Spatio-Temporal Dynamics of Land Use/Land Cover in the Western Highlands of Cameroon (WHC) and Its Implications on Water Resources Management ()
1. Introduction
For a long time, environmental issues have not been taken seriously, with concern restricted to the scientist or activist communities, with no one taking responsibility for the damage being done [1]. Cities in Sub-Saharan Africa (SSA) are expanding at unprecedented rates, dramatically altering the landscape with resources either being changed or exhausted [2]. The current world population of 7.6 billion is expected to reach 8.6 billion in 2030, 9.8 billion in 2050, and 11.2 billion in 2100 [3]. This increase will therefore require increased food production, which will need to occur in the context of mounting water scarcity, likewise decreasing the area under forest cover due to deforestation and environmental degradation.
Forests are one of the great inherited resources of the earth and play a key role in the economic development of many countries through the provision of goods and services for human welfare. Forests are valuable in providing food, shelter, fuel, medicinal ingredients, and paper that human beings and other animals depend on [4]. They provide habitat for the largest assemblage of species of any terrestrial ecosystem and store about 25 % of the total carbon in terrestrial systems, which otherwise would reside in the atmosphere as greenhouse carbon dioxide (CO2) gas [5]. Moreover, they also provide a unique habitat for a number of rare and endangered species [6]. Besides fuel wood and timber, the forest ecosystem also supports a wide range of Non-Timber Forest Products (NTFPs), such as wild fruits, nuts, rattan, and medicinal plants, which support the livelihood of many indigenous forest dwellers [7]. Forests play a direct role in nutrient cycling and soil formation as well as pest and pollution control. They also play a very important role in the hydrological cycle as their presence or absence affects the amount of runoff and the amount of water infiltrating into the soil [8]. Susswein et al. [9] also reported that forests regulate climate through the exchange of water, energy, and momentum with the atmosphere as well as maintain water supply and quality by protecting watersheds from soil erosion, floods, and drought. They therefore supply, store, and retain water in watersheds and natural reservoirs by regulating the water flow and protecting against storms, erosion, and floods [10].
Savanna regions may contain significant areas of forest, usually in riparian environments [11]. Gallery forests represent one of the few examples of naturally fragmented tropical forest patches [12]. These are narrow strips of forest associated with creeks and rivers, forming continuous strips along streams in an otherwise unforested landscape [11]. They are highly prone to fragmentation and destruction due to human impact, causing a great threat, owing to Land Use Changes [13].
According to FAO [14], global forest cover drastically decreased from 4128 million hectare in 1990 to 3999 million hectare in 2015, corresponding to a drop from 31.1 to 30.6% (1.5% loss); This trend has been recorded in Cameroon with a 1.53 million ha loss from 2001-2020 corresponding to a 4.9 % decrease due to Land Use changes.
The WHC is considered the breadbasket of Cameroon due to its agricultural and livestock rearing potential. However, this is one area of the country noted for a gradual degradation of the agro-sylvo-pastoral resource base because of irrational exploitation. It is estimated to have a surface area of 31,192 km2 [15] and has been witnessing massive population growth in the last decades, with a density of about 128 persons/km2 [16].
The reservoir in the Bamendjin area, which constitutes a major watershed, has been very useful to the growing local population in the pursuit of their developmental activities. The Bamendjin Dam, which is built across the River Noun, reinforces the hydroelectric plant and is meant to boost the main electricity production plant on the River Sanaga at Edea by holding back the River Noun and its braided tributaries in the flood plain during the dry season. It constitutes an abundantly vast “water empire” [17] which serves not only as a major reservoir for water supply but also fosters agriculture, grazing, and fishing activities [18]. Since its construction, waters held behind have generated a number of problems on the one hand, and set into motion a number of competing water and land use systems on the other hand [19].
Owing to the combined effect of increasing population, uncontrolled urbanization, agricultural expansion, climate change, and the persistent aggressive cycles of poverty, people are compelled to encroach on forests and water sources for livelihood sustenance [14]. Regardless of their importance, irrational exploitation has resulted in unprecedented destruction and continued degradation of these natural resources [12]. The watershed in the Bamendjin area has been witnessing remarkable degradation, which stems largely from the drivers of LULCC. These drivers at different spatial scales [20] affect ecosystem functions and services, such as water regulation [21], water quantity [22], and water quality [23]. There is thus an urgent need to assess the status of forests in this priority area in a bid to understand land cover types that may have been affected over time by different drivers. This will help in awareness raising and capacity building of the population of this priority area, as well as streamline policy efforts towards specific sustainable development goals 6 (sustainable water resources), 6.6 (catchment restoration), 11 (sustainable cities and communities), and 15 (land sustainability). This study, therefore, sought to evaluate relative changes in land use/ land cover, identify drivers of change in areal extent, and assess the effects of these changes on water resources of the Western highlands of Cameroon (WHC) over a 36-year span.
2. Materials and Methods
2.1. Description of Study Area
WHC is situated in the Savanna ecological zone of Cameroon, specifically in the highland plateaus of the West and Northwest regions. It is located between Latitudes 5˚40ʺ - 6˚80ʺ North and Longitudes 10˚20ʺ - 10˚40ʺ East [24] (Figure 1).
Figure 1. Location of study area in the western highlands of Cameroon.
The climate is a tropical Soudanian type with annual average rainfall from 1500 - 2600 mm and an annual average temperature range of 14˚C and 28˚C [25]. Altitude ranges from 1500 m - 2500 m asl [26]. The soils are Sandy clay ferruginous soils with pH range of 4.5 - 5.0. Vegetation varies from the low plateaus to the high plateaus due to altitudinal zonation in the drainage basin. The principal vegetation consists of Grasses with herbs [27] and is usually dominated by Pennisetum purpureum, Imperata cylindrical [28], and arboreal plants. On the high plateaus (above 1,500 meters), remnants of dense vegetation dominated by Symphonia globulifera, Strombosia schefflera, Albizia gumnifera, and Allophylus bullatus, exist in inland valleys and slopes [29]. A large part of the plateau is covered by “prairie” vegetation, which makes this area one of the main valuable grazing lands in Cameroon. On the low plateaus (1,000 - 1,500 meters), remnants of humid forests occur and are protected in some areas as sacred places for traditional sacrifices typically found around palaces [30].
The economy of the region depends largely on agriculture and livestock, which are the major activities carried out and occupy almost the totality of the land. Traditional or subsistence agriculture, which entails slash and burn shifting cultivation, is the most common agricultural practice dominating. Approximately 80% of the entire population is engaged in this agricultural practice, while approximately 6% of the industrious population practices livestock farming [16]. The Project area is approximately 2025 km2 around the Bamendjim reservoir. It is situated within six divisions in Cameroon: Ngoketunjia, Mezam, Noun, Mifi, Bamboutos, and Menoa.
2.2. Land-Use and Land Cover Assessment
2.2.1. Data Acquisition and Image Pre-Processing
Satellite images selected for the time series analysis of LULCC in the Western Highlands of Cameroon included Landsat Thematic Mapper (5TM, 1984), Landsat Enhanced Thematic Mapper Plus (7ETM+, 2001, 2010), and Landsat Operational Land Imager with Thermal Infrared Sensor (8OLI-TIRS, 2020) (Table 1). These images were accessed freely from the Global Visualization Viewer of the United States Geological Survey (USGS) (http://www.usgs.gov/ downloaded on 18th and 23rd December 2021) and geo-rectified to UTM (Universal Transverse Mercator) WGS (World Geodetic System) 84 with radiometric corrections and spatial resolutions of 30 m and 15 m for the Panchromatic band. These are orthorectified Landsat images that were easily integrated into a Geographic information system (GIS). The images were acquired approximately at the beginning of the dry season of the selected years, ensuring low cloud cover and minimizing variabilities in phenological stages of plant cover. Figure 2 summarizes all the methods and analyses used in producing the land use and land cover maps and change detection.
The downloaded images were imported into ERDAS Imagine® version 10.1 band after band for different band combinations. Darkest pixel corrections were done to enhance radiometric and geometric values and image visualization. Co-registration was done using the 1984 image as the base image, onto which the other images were warped using the image-to-image registration method and first-order polynomial warping function in ERDAS Imagine® version 10.1. False colour composites bands (752) of the different years depicting the vegetation image pixels were trained and categorized into appropriate classes [31]. Images were then resized to the same spatial resolution and later subsetted into the study area to determine the land use and land cover types by classifying images.
Table 1. Characteristics of images used for a time series analysis.
Sensor |
Acquisition Date |
Spatial Resolution |
Path/Row |
Landsat TM 5 |
29/11/1984 |
30m |
186/56 |
Landsat ETM 7 |
05/02/2001 |
30m |
186/56 |
Landsat ETM 7 |
16/01/2010 |
30m |
186/56 |
Landsat 8 |
04/02/2020 |
30m |
186/56 |
Figure 2. Methodological flow chart used in producing the land-use land cover maps and change detection.
2.2.2. Image Classification and Accuracy Assessment
Unsupervised classification was done using the ISO-Data technique in ERDAS Imagine® version 10.1. with computerized assignment of spectral signatures to the various land use and land cover types. At least 200 Ground Control Points (GCPs) were automatically generated and used during the field phase for land use land cover verification. A ground truthing survey was done to obtain ground reference data to test the reliability of the unsupervised classification and understand the general land cover classes of the study area. During the survey, GPS coordinates of ground points were recorded in the different land use and land cover types and later used to transform former land use and land cover types before performing a supervised classification. Data (GCPs) obtained from field visits were input into the supervised classification using the Maximum Likelihood Parametric Classifier (MLC) and Nearest Neighbor Algorithm (NNA) to identify various features and/or land use classes on the images based on assigned spectral signature [32] [33]. These training signatures were created with at least 10,000 pixels per land use and land cover class (Heo and Fitzhugh 2000) and automatically classified all pixels in an image into nine (9) land use and land cover categories. These categories included: (1) Dense Forest, (2) swamps and Gallery Forest, (3) montane forest, (4) Built up areas, (5) Farmlands, (6) water bodies, (7) Savanna, (8) degraded forest, and (9) Bare soils (Table 2).
The classification was later polygonised and exported to ArcGIS version 10.1 for further processing using Ground Control Points as validation set. All ground control points were overlain on the classified maps for verification and determination of the accuracy of fit according to the method of Cogolton [34]. The producer, user, and overall accuracies, as well as the Kappa coefficient statistic, were computed from the resulting confusion or error matrix to test the accuracy of the classified maps of 1984, 2001, 2010 and 2020 following the method of Disperati and Virdis [35];
(1)
(2)
(3)
The kappa statistic value is a measure of the agreement between classification and reference data [36].
(4)
where:
= Probability of agreement and
= Probability of random agreement.
Table 2. Land use/Land cover categories and their descriptions.
S/N |
LULC Category |
Description |
1. |
Dense forest |
Naturally occurring stands of continuous cover of evergreen and semi-deciduous trees clustered together. The forest is intact, high-stemmed with relatively little or no disturbance. |
2. |
Gallery forest and swamps |
Forests that form as corridors along rivers or wetlands and dominated by Raphia, mature natural trees, and tall grasses. |
3. |
Montane forest |
Forest types situated at high altitudes and on mountain slopes with a mixed species of broad leave evergreen trees and epiphytic plants. |
4. |
Built-up Areas |
Consist of human construction dedicated to residential and commercial sectors, as well as
infrastructure (roads, playgrounds, etc.) |
5. |
Farmland |
All Cultivated lands for both annual and perennial crops, as well as juvenile cash crop plantations. |
6 |
Water bodies |
Rivers, streams, lakes, dams, reservoirs in wetlands, and fish ponds. |
7. |
Savanna |
Grassland ecosystem characterized by the stunted trees sufficiently spaced with little or no canopy cover. |
8 |
Degraded forest |
Forestland with sparse crown cover of 15% - 60%, with relatively low stem density indicating
degradation caused by planned or uncontrolled logging, and agricultural activities. |
9. |
Bare ground |
Areas void of vegetation, water, settlements or farmlands with little or no vegetation cover,
including open fields, bare slopes, and exposed rocky escarpments, bare soils, quarry areas,
exposed sand bands, and eroded soil deposits. |
2.2.3. Post Classification Change Detection
LULCC detection was performed using change detection analysis in a post-classification approach. The technique used the GIS overlay in ArcGIS version 10.1 based on generated vector themes of the different years. LULC change over time was quantified by evaluating gains and losses of specific LULC categories. Change detection was computed for the different LULC category in terms of (i) surface area (ha) for the years 1984, 2001, 2010 and 2020, (ii) area change (ha) for 4 periods (1984-2001, 2001-2010, 2010-2020 and 1984-2020), (iii) % area change per period. Since the duration per period vary, the change estimates were annualized for easy comparison and evaluation as follows: (iv) annual rate of change, and (v) % annual rate of change. These estimations of change were computed based on the formulae of Kashaigili [37]:
(5)
(6)
(7)
where
= period in years between the first and second scene acquisition data.
2.3. Assessment of Drivers of LULC Changes within Local Communities in WHC
Based on the hydrological map of the study area, some main streams identified were selected, and the population around these streams was targeted for assessment of impacts of LULCC on water resources through the use of semi-structured, open-ended questionnaires. A total of 201 questionnaires were administered purposefully and proportionally to respondents in 12 localities in 4 divisions namely; Foumban (35), Foumbot (35) and Mangoum (20) (Noun Division), the Baleng community made up of Tchada (15), Dionku (15), Fampi (15) and Tyo (15) (Mifi Division), Bameboro (20), Batsuetim (15), Batossusong (15) and Galim (30) (Bamboutos Division) and Bamendou (20) (Menoua Division). These localities were chosen to represent the full range of LULC changes observed and for their proximity to key water bodies. Mostly people of 45 years and above, and considered to have a good knowledge of the area, were targeted. The questionnaires were partitioned into three themes with specific indicator variables for assessment: (a) General demography (b) Local perception of LULCC and water resource management, and (c) Assessment of drivers of change (farming, NTFP, and forest exploitation) that affect water supply. The Chi-Square statistic (χ2) was used to evaluate a significant difference (p ≤ 0.05) between indicator variables for a given theme.
3. Results
3.1. Land Use/Land Cover in the WHC
Figure 3. Classified land use and land cover maps of the Western Highland Cameroon (WHC) for the year 1984 (a), 2001 (b), 2010 (c), and 2020 (d).
The overall accuracy assessment of the supervised maps was 98.14%, 93.32%, 89.85% and 88.77%, respectively, for years 1984, 2001, 2010, and 2020 with a Kappa index of 0.95, 0.93, 0.89, and 0.93, respectively (Table 3). These values indicate a significant agreement with reference data and satisfy the minimum accuracy threshold of 85% needed for efficient land use/cover change analysis and modelling [38]. Supervised classification maps of the nine major LULC categories for the years 1984, 2010, 201, and 2020 are presented in Figure 3, and their respective surface area and percent contribution are in Table 4. In 1984, the dense forest (48100 ha), savannah (37258 ha), and swamps and gallery forest (32000 ha) were dominant LULC, accounting for 73.4% of the total area. There was a complete shift in 2020, with savannah (49898 ha), farmland (24425 ha), and degraded forest (20016 ha) accounting for 59.0% of the total area. Over the years, farmland, settlement, bare ground, and degraded forest increased steadily in surface area in the classification maps of 1984, 2001, 2010, and 2020 (Figure 4 and Table 3). On the contrary, the dense forest, swamp, and gallery forest, water bodies, and montane forest were on the decline.
Table 3. Accuracy assessments for classification maps of the different land use/land cover classes for 1984, 2001, 2010, and 2020.
LU/LC classes |
1984 |
2001 |
2010 |
2020 |
PA (%) |
UA (%) |
PA (%) |
UA (%) |
PA (%) |
UA (%) |
PA (%) |
UA (%) |
Dense Forest |
98.23 |
97.88 |
91.55 |
92.44 |
89.43 |
83.79 |
91.35 |
93.63 |
Savannah |
98.77 |
98.55 |
95.57 |
96.62 |
88.20 |
91.98 |
94.89 |
96.07 |
Settlements |
99.92 |
97.52 |
94.18 |
97.17 |
94.31 |
97.43 |
94.18 |
97.22 |
Swamp & Gallery Forest |
98.54 |
96.48 |
89.07 |
90.87 |
83.39 |
89.95 |
89.93 |
89.63 |
Farmland |
99.05 |
96.41 |
90.28 |
95.63 |
86.33 |
88.01 |
90.29 |
95.62 |
Water Bodies |
100.00 |
100.00 |
100.00 |
94.97 |
100.00 |
95.34 |
100.00 |
95.06 |
Bare ground |
92.08 |
99.58 |
90.37 |
85.60 |
86.73 |
86.01 |
90.41 |
85.60 |
Montane Forest |
98.15 |
98.40 |
93.32 |
91.80 |
88.42 |
91.68 |
94.16 |
92.82 |
Degraded Forest |
97.40 |
98.95 |
95.70 |
92.93 |
91.95 |
86.85 |
95.42 |
92.48 |
Overall Accuracy |
98.14 |
93.32 |
89.85 |
93.32 |
Kappa Index |
0.95 |
0.93 |
0.89 |
0.93 |
Note: PA = Producer’s Accuracy, UA = User’s Accuracy.
Table 4. Area of land-use/land cover category in the WHC (1984-2020).
LULC category |
1984 |
2001 |
2010 |
2020 |
Area (ha) |
% |
Area (ha) |
% |
Area (ha) |
% |
Area (ha) |
% |
Dense Forest |
48,100 |
30.1 |
39,259 |
24.5 |
32,218 |
20.1 |
21,645 |
13.5 |
Savannah |
37,258 |
23.3 |
40,167 |
25.1 |
44,848 |
28.0 |
49,898 |
31.2 |
Settlement |
2502 |
1.6 |
3789 |
2.4 |
4210 |
2.6 |
8965 |
5.6 |
Swamp and Gallery Forest |
32,000 |
20.0 |
29,789 |
18.6 |
25,400 |
15.9 |
21,028 |
13.1 |
Farmland |
8752 |
5.5 |
15,298 |
9.6 |
20,410 |
12.8 |
24,425 |
15.3 |
Water Bodies |
10,247 |
6.4 |
9612 |
6.0 |
7897 |
4.9 |
6587 |
4.1 |
Bare ground |
2577 |
1.6 |
2872 |
1.8 |
3687 |
2.3 |
3985 |
2.5 |
Montane forest |
9564 |
6.0 |
7982 |
5.0 |
4521 |
2.8 |
3451 |
2.2 |
Degraded Forest |
9,000 |
5.6 |
11,232 |
7.0 |
16,809 |
10.5 |
20,016 |
12.5 |
Figure 4. Trends in LU/LC categories for the years 1984, 2001, 2010, and 2020.
3.2. Patterns of LULC Change in the Western Highland of Cameroon (WHC)
Land cover statistics throughout the time span of 36 years showed varying dynamics, with all nine categories recording above 33% net change (loss or gain). Settlement, farmland, Bare ground, degraded forest, and bare soil and savannah recorded positive change of 258.31%, 179.08%, 122.40%, 54.64% and 33.93% respectively (Table 5). This gain probably resulted from a loss in montane forest (−63.92%), dense forest (−55.00%), and swamp and gallery forest (−34.29%). Water bodies equally recorded a total area loss of 3660 ha (−35.73%) throughout the period (Table 5, Figure 5).
Table 5. Surface Area change of the different land-use/land cover categories in the WHC (1984-2020).
LULC category |
1984-2001 |
2001-2010 |
2010-2020 |
1984-2020 |
Area change (ha) |
% |
Annual rate (%/yr) |
Area change (ha) |
% |
Annual rate (%/yr) |
Area change (ha) |
% |
Annual rate (%/yr) |
Area change (ha) |
% |
Annual rate (%/yr) |
Dense Forest |
−8841 |
−18.38 |
−1.08 |
−7041 |
−17.93 |
−1.99 |
−10573 |
−32.82 |
−3.28 |
−26455 |
−55.00 |
−1.53 |
Savannah |
2909 |
7.81 |
0.46 |
4681 |
11.65 |
1.29 |
5050 |
11.26 |
1.13 |
12640 |
33.93 |
0.94 |
Settlements |
1287 |
51.44 |
3.03 |
421 |
11.11 |
1.23 |
4755 |
112.95 |
11.30 |
6463 |
258.31 |
7.18 |
Swamp and Gallery Forest |
−2211 |
−6.91 |
−0.41 |
−4389 |
−14.73 |
−1.64 |
−4372 |
−17.21 |
−1.72 |
−10972 |
−34.29 |
−0.95 |
Farmland |
6546 |
74.79 |
4.40 |
5112 |
33.42 |
3.71 |
4015 |
19.67 |
1.97 |
15673 |
179.08 |
4.97 |
Water Bodies |
−635 |
−6.20 |
−0.36 |
−1715 |
−17.84 |
−1.98 |
−1310 |
−16.59 |
−1.66 |
−3660 |
−35.72 |
−0.99 |
Bare ground |
295 |
11.45 |
0.67 |
815 |
28.38 |
3.15 |
298 |
8.08 |
0.81 |
1408 |
54.64 |
1.52 |
Montane forest |
−1582 |
−16.54 |
−0.97 |
−3461 |
−43.36 |
−4.82 |
−1070 |
−23.67 |
−2.37 |
−6113 |
−63.92 |
−1.78 |
Degraded Forest |
2232 |
24.80 |
1.46 |
5577 |
49.65 |
5.52 |
3207 |
19.08 |
1.91 |
11016 |
122.40 |
3.40 |
![]()
Figure 5. Change in surface of the different LULC categories within WHC.
Between 1984 and 2001, farmland and settlement recorded the highest increase in surface area of 6546 ha (74.79%) and 1287 ha (51.44%), respectively (Table 5, Figure 5). This was followed by degraded forest with an increase of 2232 ha (24.80%) and bare ground 295ha (11.45%). Savannah, with the second-highest surface area, also witnessed an increase of 2909 ha (7.81%). All three categories of forest witnessed reductions in surface area, with dense forest losing 8841 ha (−18.38%), montane forest 1582 ha (−16.54%), and swamps and gallery forest 2211 ha (−6.91%). Water bodies equally recorded a reduction of 635 ha (−6.20%). Surface area of degraded forest, bare ground, and savannah approximately doubled between 2001 and 2010 with an increase of 5577 ha (49.65%), 815 ha (28.38%), and 4681 ha (11.65%). Farmland and settlements equally increased in surface area by 5112 ha (33.42%) and 421 ha (11.11%), though less than previous estimates (Table 5, Figure 5). Loss in surface area was almost threefold for montane forest, swamp, and gallery forest, and water bodies recording 3461 ha (43.36%), 4389 ha (14.74%), and 1715 ha (17.84%) respectively. Loss in the dense forest was not different from the 1984-2001 estimate (Table 5). The greatest changes were recorded in all LULC categories between 2010 and 2020, with a very high increase in settlement of 4755 ha (112.95%) and relatively lower for degraded forest and farmland, and savannah. Loss in dense forest of 10573 ha (32.82%) between 2010 and 2020 was, however, the highest record in the 36-year span.
The overall annual rates of change show that settlements (7.18%/yr), farmland (4.97%/yr), and degraded forest (3.40%/yr) accounted for the dynamism of the LULC (Table 5, Figure 6). These gains were accounted for by losses in the montane forest (−1.78%/yr), dense forest (−1.53%/yr), water bodies (−0.99%/yr), and swamp and gallery forest (−0.95%/yr). This trend was the same for the different periods (1984-2001, 2001-2010, and 2010-2020), though at varying rates. The period 2001 - 2010 recorded the highest annual rates of change (Figure 6).
Figure 6. Annual rate of land cover change for the different LULC categories within WHC.
3.3. Assessment of Drivers of LULC Changes within Local Communities in WHC
Majority of the respondents were males (71.6%), and a good number of them had no formal education (41.8%), with farming (63.8%) as the main occupation of the population. A summary of respondents’ views on local perception of LULC change and water supply is presented in Table 6 and Table 7. Water source for daily needs was very variable, with the majority relying on wells (32.8%), streams (27.9), rivers (13.9%), and springs (9.5%). Supply was more acute in the dry season (93%), and the impact in terms of shortages was felt 15-25 years ago. The possible reasons for this drop in supply include: increasing local population (57.2%), poor catchment management (20.9%), and as a result of loss of some feeder streams (20.9%). The opinion of the respondents with respect to the reasons for the loss of major water sources also varied significantly (P < 0.001), with high concentration of houses and farmlands around major catchments (20.9%) and poor farming practices along water courses (20%) being the most cited reasons. Almost all of the respondents lamented that the water crises were affecting their personal and communal life by hampering domestic activities (39.8%), increasing disease incidence (27.5%), lateness of school pupils (17.4%), and reduction in crop yields (14.9%).
Table 6. Assessment of water supply and usage within the WHC.
Parameter |
Category |
Frequency |
Percentage frequency |
X2-statistics |
P-value |
Water source for daily needs |
Catchments |
28 |
13.9 |
73.1 |
<0.001*** |
Rivers |
19 |
9.5 |
Springs |
20 |
10.0 |
Streams |
56 |
27.9 |
Taps |
12 |
6.0 |
Wells |
66 |
32.8 |
Season with acute water shortages |
Both |
14 |
7.0 |
148.9 |
<0.001*** |
Dry season |
187 |
93.0 |
Duration of
experiencing water
crises sources |
15 years |
71 |
35.3 |
16.1 |
<0.001*** |
20 - 25 years |
88 |
43.8 |
35 - 40 years |
42 |
20.9 |
Possible factors
contributing to water crises |
High concentration of houses and farmland around major catchments |
17 |
8.5 |
133.1 |
<0.001*** |
High temperature |
15 |
7.5 |
Increasing local population |
55 |
27.4 |
Loss of forest |
1 |
0.5 |
Loss of some water sources that feed main stream |
42 |
20.9 |
Over population |
14 |
7.0 |
Poor catchment management |
42 |
20.9 |
Poor farm practices around water sources |
14 |
7.0 |
Reduced volume from source |
1 |
0.5 |
Effects of water shortage on communal life |
Incidence of diseases |
55 |
27.4 |
32.1 |
<0.001*** |
lateness of school pupils |
35 |
17.4 |
Hampers domestic activities |
81 |
40.3 |
Reduction of crop yields |
30 |
14.9 |
A high proportion of the respondents (79.1%) noticed changes in the quality of the vegetation of the area, with more than half (56.7%) reporting a decrease in forest cover and savanna over the years. Most of the respondents (83%) affirmed the following anthropogenic activities to be responsible for the degradation of the vegetation and gradual loss of forests: opening of new farms (33.3%) and human settlement (29.9%). There was a clear indication of a lack of awareness on the link between forest cover change and water availability among the respondents (52.7%), though some were of the opinion that agricultural activities reduce forest cover (Table 7).
Table 7. Respondent views on the relationship between forest loss and provision of water.
Parameter |
Category |
Frequency |
Percentage frequency |
X2-statistics |
P-value |
Notice any change in vegetation in the course of time |
No idea |
22 |
10.9 |
189.5 |
<0.001*** |
No |
20 |
10.0 |
Yes |
159 |
79.1 |
Observed changes in
vegetation |
Forest cover is decreasing, and the savannah is decreasing |
114 |
56.7 |
138.4 |
<0.001*** |
Forest cover is decreasing, and
savannah is increasing |
61 |
30.3 |
Forest cover is increasing, and
savannah is increasing |
12 |
6.0 |
No response |
14 |
7.0 |
Reasons for observable changes in vegetation |
Construction of houses |
60 |
29.9 |
14.8 |
0.002** |
Opening of new farms |
67 |
33.3 |
Over grazing |
34 |
16.9 |
Poor farming practices |
40 |
19.9 |
Possible link between
forest cover change and water availability |
No |
39 |
19.4 |
36.2 |
<0.001*** |
No idea |
106 |
52.7 |
Yes |
56 |
27.9 |
Agricultural practices
affect forest cover |
No |
48 |
20.9 |
7.8 |
0.008** |
No idea |
69 |
37.3 |
Yes |
84 |
41.8 |
Farming and extractive use of the surrounding forest were considered the key agents of change, and a summary of the respondents’ views is presented in Table 8 and Table 9, respectively. Approximately 60.7% of respondents were involved in farming and had 4 - 6 farms (46.8%) per household, with farm sizes of 0 - 1 ha (53.2%). The majority of these farms were aged 15 years and above (46.8%). Acquisition of these farms was either inherited (66.7%), bought (20.4%), or rented (12.9%), with slash/burn being the most cited method of establishment (62.2%). Agroforestry was also a common practice carried out by the majority (60.7%) of the farmers with intercropping of fruit trees and other valuable tree species for bee farming.
Table 8. Assessment of farming as key drivers of change in the western highland Cameroon.
Parameter |
Category |
Frequency |
Percentage frequency |
X2-statistics |
P-value |
Dependence on the forest as a source of livelihood |
No |
69 |
34.3 |
19.74 |
<0.001*** |
Yes |
132 |
65.7 |
Main activities carried out around the forest |
Farming |
122 |
60.7 |
76.87 |
<0.001*** |
Fuelwood and NTFPs |
57 |
28.4 |
|
Tapping |
22 |
10.9 |
|
|
Number of farms cultivated per household |
1 - 3 |
62 |
30.8 |
18.48 |
<0.001*** |
4 - 6 |
94 |
46.8 |
6 and above |
45 |
22.4 |
Method of farm acquisition |
Bought |
41 |
20.4 |
0.102.18 |
|
Inherited |
134 |
66.7 |
Renting |
26 |
12.9 |
Duration of farming |
0 - 5 yrs |
49 |
24.4 |
16.93 |
<0.001*** |
15 yrs and above |
94 |
46.8 |
5 - 10 years |
58 |
28.9 |
Approximate size of farm |
0 - 1 ha |
107 |
53.2 |
59.22 |
<0.001*** |
1 - 3 ha |
75 |
37.3 |
above 5 ha |
19 |
9.5 |
Method used to establish farms |
Clearing |
33 |
16.4 |
153.61 |
<0.001*** |
Ploughing |
12 |
6.0 |
slash/burning |
125 |
62.2 |
Spraying |
31 |
15.4 |
Location of farm(s) |
Around houses |
65 |
32.3 |
8.65 |
0.034* |
Around the forest |
45 |
22.4 |
flood plains |
37 |
18.4 |
Savannah |
54 |
26.9 |
Approximate distance to the farm(s) |
0 - 500 m |
65 |
32.3 |
5.94 |
0.051ns |
1 - 5 km |
54 |
26.9 |
500 - 1 km |
82 |
40.8 |
Practice Agroforestry |
No |
70 |
34.8 |
18.51 |
<0.001*** |
Yes |
131 |
65.1 |
Ethnobotanical assessment of the extractive uses of the surrounding forests affirmed high dependence on the forest as a source for livelihood (Table 9), medicinal plants (34.4%), and food (33.3%) being the most significant NTFPs collected. Noticeable changes in vegetation were also observed with core NTFPs on the decline (42.8%) or had disappeared (33.0%), resulting in non-significant shifts of collection points either at the forest centre (40.8%), forest edge (30.8), or at random (28.4%).
Table 9. Ethnobotanical assessment on extractive use of surrounding forests.
|
Category |
Frequency |
Percentage frequency |
X2-statistics |
P-value |
Dependence on the forest as a source of livelihood |
No |
69 |
34.3 |
|
|
Yes |
132 |
65.7 |
End-use of NTFPs
collected from forests |
Crafts |
14 |
7.0 |
39.50 |
<0.001*** |
Food |
70 |
33.3 |
fuel wood |
50 |
24.9 |
medicinal plant |
67 |
34.8 |
Types of NTFPs collected from the forests |
Bamboo |
24 |
10.7 |
47.95 |
<0.001*** |
Medicinal plants |
103 |
46.2 |
Mushroom and snails |
70 |
31.4 |
Raffia wine |
26 |
11.7 |
Plant forms harvested for medicinal uses |
Herbs |
73 |
36.3 |
0.925 |
0.630ns |
Shrubs |
66 |
32.8 |
Trees |
62 |
30.8 |
Plant parts used |
Barks |
43 |
21.4 |
112.3 |
<0.001*** |
Leaves |
71 |
35.3 |
Rhizomes |
13 |
6.5 |
Roots |
6 |
3.0 |
Sap |
13 |
6.5 |
Seeds |
17 |
8.5 |
Whole plant |
38 |
18.9 |
Collection points |
Forest centre |
82 |
40.8 |
5.22 |
0.073ns |
Forest edge |
62 |
30.8 |
Random |
57 |
28.4 |
Noticeable changes in availability of plants in
forest |
No |
45 |
22.4 |
137.0 |
<0.001*** |
No response |
13 |
6.5 |
Yes |
143 |
71.1 |
Level of availability |
Decreasing |
92 |
42.8 |
61.6 |
<0.001*** |
Disappeared |
71 |
33.0 |
Increasing |
24 |
11.2 |
Undecided |
28 |
13.0 |
4. Discussion
4.1. Land-Use Land Cover Dynamics
Nine principal land use/ cover classes were identified in this study, with some showing gains in surface area while others showing losses. The decrease in forest cover (dense, montane, and swamp, and gallery forests) observed in this study could be attributed to their conversion to agricultural fields and settlement to ensure food security and accommodate the increasing human population, respectively. Population growth, coupled with land scarcity, has forced farming households to expand their agricultural fields into natural environments in order to increase their agricultural production and cover their daily household needs. Similar outcomes were observed by Ewane et al. [20], who examined implications of agricultural expansion and land use/ cover change on local land use planning and sustainable Development, and reported an increase in built-up areas and farmland as a result of an increase in human population. There has been continuous farming on the slopes around the mountain, as Ndzeidze [29], also reported that very little of this highland montane forest remains, as the forests have been cleared over the years in the quest for farming, settlements, and grazing land. A study in the Bamboutos watershed to assess the impact of land use/ cover change on water quality reported the conversion of gallery forest, wetlands and swamp, into farm land, commercial vegetable farming and market gardening with residential settlement as well, while the savanna vegetation is transformed into animal grazing sites [39]. In this same study, although the natural forest was being converted to arable farming and settlement expansion, the natural forest was also highly regarded and protected for its cultural values, as families’ shrines and the village’s secret forest were used for rituals and sacrifices. In addition to being lost to farmland and settlement areas, ecological land (open primary forest land, open secondary forest land, grass land, and bare rock area) has been transformed into tea and eucalyptus plantations that were established for commercial purposes to supply ready-to-consume beverages and market electricity poles for the national electricity company over the past 30 years [20]. Almayehu et al. [31] delineated the following as direct causes of land use and land cover change in the Somodo watershed; illegal logging and fuelwood extraction, expansion of plantation, expansion of settlement, agricultural expansion, and construction of infrastructures such as schools, roads, and research center.
Land scarcity largely accounts for intensive cultivation with no fallow periods. The massive decline in forest cover is often associated with agricultural expansion in the periphery of the forest, while timber exploitation and charcoal production are other problems that contribute to the decline in forest cover [40]. Forests remain potential sites for opening new farms, as the best crop yields, they reported, were obtained from farms inside the forest and at the forest edge. This could be attributed to high organic matter content in forest soils, caused by litter decomposition. Gallery forests also have alluvial soils that are far more fertile than the soils of the savannah. This is due to nutrients in the stream associated with organic matter from litter decomposition. This has resulted in a general downward trend in the area covered by these forests on a spatiotemporal basis. Owing to the scarcity of land, savannahs are also gradually being converted to farmland.
4.2. Drivers of LULC Change and Implications for Water Resources Management
From the socio-economic survey, farming, collection of fuel wood, and NTFPs were the main economic activities carried out by the population, and these activities have largely accounted for the degradation of the forest in the WHC. The main drivers of LULC in the WHC are intensive subsistence farming, collection of fuel wood, increased construction of housing, unregulated and overgrazing, unsustainable farming practices, and overexploitation of NTFPs. The increased pressure could be attributed to population growth, hence increased settlement development, expansion of farmland through slash-and-burn agricultural practices, forest clearing, and spraying has led to high deforestation rates, destruction of water catchments, and loss of soil fertility. Momo et al. [12] made similar observations while studying land use/ land cover and anthropogenic causes around the Koupa-Matapit area. The degradation of forests in this region is a result of the introduction of sun-loving crops, which encourages the clearing of shade, hence aggravating forest loss over time. Our result concurs with the observations made by Teshome et al. [41], who investigated the dynamics of land cover in the Muger sub-basin, Oromia, Ethiopia, and identified the main factors that are responsible for the land cover change as well as the effects of these changes. Their study revealed agricultural expansion ranking first as the main driver of land cover change, followed by population growth, wood extraction, expansion of settlement, and infrastructure development. The noted majority of the farmers owned farmland of sizes 0-1ha and had approximately 4-6 farms per household, which is estimated at 6ha per household. Thus, the scarcity of favourable land for farming as a result of population growth increases pressure, deforestation of the nearby forested (dense, montane, swamps and gallery forests) areas, and savannah vegetation all contribute to the degradation of the watershed and water resource. Field survey advanced reason for change in vegetation being construction of new houses, opening of new farm lands, overgrazing, and poor farming practices, noting significant differences. Our findings correspond with those of Kometa and Ebot [18], who recounted that a reduction in pasture/fresh grass for livestock resulted from overgrazing due to an increase in the number of animals, not only resulting in vegetal cover degradation, but also degradation of some watershed ecosystems. River valleys in the Bamboutos watershed are being altered into intensive marketable farming sites due to their low relief, fertile soil, and water supply availability all year round, swamps and wetlands now occupied by market garden farms [18] [38]. The nucleation of this build-up houses, coupled with the intensive farming and migration of livestock into swamps and gallery forest areas during the dry season, results in the shrinking of the water resource and land use conflict. The decrease in forested land (dense, montane, swamps, and gallery forests), with increases in farmlands, settlement, barren land savannah, and water bodies, and poor catchment management, poor farming practices around water sources highlight the need for sustainable water resource management.
The increased exploitation of the various NTFPs leaves a significant impact on the forest ecosystem and its services. Indeed, the removal of leaves, barks, and roots could cause a vulnerability of plants or extinction of rare species, which in effect could result in forest degradation through loss of biodiversity [42]. Moreover, the almost daily human presence within the forest and harvesting of NTFPs and other wood products are at the origin of the loss or the disappearance of some valuable species [43]. The main drivers of deforestation and degradation are therefore the combination of several factors, including the expansion of peasant agriculture, poor farm practices, and the construction of houses. The socio-economic survey identifies slash-and-burn as the primary method for farm establishment that must be accounting for observed increase in the surface area of degraded forests and savannah vegetation.
The increasing human population, coupled with the scarcity of favourable land for farming, grazing and construction of settlements causing the agricultural front to migrate to previously unexploited areas such as the banks of water courses [44] These activities are all taking their toll on the watershed appear to be the drivers of watershed degradation putting the supply of fresh water at risk [18]. Indicators such as the decline in water quantity and the increase in water-related illnesses are active issues that draw attention to questions on the environmental quality and degradation. Other issues, such as the seasonality of streams and scarcity of potable water in the dry season, are just part of the chain of problems originating from the degradation of the watershed. It is worth noting that the long cycles of poverty in the WHC intensify the watershed degradation process [45]. The case of the region is remarkable due to its high population density and a rural economy that depends on subsistence agriculture, which yields little income and can barely sustain livelihoods [45] [46]. The varied nature of the physical and human landscape thus constitutes the framework for the spatial processes, which considerably account for the watershed degradation [18]. These agrarian changes in response to pedoclimatic changes weaken natural ecosystems, more especially, the water ecosystems, including the gallery forests that naturally surround rivers and streams. Climate change and human induced changes are considered the two major factors impacting hydrological processes [47] [48]. Result showed that the WHC hydrological basin is faced with decreasing stream flows as evident in the shrinking of the surface area of water bodies. Climate variability has led to rising temperatures, increased evapotranspiration and changing precipitation patterns that play vital role in streamflow trends and surface area of water bodies, while human activities have changed the temporal and spatial distribution of water resources. Though the GIS analyses could not pick streams, the gradual decline in the surface area of the Bamendjin dam was witnessed.
These findings advocate that the stability, functioning, and resilience of the WHC ecosystem have been significantly undermined over the past 40 years. There is an urgent need for the sustainable management of the LULC, restoration of degraded ecosystems, especially water resources in the study area, and policy reforms promoting Participatory Land Use Planning (PLUP). The orientation Law on land use and sustainable development planning in 2011 by the Ministry of Economy, Planning and Regional Development (MINEPAT), Cameroon, laid no emphasis on local land use planning; however, it is expected to develop a national framework document for the PLUP. In Cameroon, local land use and sustainable development plans are prepared at the level of the village communities and adopted by the council of the municipalities concerned, for integration into the regional and national land use plans. This is according to the 2011 Orientation Law on Land Use and Sustainable Development Planning. Hence, the project on Mount Bamboutos landscape, initiating the PLUP, serves as part of the preliminary phase of the national land use plan planning, ensuring the planned allocation zones of the proposed local land use are integrated into the council, regional, and national land use plan documents. Stimulated land tenure security towards achieving REDD+ and sustainable development in Cameroon is apparent from the successful development, validation, and integration of local land use plans into council, regional, and national land use plans [20].
5. Conclusions
Over the 36 years span, there has been a constant decrease in area of the different land cover owing to an increase in area of the different drivers of change. Dense forest, montane forest, gallery forests, and water bodies witnessed a constant decrease in trend in terms of surface area, while Savannah, farmlands, built-up areas, degraded forest, and bare soils increased over the years.
Rapid population growth of the WHC in the last decades has resulted in the clearance of natural vegetation to allow room for settlement and agricultural land for food security. Therefore, the need for service and fuel wood, agricultural land, and settlements is at the heart of the multifaceted pressures on the forests and land cover change. Therefore, the stability, diversity, functioning, and resilience of the ecosystem have been disturbed for more than 40 years, with significant degradation of the forest and forest resources.
These drivers have, in effect, ushered in profound degradation of the Bamendjin watershed, putting it under serious threat. The high levels of poverty and the over-dependence on agricultural activities only worsen the situation. Water scarcity remains an issue of major concern and a hindrance to development. The fact that watersheds are zones of social and economic conflicts undermines efforts to improve the lives of large numbers of the poor who depend on these resources for their survival and economic well-being, as land users have different goals for the available natural resources. The results provide important baseline information to land managers, land users, and large land owners on the urgent need for land use planning of the WHC. Climate change and human induced changes are considered the two major factors impacting hydrological processes. Result showed that the WHC hydrological basin is faced with decreasing stream flows as evident in the shrinking of the surface area of water bodies. The long-term data record of monthly water-surface area and climatic can serve as base data for studies that may provide scientific foundation for understanding the potential confounding impacts of climate change and anthropogenic activities on local hydrology and providing strategies for effective control by changing local land-use practices.
The Bamendjin watershed holds much promise for the sustainable development and the sustenance of local livelihoods in the area. Given the present dispensation, the logical way forward remains the harmonious observation and management of this very crucial natural resource system. It is important that awareness be raised on the indicators of watershed degradation and the importance of watershed management. This will reflect an appropriate approach for the sustainable use and management of natural resources. Institutional and organizational agreements should be enforced and better managed in order to ensure effectiveness in the watershed programmes. A balance between population and carrying capacity of upland watersheds must be achieved with the intention to protect and conserve some of our catchments. In a diversified relief watershed like the WHC which is fast degrading, there is need for significant policy and programme responses and solutions which must not ignore the basic needs of people living in the uplands. Incentives may be needed to drop existing land-use practices for environmentally acceptable and sustainable resource utilization. Land tenure and the complexity of user rights that constrain the development of land use practices needed to stabilize forest and rangelands should be revisited. Finally, multiple use options employing agroforestry, production forests, protection of forests on critically steep slopes, capitalization of payment ecosystem services and other practices which provide needed goods and services and sustainable use of water resources must be promoted in appropriate areas of the communities.
Funding
This research received no external funding.
Data Availability Statement
The data presented in this study are available on request from the corresponding author.
Author Details
Department of Plant Science, University of Buea, Buea, Cameroon and School of Geoscience, University of South Florida, Tampa, USA.
Authors’ Contributions
Conceptualization, methodology and investigations: G.B.C., N.A., and B.A.F.; data curation and formal analysis, B.A.F; C.M.; L.D.A.; and G.B.C., writing—original draft preparation, B.A.F.; writing—review and editing, G.B.C., N.A., B.A.F.; C.M.; and L.D.A.; funding acquisition, B.A.F.; N.A.; and G.B.C. All authors have read and agreed to the published version of the manuscript.
Acknowledgements
The authors acknowledge with gratitude the financial, logical and material support provided by family members and friends. We are grateful to Zed Ngendoh Sanga for sourcing the images and his invaluable remote sensing techniques, Elias Ndive, Fritz Mukow and community members for assisting during ground truthing and socioeconomic survey. We thank the staff at UNVDA Ndop and Bamendjin offices for providing long-term information vital for this study.