lu Y (2023) A comparative study on intra-annual clas- de pollution à Lubumbashi. Ann Fac Sci Agron Université de
sification of invasive saltcedar with Landsat 8 and Landsat 9. Int Lubumbashi 1(2):19–25
J Remote Sens 44(6):2093–2114. https://doi.org/10.1080/01431 Mutombo CS, Bakari SA, Ntabaza VN, Nachtergael A, Lumbu JBS,
161.2023.2195673 Duez P, Kahumba JB (2022) Perceptions and use of traditional
Lisa P, Christian L, Karen B, Blake W (2021) Armed conflict and African medicine in Lubumbashi, Haut-Katanga province (DR
cross-border asymmetries in urban development: a contextual- Congo): a cross-sectional study. PLoS One 17(10):e0276325.
ized spatial analysis of Goma, Democratic Republic of the Congo https://doi.org/10.1371/journal.pone.0276325
and Gisenyi, Rwanda. Land Use Policy 109:105711 Mwitwa J, German L, Muimba-Kankolongo A, Puntodewo A (2012)
Liu C, Frazier P, Kumar L (2007) Comparative assessment of the meas- Governance and sustainability challenges in landscapes shaped
ures of thematic classification accuracy. Remote Sens Environ by mining: mining-forestry linkages and impacts in the Copper
107:606–616 Belt of Zambia and the DR Congo. For Policy Econ 25:19–30.
Lusambo LP, Lupala ZJ, Midtgaard F, Ngaga YM, Kessy JF, Abdal- https://doi.org/10.1016/j.forpol.2012.08.001
lah JM, Mombo F, Nyamoga GZ (2016) Increased biomass for N’Tambwe Nghonda DD, Muteya HK, Kashiki BKWN, Sambiéni KR,
carbon stock in participatory forest managed miombo woodlands Malaisse F, Useni SY, Masengo KW, Bogaert J (2023) Towards
of Tanzania. J Ecosyst Ecogr 6(2):1–6 an inclusive approach to forest management: highlight of the
Malaisse F (2010) How to live and survive in Zambezian open forest perception and participation of local communities in the manage-
(Miombo ecoregion). Les Presses Agronomiques de Gembloux, ment of miombo woodlands around Lubumbashi (Haut-Katanga,
Gembloux D.R. Congo). Forests. https://doi.org/10.3390/f14040687
Mama A, Sinsin B, De Cannière C, Bogaert J (2013) Anthropization et Ndenge TC, Perfect-Mrema J (2022) Dependency, Distribution and
dynamique des paysages en zone soudanienne au nord du Bénin. Diversification (3Ds) of livelihoods in the Miombo woodlands:
Tropicultura 31(1):78–88 learning from Tanzania. Int J Sustain Dev 29(5):429–439. https://
Mama A, Bamba I, Sinsin B, Bogaert J, De Cannière C (2014) Défor- doi.org/10.1080/13504509.2022.2032461
estation, savanisation et développement agricole des paysages de O’Neill RV, Krummel JR, Gardner REA, Sugihara G, Jackson B,
savanes-forêts dans la zone soudano-guinéenne du Bénin. Bois DeAngelis DL, Graham RL (1988) Indices of landscape pattern.
Forêts Trop 322:66–75. https://doi.org/10.19182/bft2014.322. Landsc Ecol 1(3):153–162. https://d oi.o rg/1 0.1 007/B
F0016 2741
a31237 Oiro S, Comte JC, Soulsby C, MacDonald A, Mwakamba C (2020)
Mas JF (2000) Une revue des methodes et des techiques de télédétec- Depletion of groundwater resources under rapid urbanisation in
tion du changement. Can J Remote Sens 26(4):349–362 Africa: recent and future trends in the Nairobi Aquifer System,
Maseko H, Shackleton CM, Nagoli J, Pullanikkatil D (2017) Chil- Kenya. Hydrogeol J 28(8):2635–2656. https://doi.org/10.1007/
dren and wild foods in the context of deforestation in rural s10040-020-02236-5
Malawi. Hum Ecol 45(6):795–807. https://doi.org/10.1007/ Ojoyi M, Mutanga O, Odindi J, Aynekulu E, Abdel-Rahman E (2015)
s10745-017-9956-8 The effect of forest fragmentation on tree species abundance and
McGarigal K, Cushman SA (2002) Comparative evaluation of diversity in the Eastern arc mountains of Tanzania. Appl Ecol
experimental approaches to the study of habitat fragmentation Environ Res 13(2):307–324
effects. Ecol Appl 12(2):335–345. https://doi.org/10.1890/1051- Olofsson P, Foody GM, Stehman SV, Woodcock CE (2013) Making
0761(2002)012[0335:CEOEAT]2.0.CO;2 better use of accuracy data in land change studies: estimating
Mganga ND (2022) Comparison on vegetation composition and struc- accuracy and area and quantifying uncertainty using stratified
ture in fire prone forest of Ilunde and fire suppressed forest of estimation. Remote Sens Environ 129:122–131. https://doi.org/
Kitwe in Tanzania. Tanzan J Sci 48(2):294–303 10.1016/j.rse.2012.10.031
Mng’omba SA, Akinnifesi FK, Mkonda A, Mhango J, Chilanga T, Olofsson P, Foody GM, Herold M, Stehman SV, Woodcock CE, Wulder
Sileshi G, Ajayi OC (2015) Ethnoecological knowledge for MA (2014) Good practices for estimating area and assessing
identifying elite phenotypes of the indigenous fruit tree, Uapaca accuracy of land change. Remote Sens Environ 148:42–57.
kirkiana in the miombo woodlands of Southern Africa. Agroecol https://doi.org/10.1016/j.rse.2014.02.015
Sustain Food Syst 39(4):399–415. https://d oi.o rg/1 0.1 080/2 1683 Payne C, Panda S, Prakash A (2018) Remote sensing of river ero-
565.2014.980486 sion on the Colville river, North Slope Alaska. Remote Sens
Moura I, Duvane JA, Silva MJ, Ribeiro N, Ribeiro Baros I (2018) 10(3):397. https://doi.org/10.3390/rs10030397
Woody species from the Mozambican miombo woodlands: Potapov PV, Turubanova SA, Hansen MC, Adusei B, Broich M, Alt-
a review on their ethnomedicinal uses and pharmacological statt A, Mane L, Justice CO (2012) Quantifying forest cover loss
potential. J Med Plant Res 12(2):15–31. https://doi.org/10.5897/ in democratic Republic of the Congo, 2000–2010, with landsat
JMPR2017.6540 ETM+ data. Remote Sens Environ 122:106–116. https://doi.org/
Moyo M, Aremu AO, Van Staden J (2015) Medicinal plants: an invalu- 10.1016/j.rse.2011.08.027
able, dwindling resource in sub-Saharan Africa. J Ethnophar- Rakotondrasoa OL, Malaisse F, Rajoelison G, Gaye J, Razafimanant-
macol 174:595–606. https://doi.org/10.1016/j.jep.2015.04.034 soa TM, Rabearisoa M, Bogaert J (2013) Identification des
indicateurs de dégradation de la forêt de tapia (Uapaka bojeri) Lubumbashi (Haut-Katanga, RD Congo) entre 1989 et 2014.
par une analyse sylvicole. Tropicultura 31(1):10–19 Tropicultura 38(1):98–108
Rannestad MM, Gessesse TA (2020) Deforestation and subsequent cul- Useni YU, André M, Mahy G, Kaleba SC, Malaisse F, Kankumbi
tivation of nutrient poor soils of miombo woodlands of Tanzania: FM, Bogaert J (2018b) Interprétation paysagère du processus
long term effect on maize yield and soil nutrients. Sustainability d’urbanisation à Lubumbashi: dynamique de la structure spa-
12(10):4113. https://doi.org/10.3390/su12104113 tiale et suivi des indicateurs écologiques entre 2002 et 2008.
Rapport semestriel de la Commune de Kasenga, 2022. Recensement Anthropization des paysages katangais 281–296
général de la population congolaise et étrangère dans la com- Useni SY, Boisson S, Cabala KS, Nkuku KC, Malaisse F, Halleux JM,
mune de Kasenga. Territoire de Kasenga, République Démocra- Bogaert J, Munyemba KF (2020a) Dynamique de l’occupation
tique du Congo du sol autour des sites miniers le long du gradient urbain-rural
Rwanga SS, Ndambuki JM (2017) Accuracy assessment of land use/ de la ville de Lubumbashi, RD Congo. Biotechnol Agron Soc
land cover classification using remote sensing and GIS. Int J Environ 24(1):14–27
Geosci 8(04):611–622. https://doi.org/10.4236/ijg.2017.84033 Useni SY, Khoji MH, Bogaert J (2020b) Miombo woodland, an eco-
Sabah SA, Ghadeer RK, Manaf FA, Jobair BA (2022) Detecting and system at risk of disappearance in the Lufira Biosphere Reserve
assessing the spatio-temporal land use land cover changes of (Upper Katanga, DR Congo)? A 39-years analysis based on land-
Bahrain Island during 1986–2020 using remote sensing and GIS. sat images. Glob Ecol Conserv 24:e01333. https://doi.org/10.
Earth Syst Environ 6(4):787–802 1016/j.gecco.2020.e01333
Sadda AS, Diouf A, Lawali S, Ouedraogo M, Bogaert J, Mahamane Useni SY, Mpanda MM, Malaisse F, Kaseya PK, Bogaert J (2023) The
A (2016) Pression anthropique et dynamique paysagère en zone spatiotemporal changing dynamics of miombo deforestation and
rurale semi-aride: cas de la commune de Tibiri, région de Maradi illegal human activities for forest fire in Kundelungu National
(Niger). Tropicultura 34(2):129–139 Park, Democratic Republic of Congo. Fire 6(5):174. https://doi.
Shapiro AC, Trettin CC, Küchly H, Alavinapanah S, Bandeira S (2015) org/10.3390/fire6050174
The mangroves of the Zambezi delta: increase in extent observed Vinya R, Malhi Y, Brown ND, Fisher JB, Brodribb T, Aragão LE
via satellite from 1994 to 2013. Remote Sens 7(12):16505–16518 (2019) Seasonal changes in plant–water relations influence pat-
Syampungani S, Geldenhuys CJ, Chirwa PW (2016) Regeneration terns of leaf display in miombo woodlands: evidence of water
dynamics of miombo woodland in response to different anthro- conservative strategies. Tree Physiol 39(1):104–112. https://doi.
pogenic disturbances: forest characterisation for sustainable man- org/10.1093/treephys/tpy062
agement. Agrofor Syst 90(4):563–576. https://doi.org/10.1007/ Wu J, Gao W, Tueller PT (1997) Effects of changing spatial scale on the
s10457-015-9841-7 results of statistical analysis with landscape data: a case study.
Turner MG, O’Neill RV, Gardner RH, Milne BT (1989) Effects of Geogr Inf Sci 3(1–2):30–41
changing spatial scale on the analysis of landscape pattern. Lands Wulder MA, Roy DP, Radeloff VC, Loveland TR, Anderson MC, John-
Ecol 3:153–162 son DM, Cook BD (2022) Fifty years of landsat science and
Tyukavina A, Hansen MC, Potapov P, Parker D, Okpa C, Stehman impacts. Remote Sens Environ 280:113195. https://doi.org/10.
SV, Kommareddy I, Turubanova S (2018) Congo basin forest 1016/j.rse.2022113195
loss dominated by increasing smallholder clearing. Sci Adv Xiao Q, Hu D, Xiao Y (2017) Assessing changes in soil conservation
4(11):eaat2993. https://doi.org/10.1126/sciadv.aat2993 ecosystem services and causal factors in the three gorges reser-
Useni SY (2019) Analyse spatio-temporelle des dynamiques voir region of China. J Clean Prod 163:172–180. https://doi.org/
d’anthropization paysagère le long du gradient urbain-rural de 10.1016/j.jclepro.2016.09.012
la ville de Lubumbashi (Haut-Katanga, République démocratique Yaovi CR, Hien M, Kabore SA, Sehoubo YJ, Somda I (2021) Uti-
du Congo). Bois Forêts Trop 342:107–108. https://doi.org/10. lisation et vulnérabilité des espèces végétales et stratégies
19182/bft2019.342.a31725 d’adaptation des populations riveraines de la Forêt Classée du
Useni SY, Chukiyabo KM, Tshomba KJ, Muyambo ME, Kapalanga Kou (Burkina Faso). Int J Biol Chem Sci 15(3):1140–1157
KP, Ntumba NF, Mpundu MM, Baboy LL, Nyembo KL (2013)
Utilisation des déchets humains recyclés pour l’augmentation de Springer Nature or its licensor (e.g. a society or other partner) holds
la productivité du maïs (Zea mays L.) sur un ferralsol du sud-est exclusive rights to this article under a publishing agreement with the
de la RD Congo. J Appl Biosci 66:5070–5081 author(s) or other rightsholder(s); author self-archiving of the accepted
Useni SY, Malaisse F, Cabala KS, Munyemba KF, Bogaert J (2017) Le manuscript version of this article is solely governed by the terms of
rayon de déforestation autour de la ville de Lubumbashi (Haut- such publishing agreement and applicable law.
Katanga, RD Congo): synthèse. Tropicultura 35(3):215–221
Useni SY, Cabala K, Halleux JM, Bogaert J, Munyemba K (2018a)
Caractérisation de la croissance spatiale urbaine de la ville de

Sikuzani, Et Al. Anthropogenic Pressures and Spatiotemporal Dynamics of Forest Ecosystems in The Rural and Border Municipality of Kasenga

Uploaded by

Copyright:

Available Formats

Sikuzani, Et Al. Anthropogenic Pressures and Spatiotemporal Dynamics of Forest Ecosystems in The Rural and Border Municipality of Kasenga

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Sikuzani, Et Al. Anthropogenic Pressures and Spatiotemporal Dynamics of Forest Ecosystems in The Rural and Border Municipality of Kasenga

Uploaded by

Copyright:

Available Formats

Landscape and Ecological Engineering (2024) 20:195–212

Anthropogenic pressures and spatio‑temporal dynamics of forest

Keywords Forests · Ecosystem services · Anthropization of landscapes · Deforestation · Migration flows

In sub-Saharan Africa, forests are a survival capital for

GW grassland and wetland, BBS built-up and bare soil

Fig. 2 Decision tree to identify

BBS Savanna GW Forest Agriculture Total 1989

BBS 0.10 0.00 0.16 0.00 0.09 0.34

BBS Savanna GW Forest Agriculture Total 1998

BBS 0.19 0.00 0.01 0.00 0.11 0.32

BBS Savanna GW Forest Agriculture Total 2004

BBS 0.26 0.00 0.00 0.00 0.08 0.35

BBS Savanna GW Forest Agriculture Total 2009

BBS 0.36 0.00 0.00 0.00 0.17 0.53

BBS Savanna GW Forest Agriculture Total 2014

BBS 0.46 0.00 0.00 0.00 0.17 0.64

Fig. 3 Identification of the land a

Relative value of spatial configuration indices

Fig. 4 Evolution of the domi- 100 1989

Largest pacth index of the forest (%)

You might also like