EN中文

Journal of Geographical Sciences >

2017 , Vol. 27 >Issue 5: 579 - 605

DOI: https://doi.org/10.1007/s11442-017-1394-4

Research Articles

Assessment of multifunctional landscapes dynamics in the mountainous basin of the Mo River (Togo, West Africa)

  • DIWEDIGA Badabate , 1, 3 ,
  • AGODZO Sampson 2 ,
  • WALA Kperkouma 3 ,
  • LE Quang Bao 4
Expand
  • 1. WASCAL Graduate Research Program in Climate Change and Land Use; Department of Civil Engineering, Kwame Nkrumah University of Science and Technology, Kumasi, Ghana
  • 2. Department of Agricultural Engineering, Kwame Nkrumah University of Science and Technology, Kumasi, Ghana
  • 3. Laboratory of Botany and Plant Ecology; University of Lomé, Togo
  • 4. CGIAR Research Program on Dryland Systems (CRP-DS), International Centre for Agricultural Research in Dry Areas (ICARDA), Amman 11195, Jordan;

Author: Diwediga Badabate, PhD, specialized in integrated soil-landscape assessment and modeling.E-mail:;

Received date: 2016-01-27

  Accepted date: 2016-06-28

  Online published: 2017-05-10

Copyright

Journal of Geographical Sciences, All Rights Reserved
Fold

Abstract

In this study, historical landscape dynamics were investigated to (i) map the land use/cover types for the years 1972, 1987, 2000 and 2014; (ii) determine the types and processes of landscape dynamics; and (iii) assess the landscape fragmentation and habitat loss over time. Supervised classification of multi-temporal Landsat images was used through a pixel-based approach. Post-classification methods included systematic and random change detection, trajectories analysis and landscape fragmentation assessment. The overall accuracies (and Kappa statistics) were of 68.86% (0.63), 91.32% (0.79), 90.66% (0.88) and 91.88% (0.89) for 1972, 1987, 2000 and 2014, respectively. The spatio-temporal analyses indicated that forests, woodlands and savannahs dominated the landscapes during the four dates, though constant areal decreases were observed. The most important dynamic process was the decline of woodlands with an average annual net loss rate of -2%. Meanwhile, the most important land transformation occurred during the transition 2000-2014, due to anthropogenic pressures. Though the most important loss of vegetation greenness occurred in the unprotected areas, the overall analyses of change indicated a declining trend of land cover quality and an increasing landscape fragmentation. Sustainable conservation strategies should be promoted while focusing restoration attention on degraded lands and fragmented ecosystems in order to support rural livelihood and biodiversity conservation.

Key words： land cover dynamics; spatio-temporal patterns; swap change; landscape fragmentation; protected areas; Mo River Basin; Togo

Cite this article

DIWEDIGA Badabate , AGODZO Sampson , WALA Kperkouma , LE Quang Bao . Assessment of multifunctional landscapes dynamics in the mountainous basin of the Mo River (Togo, West Africa)[J]. Journal of Geographical Sciences, 2017 , 27(5) : 579 -605 . DOI: 10.1007/s11442-017-1394-4

1 Introduction

Controlling the dynamics of land resource for global change mitigation remains ever of greater importance at global scale as well as national and local levels. Estimates indicated that forest areas have decreased of about 3.1% over the past 25 years and the largest deforestation and forest degradation occur in tropical areas of America and Africa (FAO, 2015). Similarly, the same report highlighted that between 2010 and 2015, the annual gain of forestlands was far below to the extent of loss, resulting in a net annual loss of 3.3 million ha of forests per annum. Regardless of the scale, processes inducing the loss of natural vegetation cover, forest fragmentation and associated land use/cover changes (LUCC) occur continuously, leading to land cover quality decline and degradation. From the perspective of the Millennium Ecosystem Assessment (MEA, 2005), the degradation of vegetation functions and services (e.g. forest degradation, habitat fragmentation), and the loss of natural land productivity (e.g. deforestation) are two major manifestations of the phenomenon of land degradation. From this view, long-term, permanent and regular monitoring of the landscape dynamics is regarded as an essential step for a real understanding of the change drivers and for modelling mindsets towards mitigation of land degradation and achievement of sustainability (Houet et al., 2010; Fan et al., 2013).
Spatio-temporal changes of landscapes are continuous processes fully maintained by both natural and human-related drivers. Human imprints on terrestrial ecosystems are of large extent (Ellis, 2011; Gaia, 2011) and of major environmental concerns related to habitat fragmentation, decline of land and ecosystem services (ESS), biodiversity loss, livelihood decrease and climate instability (Balthazar et al., 2015; Schleuning et al., 2011). The drivers of these forest fragmentation and deforestation are mainly human-induced such as agriculture, large-scale forest harvest, and small-scale forest disturbances (Lambin et al., 2003; Damnyag et al., 2013; Specht et al., 2015). The increasing demand for agricultural lands and forest products induced the reduction of land conservation capacities and the ESS provision, even in protected areas that have such devoted conservation role (Castro et al., 2015; Vedeld et al., 2012; Traoré et al., 2012; Folega et al., 2014b; Damnyag et al., 2013). In this “Anthropocene” epoch, the complexity of interactions between human and environmental systems and subsequent effects have raised much more awareness (Ellis, 2013; Ellis et al., 2013).
A recent growing concern in scientific communities is how much land has changed and how this dynamic is going to affect the future of land resources. Therefore, spatial patterns of land degradation, its causal factors and potential impacts have been investigated through various approaches, scales and dimensions (Zheng et al., 2014; Kim et al., 2014; He et al., 2014; Vu et al., 2014; Le et al., 2012). Geographic information systems (GIS) and remote sensing (RS) approaches have evolved increasingly to support monitoring landscape dynamics at different scales for better planning and management (Kennedy et al., 2009; Griffiths et al., 2012; Portillo-Quintero et al., 2012). These satellite archives such as Landsat data have brought new insights into the approaches of understanding of LUCC, and monitoring deforestation and forest degradation (DFD) processes. The successes of the application of GIS and RS in the assessment of landscape dynamics remain in the availability of earth observation data as well as the multitude of methods for land cover mapping and degradation assessment (Zhou et al., 2012; Kim et al., 2013; Gounaridis et al., 2014 He et al., 2014; Kim et al., 2014; Zheng et al., 2014; Zhou et al., 2014a). These data and methods offer great potential to cover various spatio-temporal scales of analyses and monitoring of LUCC (Farooq, 2012; Rogan and Chen, 2004). Change detection analyses also brought significant contribution to the understanding of the processes involved in landscape dynamics. Furthermore, habitat quality assessment and landscape fragmentation analyses emerged from the combination of these tools and data to provide more appraisal to landscape dynamics and impacts on ecosystems. Landscape metrics are ecological indicators used to quantify the composition and spatial configuration of landscapes (Uuemaa et al., 2013; Mander and Uuemaa, 2010; Peng et al., 2010). Thus, as an essential approach in quantifying landscape spatial patterns with distinct ecological implications, landscape metrics help in the analysis of LUCC patterns and related ecological effects (Walz, 2011; Uuemaa et al., 2013). Thus, the integration of multi-temporal satellite data in combination with GIS and field data showed great insights in addressing landscape change and degradation at various scales.
In Togo, several recent studies have been undertaken using GIS and RS to assess and monitor the changes in land resources at national and local scales (Folega et al., 2014b; Folega et al., 2014a; Badjana et al., 2014; Folega et al., 2011; Folega et al., 2015). Human activities especially agricultural expansion, illegal tree logging and incursions in PA as well as settlement enlargement in rural and semi-urban areas have been identified as responsible for most of the changes (Fontodji et al., 2011; Dourma et al., 2009; Kokou et al., 2009). These studies have shown that current trends of land resources do not favor the functional services of the different ecosystems. Therefore, continual and complementary studies are encouraged to deepen the knowledge on the processes and determinants of LUCC for supporting integrated land management. Especially, the monitoring of the processes involved in land dynamics in rural mountainous areas of the Mo River Basin have not yet gained enough attention although they are the location of numerous social and ecological interactions (Dourma et al., 2009). Therefore, this study focused on the analysis of the historical landscape dynamics in the Mo River Basin in order to provide up to date information on landscape transformation to support management and planning. With a glance at land management regime, the study specifically aimed at (1) identifying and mapping of the major land/use cover types for the dates 1972, 1987, 2000 and 2014; (2) determining the types and processes of landscape dynamics as well as their rate of occurrence; and (3) assessing the landscape fragmentation and habitat loss over time using FRAGSTAT-based landscape metrics. With regard to the social and ecological importance of the natural landscapes in the area, knowledge on the historical LUCC processes is critical to ensure better future strategies of land management and rural development.

2 Study area

The study is carried out in the Mo watershed (Figure 1), a sub-basin of Volta Basin (West Africa). The area is particularly sensitive as it contains great parts of the Fazao - Malfakassa National Park, which is likely to be subjected to more human pressures (Woegan, 2007). The population are mainly rural and farming is the main activity. In 2010 the central region, embedding the Mo River Basin, is the region with the lowest density of population (47 inhbts/km2 in 2010 versus 21 inhbts/km2 in 1981) compared to 109 inhbts/km2 at the national level (DGSCN, 2010). The Mo basin is part of the Ecological zone 2 of Togo (Ern, 1979), characterized by a mosaic of mountainous dry and open forests, guineo-soudanian savannahs and
agro-ecosystems within a human-transformed landscape (UPA) (Dourma et al., 2009). In 1990-1991, with the socio-political crisis in Togo, the population in these areas coming back from the south have increased. Serious inadequate land management took place within this incorrect multiple use of lands. Additionally, local people used to withdraw wood and non-woody products, inducing forest degradation, and in the worst case, deforestation and desertification. Some parts of the hilly lands have an altitude greater than 800 m, especially in Aledjo Mounts. The climate includes cool nights at upland regions, a rainy season from April to October. Mean annual rainfall in the area is between 1200-1300 mm with an irregular spatio-temporal distribution. Mean minimal and maximal temperatures reach respectively 19ºC in January with the Harmattan winds and 30ºC in April. The rivers/streams network is heavily developed in accordance with the mountainous relief. Foremost of the land uses in the area is small-scale subsistence farming, pasture lands, PA and built-up areas. The prominent environmental issues are land degradation due to overgrazing, unsustainable agricultural land use, fuel wood harvesting and charcoal production (Wala et al., 2012). Illicit incursions for hunting and tree logging in PA are also concerns that cause conflicts between land users and state agencies protecting lands (Dourma et al., 2009; Woegan, 2007). In addition, protection legislation regarding the PA is weak, and non-inclusive. The catchment is relatively important for tourism and one of the breadbaskets (crop production) of the country.
Figure 1 Location of the study area

3 Data sources and methods

3.1 Mapping the land-use and cover for the period 1972-2014

Land-use and cover maps of 1972, 1987, 2000 and 2014 for the Mo river basin were obtained from the classification of historical Landsat archives (Jianjun et al., 2005; Pattanaik et al., 2011). Single ortho-rectified images of Landsat 8 (03 December 2014), Landsat ETM+ (04 December 2000) and Landsat TM (30 October 1987) free of cloud were collected at the path 193/row 054 (https://earthexplorer.usgs.gov). A pair of Landsat MSS (10 November 1972) was downloaded from the same source at the path 207/row 053 and path 207/row 054. These data are acquired for the time corresponding to the early dry season onset (October to December), enabling the clear distinction between land use/cover (LUC) types, especially agricultural fields and typical savannahs in the landscape (Traore et al., 2014; Ruelland et al., 2010).
Developing LUC maps was the process of clustering and assigning similar pixels into classes (Rojas et al., 2013). Therefore, to reduce the effects of typical similarities between closer cover types such as savannahs and woodlands, which look similar in savannah-dominated landscapes, a transformation was necessary. Since the study is interested in vegetation cover mapping, normalized difference vegetation index (NDVI) was computed as independent layer not only to reduce the effects of topography but also to measure the distribution of vegetation health over the landscape of interest (Braimoh and Vlek, 2004). NDVI is widely used as a powerful indicator of vegetation greenness, and less sensitive to topographic factors in mountainous areas (Matsushita et al., 2007; Diallo et al., 2010). Original bands were combined with the NDVI layer to perform the pixel-based supervised classification using Maximum Likelihood algorithm in ENVI 4.7 image processing software.
Though topography is a common source of biases in LUC classification in mountainous areas, DEM data were not integrated into the classification process, as the maximal elevation above sea level which is around 850 m, does not really provide significant hill shade at the sensor passing time (Diallo et al., 2010). The combined layers were registered to UTM WGS 84 projection system and used to extract spectral signatures for the classification of the respective images (Braimoh, 2004; Wittig et al., 2007; Gutiérrez Angonese and Grau, 2014).
Six main categories were defined based on the LUC classification scheme from the national vegetation map (Afidégnon et al., 2003). The classification scheme was supplemented by the United Nations Food and Agriculture Organization Land Cover Classification System in order to better consider the physiognomy-structural conditions (vertical and horizontal arrangements as well as land use affected on the cover types) of the vegetation types.
(1) Forests: close canopy vegetation including the riparian forests along streams and dry forests in lowlands. The canopy cover exceeds 60% with an understory layer.
(2) Woodlands: open canopy vegetation including woody savannahs and woodlands. The canopy cover comprises between 30% to 60% and do not possess understory vegetation making their cover less thick compared to forests. Trees are higher than 5 m.
(3) Savannahs: Treeless open canopy vegetation composed of tree savannahs, shrubs, and scattered grasslands. Generally, tree height is lower than 5 m. They have a bush or grass dominant layer with woody coverage less than 30%. This category includes old fallows.
(4) Agricultural land: cultivated (including cereal crops, vegetable crops and fruit orchards) and non-cultivated (farm fallows less than 3 years and parklands) lands;
(5) Built-up areas: areas occupied by residential settlements as well as paved surfaces.
(6) Water: surface water bodies including rivers and reservoirs.
Paved surfaces and bare rocks are mostly confused among settlements and agricultural lands since they reflect in the quite similar range.
For each abovementioned LUC type, training areas were developed independently and the spectral characteristics of each training sample were checked through the separability tests of Jeffries-Matusita and Transformed Divergence (Zhou et al., 2008; Braimoh, 2004). The output values of the separability tests range from 0 to 2, with 0 indicating poor separability and 2 a total separability i.e. the signatures have no similarity among them. For this study, the both separability measures between the defined classes were acceptable.
Collection of reference data to assess the accuracy of historical maps is often a tremendous issue in data scarce areas (Wilson and Sader, 2002). While the collection of such data is less hectic for recent and current images, it is very difficult for images of long history especially due to the lack of reliable data such as aerial photographs (Biro et al., 2013; White et al., 2013; Zhou et al., 2008). Therefore, the data dearth for the current study constrained to the use of different sources of information. Ground truth information for accuracy assessment of classified map of 2014 relied on the use of 177 field-registered GPS coordinates collected based on a random sampling. These points were collected during field campaigns between February and May 2014 corresponding to the dry season and matching the acquisition season of the image. Validation samples of at least 45 pixels were composed of either the raw GPS points or a blend with homogenous polygons around the GPS points. For the validation of LUC map of 2000, we referred to the available topographic map at 1/200000 (IGN, 1986) and vegetation map of Togo (Afidégnon et al., 2003) combined with LUC maps of 2005 and 2009 from GlobCover project (Bicheron et al., 2008). The reclassification of GlobCover images was performed to meet the classification system used in the current study. Historical Google Earth images were helpful in the creation of these validation information. For the early date images (1987 and 1972), homogeneous areas were selected to create representative validation samples based on the detection of unchanged areas (persistent pixels) along the time series. Hence, using the geographic link tool from ENVI software, the validated maps of 2000 and 2014 were overlaid with each of the classified maps for 1987 and 1972 to collect randomly the validation data from the raw unclassified images (Biro et al., 2013; Waiswa, 2011; Lung and Schaab, 2010). Background knowledge of the study area as well as qualitative information from local informants (local elders) were also helpful in the selection of these validation samples sites.
A confusion matrix, the overall accuracy, the Kappa index of agreement were reported for each LUC map. Ultimately, some post-classification analyses were performed to minimize classification errors due to image registration and georeferencing of satellite images. A clump of 3 x 3 window was applied to all output maps to eliminate the “salt and pepper” polygons (Petursson et al., 2013; Zhai et al., 2015).

3.2 Analyses of change and patterns of land use cover types

The output images were exported to GIS software for change detection analyses between the four individual maps of the basin. Post-classification comparison adopted to detect changes in land-cover types was based on pairwise overlay (bi-temporal analyses) of individual LUC maps (Braimoh and Vlek, 2004; Pang et al., 2013; Pang et al., 2010; Badjana et al., 2014). Class statistics, transitions analyses, conversion categories, and annual rates of occurrence were computed from the output LUC maps for each LUC type and transition category (Zhou et al., 2008; White et al., 2013; Tfwala et al., 2012). Statistics were produced for the three transition periods 1972-1987; 1987-2000; 2000-2014 and the overall period 1972-2014.
We detected the gross gains (total gains), gross losses (total losses), net change (i.e. changes in land quantity), and swap changes (i.e. changes in land location) for each LUCC from a pairwise conversion matrix (Braimoh, 2004; Pontius et al., 2004; Schmitt-Harsh, 2013; Carmona and Nahuelhual, 2012). The gross gain for a category i is expressed as the summative value of all areas gained from other LUC types at a final date. Inversely, the gross loss of a LUC category i is the summative value of all areas converted from i into other LUC types. For each LUC type, the total change area was calculated as the sum of all areas affected by changes (i.e. gross gains + gross losses). Hence, the net change was calculated as the simple difference between gross gains and losses during a transition period. Meanwhile, the swap change was derived by subtracting the absolute net change from the total change for the specific LUC type during a given transition period.
The annual net change for each LUC type was calculated according to Equation 1 below (Carmona and Nahuelhual, 2012; FAO, 1996):
where CR is the annual rate of change of a cover type, Sf is the area of the targeted LUC type at the final time Tf; Si is the area of the same targeted LUC type at the initial time Ti; ln is the natural log function.
Furthermore, we detected the important LUC changes and persistence among the different LUC types for all the transition periods following Pontius et al. (2004). This methodology assumes there is randomness in the landscape transitions when land categories gained from other categories in proportion to the availability of the other losing categories, or reciprocally. Meanwhile the systematic transitions base on the interpretation of the transition proportions relative to the sizes of the categories. A transition is assumed random when the difference between the expected and the actual transition proportions is close to zero while any large value indicates systematic landscape transition (Pontius et al., 2004; Schmitt-Harsh, 2013; Romero-Ruiz et al., 2012). For this study, transitions with an absolute difference value higher than or equal to 0.5 were considered as the most important systematic changes.
In the process of computing proportions for accounting for systematic and random changes, three important variables are used in determining the random and systematic transitions; viz. the observed (actual) transition values, the expected land gains and losses under random processes of gain and loss (Romero-Ruiz et al., 2012; Teferi et al., 2013; Gutiérrez Angonese and Grau, 2014). Whilst the observed transitions were computed from the actual values in the cross-tabulation matrix between two times, equations 2 and 3 were used to calculate the expected gain (Gij) and expected loss (Lij) of each transition under a random process of gain or loss (Schmitt-Harsh, 2013; Pontius et al., 2004; Nakakaawa et al., 2010).
where Gij denotes the expected transition from category i to j under random processes of gain, P+j is the proportion of the landscape in category j in the final time; Pjj is the observed persistent proportion of the category j; Pi+ is the total area of category i at initial time; Lij is the expected transition from category i to j under random processes of loss; and Pii is the proportion of the category i that showed persistence between the two times.
Loss-to-persistence ratio L(-) and gain-to-persistence ratio G(+) were also calculated using Equations 4 and 5 to assess the vulnerability of the land classes to transition (Ouedraogo, 2010; Nakakaawa et al., 2010; Braimoh, 2004; Romero-Ruiz et al., 2012). L(-) value for a cover category higher than 1 indicates a high vulnerability of that category to be converted into other categories (Gutiérrez Angonese and Grau, 2014). Meanwhile, G(+) indicates the tendency of a cover category to gain more from other cover types. These ratios are expressed as follows:

3.3 Landscape fragmentation and degradation using landscape metrics

Spatial patterns in wild and human-impacted systems are often characterized using not only long-term spatial changes in vegetation cover but also the use of landscape metrics (LMs) helps in identifying and measuring landscape dynamic (Schindler et al., 2013; Renetzeder et al., 2010; Uuemaa et al., 2013; Mander and Uuemaa, 2010; Kang et al., 2013). Hundreds of LMs available for characterizing landscape patterns (i.e. composition and structure) and dynamics but we only selected 5 metrics (Table 1) (Zhang et al., 2013; Wu et al., 2014; Weng, 2007; Wang et al., 2013; Wala et al., 2012) to evaluate landscape fragmentation at class level using FRAGSTATS 4.2.1 (McGarigal and Ene, 2013). Using 8 cell neighborhood rule, at only class and landscape levels, LMs were calculated for the four epochs at the whole watershed scale without any sampling.
Table 1 Land cover-based landscape metrics for approximating ecosystem services (ESS) change (McGarigal and Marks, 1995; McGarigal et al., 2002
Indices & Acronyms Meaning of index - Measured ESS
Number of patches (NP) NP ≥ 1, without limit. NP = 1 when the landscape contains only 1 patch of the corresponding patch type. Measure of the extent of class fragmentation
Patch density (PD) PD has the same basic utility as number of patches as an index, except that it expresses number of patches on a per unit area basis that facilitates comparisons among landscapes of varying size. PD > 0, constrained by cell size.
Largest patch index (LPI) Largest patch index quantifies the percentage of total landscape area comprised by the largest patch. As such, it is a simple measure of dominance. 0 < LPI ≦ 100; LPI approaches 0 when the largest patch in the landscape is increasingly small.
Patch cohesion index - COHESION Patch cohesion index measures the physical connectedness of the corresponding patch type. 0 < COHESION < 100; COHESION approaches 0 as the proportion of the landscape comprised of the focal class decreases and becomes increasingly subdivided and less physically connected.
Aggregation index- AI 0 ≦ AI ≦ 100. Aggregation index is calculated from an adjacency matrix, which shows the frequency with which different pairs of patch types (including like adjacencies between the same patch types) appear side-by-side on the map.

4 Results

4.1 Historical land use/cover change and geographic patterns

During the last four decades, major transformations affected the landscapes of the Mo river basin (Figures 2 and 3). On the spatial angle, the central and the northeastern areas of the basin were dominantly covered by human systems (croplands and settlements), especially in 2000 and 2014. Most of the greenest areas from 1972 to 2014 lay within the PA, especially the Fazao-Malfakassa National Park in the southern and western parts of the basin. Savannahs and shrubs were the most scattered LUC types from the earlier dates to the most recent years. Some scattered patches of forests and woodlands were located within the most human-dominated parts, decreased markedly from 1972 to 2014. Meanwhile, the emergence of forest was marked along river network within the PA whereas in the free access lands, their cover decreased constantly over time. Built-up areas mostly developed in the eastern parts along the main road network.
Figure 2 Historical patterns of LUC types in the Mo River basin^(Note: The legend is dedicated to the four LUC maps)
Figure 3 Historical LUCC in the Mo River basin ^Note: The legend is dedicated to the four LUCC maps
In terms of the areal distribution, natural vegetation dominated the landscapes and decreased from 99% in 1972 to 98%, 96% and 91%, respectively in 1987, 2000 and 2014 (Tables 2a, 2b, 2c and 2d). The most dominant LUC types over time were woodlands and savannahs. While the total area of the woodlands decreased from 94,829 ha in 1972 to 40,527 ha in 2014, savannahs increased from 45,000 ha (30%) in 1972 to about 79,000 ha (53%) in 2014. The proportion of forests was about 6% of the total area in 1972 and slightly increased to 6%, 9% and 11% for 1987, 2000 and 2014, respectively. Over the period 1972-2014, there was an increasing forest coverage. Croplands significantly increased from 0.2% in 1972 to 8% in 2014. Similar to the croplands, built-up areas also showed an increasing trend all over the years but still exhibit very low values. All LUC types experienced unidirectional changes between 1972 and 2014, with the exception of water areas which varied in relation to the water levels of Aleheride and Aledjo dams and the riverbeds. In general, the class statistics indicated high wildness and rurality of the study area where agricultural expansion and settlement growth are increasing the trend of human-appropriation of the landscapes.

4.2 Change rates, persistence, gains and losses of land use/cover types

Forests gained about 791 ha, 3497 ha and 3712 ha respectively for the three transition periods. Though forests gained over all the periods, the annual rates of gain decreased during the second and third transition periods (Tables 2b and 2c). The swap changes for forest coverage were of 7.2%; 6.9% and 5.9% for the ordered transition periods. The swapping values were quite higher than the respective absolute net changes (0.5%; 2.4% and 2.5%), suggesting that forest regeneration (gain from other cover types) and degradation or deforestation (loss to other types) during all the transition periods affected much more spatial coverage than the proportions revealed through the net changes. However, the overall net (5.4%) and swap (6.1%) changes over the period 1972-2014 were close, indicating that forest change processes (loss and gain) affected mostly great proportions of its initial spatial coverage in addition to the net gain of about 0.7% of new forest areas.
Table 2 Areal distribution of the four main LUC types and their propotions of changes
Similar to forests, savannahs showed a constant increase in coverage passing from a net gain of 3051 ha (1.3% of initial areas) during the period 1 to approximated net gains of 1876 ha (1.3%) and 29028 ha (19.6%) during the transition periods 2 and 3, respectively (Tables 2). However, land transformation processes affected much more coverage (high swap values of 34.6%, 34.7% and 21% for periods 1, 2, and 3, respectively). The transition period 2 showed the lowest annual rate of gain (0.3%) while the period 3 experienced the highest processes of savannah gain (3.3%). In overall, the net (23%) and swap (25%) changes showed that the processes of savannah gain and loss affected quite same landscapes, even if there was an additional 2% of savannah cover change not captured through the net change. This category has the highest value of swapping (25.11%), suggesting a constant transformation (losses to and gains from other categories) of savannahs during the overall period, which affected about 52.3% of the total changes in this category.
In contrast, woodlands experienced net losses of about 3.9% (5738 ha), 5.4% (7987 ha) and 27.3% (40,519 ha) of their initial coverage during the three periods, respectively (Tables 2). Similar to savannahs, woodlands experienced the high swap changes of 38.2% and 35.6% during the first two periods, indicating that woodlands experienced high exchanges of coverage location (gain and loss) with other categories. However, the swapping in woodlands decreased during the third period (12.3%) because this cover type considerably decreased over time while gained only 6.2% from other categories. Between 1972 and 2014, woodlands had a net loss of about 36.7% out of 49.3% of total change at an annual rate of -2%. The highest rate of woodland loss occurred during the period 1987-2000. The swapping were quite fair (12.7%) suggesting that the changes in woodland locations did not occur much more as revealed from the transition periods. During the overall time, the swapping for woodlands represented about 26% of the total change, which is lower than the decreasing swap changes for this category during transition periods (91%, 87%, and 31% for the 3 periods, respectively).
During the three transition periods, transformations affecting cultivated land consisted of both swap and net changes (Tables 2). While net gains of croplands increased over time, the swapping affected additional areas passing from 0.3% between 1972 and 1987 to 2.4% and 3.6% for the transition periods 2 and 3, respectively. Thus, cultivated lands showed high tendency to gain from other land categories much more than to lose, suggesting that more intensive as well as extensive cultivation occurred in the Mo basin landscapes over the past 42 years. Over all the three time intervals, similar trends were observed for human settlements as their expansion induces agricultural lands in the meanwhile. For all the three periods, swap changes of these settlements far exceeded the observed net changes, meaning that all the areas affected by settlement expansion were underestimated by the net change results.

4.3 Land use cover persistence, systematic and random conversions

Land cover persistence dominated the landscape (more than 50% as sum of first lines of diagonal values in Tables 3). The persistence constantly decreased and accounted for 56%, 55%, and 50% of the landscape, respectively for the ordered 3 transition periods. About 59% (100-sum of the diagonal entries in first lines, in Tables 3) of the landscape did change during the overall 42-year period. Natural vegetation dominated the persistence with woodlands exhibiting the highest but decreasing persistence over time. Meanwhile, the persistence of forests and savannahs constantly increased for all the periods. These categories gained much more from the transformation of woodlands, which persist despite the increasing pressures.
Table 3 Matrices for the four periods under investigation in the Mo basin
(a) Period 1 (1972-1987)
Year 1987
Forests Woodlands Savannahs Croplands Total 1972 Gross loss L(-)
Year
1972		 Forests 2.05 3.22 0.36 0.01 5.64 3.60
2.05 (0.00) 2.98 (0.24) 1.55 (-1.18) 0.08 (-0.07) 6.66 (-1.02) 4.61 (-1.01) 1.76
2.05 (0.00) 2.30 (0.92)* 1.24 (-0.88)* 0.05 (-0.04) 5.65 (0.00) 3.60 (0.00)
Woodlands 3.54 40.88 18.89 0.50 63.85 22.96
2.80 (0.74) 40.88 (0.00) 17.74 (1.15) 0.87 (-0.37) 62.37 (1.48) 21.49 (1.48) 0.56
3.55 (0.00) 40.88 (0.00) 18.56 (0.34) 0.78 (-0.28) 63.85 (0.00) 22.96 (0.00)
Savannahs 0.59 15.78 12.97 0.84 30.28 17.31
1.33 (-0.74) 15.99 (-0.21) 12.97 (0.00) 0.41 (0.43) 30.75 (-0.46) 17.77 (-0.46) 1.33
1.58 (-0.99)* 15.35 (0.44) 12.97 (0.00) 0.35 (0.49) 30.28 (0.00) 17.31 (0.00)
Croplands 0.00 0.07 0.09 0.01 0.17 0.16 16
0.01 (-0.01) 0.09 (-0.02) 0.05 (0.04) 0.01 (0.00) 0.16 (0.02) 0.15 (0.02)
0.01 (-0.01) 0.10 (-0.03) 0.05 (0.03) 0.01 (0.00) 0.17 (0.00) 0.16 (0.00)
Total 1987 6.18 59.98 32.34 1.37 100.00 44.09
6.18 59.98 (0.00) 32.34 (0.00) 1.37 (0.00) 100.00 (0.00) 44.06 (0.02)
7.19 (-1.01) 58.66 (1.32) 32.84 (-0.50) 1.20 (0.17) 100 (0.00) 44.09 (0.00)
Gross gain 4.13 19.09 19.37 1.36 44.09
4.13 (0.00) 19.09 (0.00) 19.37 (0.00) 1.36 (0.00) 44.08 (0.00)
5.14 (-1.01) 17.77 (1.32) 19.87 (-0.50) 1.19 (0.17) 44.08 (0.00)
G(+) 2.01 0.47 1.49 136
(b) Period 2 (1987-2000)
Year 2000
Forests Woodlands Savannahs Croplands Total 1987 Gross loss L(-)
Year 1987 Forests 3.22 2.52 0.42 0.01 6.17 2.95
3.22 (0.00) 2.74 (-0.23) 1.70 (-1.28) 0.18 (-0.17) 7.85 (-1.68) 4.63 (-1.68) 0.92
3.22 (0.00) 1.76 (0.76)* 1.08 (-0.66)* 0.10 (-0.09) 6.17 (0.00) 2.95 (0.00)
Woodlands 4.77 36.81 17.25 1.09 59.99 23.18
3.39 (1.38) 36.81 (0.00) 16.49 (0.76) 1.74 (-0.65) 58.57 (1.42) 21.76 (1.42) 0.63
4.35 (0.42) 36.81 (0.00) 17.16 (0.09) 1.54 (-0.45) 59.99 (0.00) 23.18 (0.00)
Savannahs 0.52 14.92 15.01 1.74 32.34 17.33
1.83 (-1.31) 14.38 (0.54) 15.01 (0.00) 0.94 (0.80) 32.24 (0.11) 17.22 (0.11) 1.15
2.23 (-1.70)* 14.25 (0.67)* 15.01 (0.00) 0.79 (0.95)* 32.34 (0.00) 17.33 (0.00)
Croplands 0.01 0.33 0.85 0.15 1.36 1.21
0.08 (-0.07) 0.61 (-0.27) 0.38 (0.48) 0.15 (0.00) 1.21 (0.15) 1.06 (0.15) 8.07
0.11 (-0.10) 0.68 (-0.35) 0.42 (0.43) 0.15 (0.00) 1.36 (0.00) 1.21 (0.00)
Total 2000 8.53 54.60 33.61 3.01 100.00 44.79
8.53 (0.00) 54.60 (0.00) 33.61 (0.00) 3.01 (0.00) 100.00 (0.00) 44.79 (0.00)
44.79 (0.00)
9.92 (-1.39) 53.57 (1.03) 33.71 (-0.11) 2.58 (0.44) 100.00
Gross gain 5.30 17.79 18.60 2.86 44.79
5.30 (0.00) 17.79 (0.00) 18.60 (0.00) 2.86 (0.00) 44.79 (0.00)
6.70 (-1.39) 16.76 (1.03) 18.70 (-0.11) 2.43 (0.44) 44.79 (0.00)
G(+) 1.65 0.48 1.24 19.07
(c) Period 3 (2000-2014)
2014
Forests Woodlands Savannahs Croplands Total 2000 Gross loss L(-)
2000 Forests 5.08 2.04 1.34 0.06 8.53 3.45 0.68
5.08 (0.00) 1.16 (0.88) 3.86 (-2.52) 0.61 (-0.55) 10.74 (-2.21) 5.66 (-2.21)
5.08 (0.00) 1.06 (0.98)* 2.06 (-0.72)* 0.32 (-0.26) 8.53 (0.00) 3.45 (0.00)
Woodlands 4.12 21.12 27.03 2.26 54.60 33.49 1.59
3.55 (0.57) 21.12 (0.00) 24.73 (2.30) 3.93 (-1.67) 53.48 (1.12) 32.37 (1.12)
5.08 (-0.96)* 21.12 (0.00) 24.49 (2.54)* 3.78 (-1.51)* 54.60 (0.00) 33.49 (0.00)
Savannahs 1.75 4.01 23.11 4.58 33.61 10.50 0.45
2.19 (-0.44) 4.57 (-0.56) 23.11 (0.00) 2.42 (2.16) 32.37 (1.23) 9.27 (1.23)
2.47 (-0.72)* 6.12 (-2.11)* 23.11 (0.00) 1.84 (2.74)* 33.61 (0.00) 10.50 (0.00)
Croplands 0.05 0.11 1.61 1.21 3.01 1.80 1.49
0.20 (-0.14) 0.41 (-0.30) 1.37 (0.24) 1.21 (0.00) 3.19 (-0.18) 1.98 (-0.18)
0.22 (-0.16) 0.53 (-0.42) 1.04 (0.57)* 1.21 (0.00) 3.01 (0.00) 1.80 (0.00)
Total 2014 11.03 27.29 53.18 8.20 100.00 0.00
11.03 (0.00) 27.29 (0.00) 53.18 (0.00) 8.20 (0.00) 100.00 49.45
12.87 (-1.84) 28.89 (-1.60) 50.81 (2.36) 7.16 (1.03) 100.00 49.45 (0.00)
Gross gain 5.95 6.17 30.07 6.98 49.45 49.45 (0.00)
5.95 (0.00) 6.17 (0.00) 30.07 (0.00) 6.98 (0.00) 49.45 (0.00)
7.79 (-1.84) 7.77 (-1.60) 27.71 (2.36) 5.95 (1.03) 49.45 (0.00)
G(+) 1.71 0.29 1.30 5.77
(d) Period 4 (1972-2014)
Year 2014
Forests Woodlands Savannahs Croplands Total 1972 Gross loss L(-)
Year 1972 Forests 2.59 1.18 1.76 0.11 5.65 3.05 1.18
2.59 (0.00) 0.99 (0.19) 2.87 (-1.11) 0.46 (-0.35) 6.95 (-1.30) 4.36 (-1.30)
2.59 (0.00) 0.94 (0.24) 1.82 (-0.06) 0.28 (-0.17) 5.65 (0.00) 3.05 (0.00)
Woodlands 5.90 20.93 33.58 3.35 63.85 42.92 2.05
5.71 (0.19) 20.93 (0.00) 32.46 (1.12) 5.21 (-1.86) 64.68 (-0.83) 43.75 (-0.83)
6.51 (-0.61)* 20.93 (0.00) 31.38 (2.20)* 4.84 (-1.49)* 63.85 (0.00) 42.92 (0.00)
Savannahs 2.52 5.16 17.73 4.67 30.28 12.55 0.71
2.71 (-0.18) 5.32 (-0.16) 17.73 (0.00) 2.47 (2.20) 28.41 (1.88) 10.68 (1.88)
2.96 (-0.43) 7.32 (-2.15)* 17.73 (0.00) 2.20 (2.47)* 30.28 (0.00) 12.55 (0.00)
Croplands 0.01 0.01 0.08 0.07 0.17 0.11 1.57
0.02 (-0.01) 0.03 (-0.02) 0.09 (-0.01) 0.07 (0.00) 0.20 (-0.03) 0.14 (-0.03)
0.01 (0.00) 0.03 (-0.02) 0.06 (0.02) 0.07 (0.00) 0.17 (0.00) 0.11 (0.00)
Total 2014 11.03 27.29 53.17 8.21 100.00 58.68
11.03 (0.00) 27.29 (0.00) 53.17 (0.00) 8.21 (0.00) 100.00 58.68 (0.00)
12.08 (-1.05) 29.22 (-1.94) 51.02 (2.15) 7.40 (0.81) 99.99 (0.01) 58.68 (0.00)
Gross gain 8.44 6.36 35.44 8.14 58.68
8.44 (0.00) 6.36 (0.00) 35.44 (0.00) 8.14 (0.00) 58.68 (0.00)
9.48 (-1.05) 8.29 (-1.94) 33.29 (2.15) 7.33 (0.81) 58.68 (0.00)
G(+) 3.26 0.30 2.00 116.29

Note (applicable to Tables 3a, 3b, 3c and 3d): each table contains both the outputs of gain and loss analyses. Each cell is subdivided into three rows and two columns of numbers. Left column of each cell: the first row contains bolded numbers that represent the actual (observed) proportions of inter-categorical transitions (persistence and transitions) of the landscape. The second row represents the expected percentage of land under random processes of gain (named Expected (+)) calculated using Equation 2, where figures in round parentheses are equal to the observed proportion minus the one expected (named Difference (+)). The third row contains italicized numbers representing the expected proportion of land under random processes of loss (named Expected (-)) calculated using Equation 3, where numbers within round parentheses represent the observed proportion minus the expected one (named Difference (-)). Extreme right column of the table contains the Loss-to-persistence ratio (L(-)) while the extreme row is the Gain-to-persistence ratio (G(+)). Numbers highlighted in gray represent systematic gain transitions; starred numbers represent systematic loss transitions.

The different types of transitions are indicated in Tables 3a, 3b, 3c, and 3d. Difference values between the observed and expected proportions of the landscapes (values in round parentheses in Tables 3) are used to detect random or systematic transitions. Values closer to zero are indicative of random transition while higher values indicate systematic transitions (But a threshold of 0.5% was considered for analytical purpose in this study). Most of the major transitions occurred between the four dominant cover categories, viz. forests, woodlands, savannahs and croplands (Tables 3). Over the three and the overall periods, transitions of forests-savannahs indicated large and negative difference values between observed and expected gains for savannahs, ranging from -1.18% (1972-1987) to -2.52% (2000-2014). These negative values indicate that savannahs did not emerge from forests. Similarly, a
negative difference value (-1.12%) was observed for the overall period 1972-2014, corroborating the general trend of systematic avoidance of the forest replacement under random gain processes for savannahs. The vulnerability of forests to loss was quite evidenced by the loss-to-persistence ratio values higher than one during the first transition and overall periods (Tables 3a and 3d). However, the gain-to-persistence ratio (G+) for forests and savannahs were higher than one, indicating that this category gained much more than persistence over time. In contrast, L- of woodlands was of 1.6 for 2000-2014 and 2.1 for 1972-2014 indicating high vulnerability to loss in recent years in comparison to the two first periods where woodlands were less vulnerable (L- < 1).
Conversely, forest gains were not associated with the replacement of savannahs (negative values of difference between observed and expected gains; -0.74%; -1.31%; -0.44% and -0.18%). In term of gains, there is a systematic mutual avoidance between forests and savannahs, in line with the trend of the transition of forests-savannahs. Meanwhile, forests systematically gained from woodlands (0.74%, 1.38%, and 0.57%) rather than savannahs, which exhibited negative values of observed minus expected gains of forests from savannahs. Mostly, savannahs gains during the four periods emerged from the replacement of woodlands, as indicated by the positive values of difference observed-expected gains (1.15%; 0.76%, 2.30% and 1.12%, respectively for the four periods). Under expected random gain, croplands systematically emerged from savannah losses solely at the rates varying between 0.8-1. In sum, under random process of gain, forests gained more from woodlands than the inverse at relatively very low expected rates (ratio values close to zero). Forests gains during all the periods did not emerge from savannahs, and inversely. Savannahs mostly gained from woodlands rather than from forests while woodlands did not gain from savannahs, except during the period 1987-2000.
On the other hand, as indicated by the positive values of difference between observed and expected losses, forest losses during the three transition periods occurred systematically towards woodlands (0.92%; 0.76%, and 0.98%, for period 1, 2, and 3, respectively) rather than savannahs (-0.88%, -0.66%, -0.72%, respectively for the periods 1, 2 and 3). However, in line with the criteria of systematic change defined in this study (threshold of 0.5%), the overall period did not show a systematic gain of woodlands from forests (0.24% lower than 0.5%). Under these random processes of loss, it is expected that the loss of savannahs was systematically converted into woodlands, except the periods 2000-2014 (-2.11%) and 1972-2014 (-2.15%), indicating that savannahs losses are not associated with woodlands replacement for these latter transition periods. The transition of croplands-natural vegetation showed that when croplands lost, they tended to be converted systematically into savannahs rather than other natural categories. This trend was most acute for the period 2000-2014 with a difference value between observed and expected loss of 0.57% and at an expecting ratio around 0.5. Neither savannahs nor croplands lost systematically into forests, as indicated by the low proportions of transition. Croplands exhibited highest values of G+ indicating the agricultural expansion occurred much more than agricultural land abandonment or conversion into other cover categories. These G+ values for croplands were far higher than L- values indicating that croplands are less vulnerable to conversion to other categories than they gain from other categories.

4.4 Landscape fragmentation and habitat quality change

In general, the landscape metrics (NP and PD) increased for forests (Table 4a), woodlands (Table 4b) and savannahs (Table 4c) from 1972 to 2014 in both PA and UPA. NP and PD for woodlands decreased from 1987 to 2014 in UPA landscapes. Meanwhile savannahs-based NP and PD increased from 1972 to 1987 and decreased constantly from 1987 to 2014. Over time, forest-LPI in PA increased markedly while the inversed trend was observed in UPA. While forest in PA constantly increased, UPA forest cover exhibited a slight improvement in COH in 2014 after a decrease from 1972 to 2000. Marked changes in landscape metrics were observed for 2014 and 2014, especially in UPA.
Table 4a Fragstat-based landscape indices for forest cover
Year Status NP PD LPI COH AI
1972 PA 55 2.20 0.28 78.31 69.96
UPA 5 0.20 0.09 71.80 75.00
1987 PA 148 5.89 2.30 91.71 59.32
UPA 27 1.08 0.05 51.83 36.23
2000 PA 177 7.09 3.72 93.47 70.44
UPA 18 0.72 0.01 27.93 19.15
2014 PA 338 13.47 4.01 93.84 69.19
UPA 62 2.47 0.37 73.45 53.77
References NP ≥ 1 PD > 0 0 < LPI ≤ 100 0 < COH ≤ 100 0 < AI ≤ 100

NP = number of patch; PD = patch density; LPI = largest patch index; COH = patch cohesion; AI = Aggregation index; PA = protected areas; UPA = unprotected areas

Table 4b Fragstat-based landscape indices for woodlands
Year Status NP PD LPI COH AI
1972 PA 28 1.12 47.24 99.59 87.19
UPA 56 2.24 23.47 98.59 84.85
1987 PA 245 9.76 36.96 99.38 68.78
UPA 787 31.35 5.61 92.72 56.09
2000 PA 296 11.86 28.16 98.56 75.43
UPA 368 14.66 2.37 90.14 62.15
2014 PA 478 19.04 3.91 89.46 63.33
UPA 99 3.94 0.06 50.19 40.07
References NP ≥ 1 PD > 0 0 < LPI ≤ 100 0 < COH ≤ 100 0 < AI ≤ 100
Table 4c Fragstat-based landscape indices for savannahs
Year Status NP PD LPI COH AI
1972 PA 75 3.01 24.43 98.51 85.60
UPA 36 1.44 55.18 99.72 88.38
1987 PA 298 11.87 26.61 98.45 76.70
UPA 98 3.90 61.11 99.79 73.75
2000 PA 257 10.30 25.52 95.06 71.86
UPA 40 1.59 65.84 99.84 82.34
2014 PA 135 5.38 37.22 99.07 81.97
UPA 40 1.59 63.27 99.76 88.92
References NP ≥ 1 PD > 0 0 < LPI ≤ 100 0 < COH ≤ 100 < AI ≤ 100

5 Discussion

5.1 Land use/cover mapping and accuracy

The LUC mapping showed that natural vegetation dominated the Mo basin for all the observational dates (1972, 1987, 2000 and 2014). At the early dates of the study period (1972 and 1987), great parts of the Mo landscapes were greener and dominated by woodlands and savannahs. Agricultural patches (360 ha over about 148,600 ha of the whole basin) were merely located around scattered human settlements in the landscapes, especially along the main roads. Despite this dominance of natural vegetation over time, woodlands showed acute areal loss while forests and savannahs substantially increased along the study period. The decline of woodlands and the expansion of savannahs indicated that wood extraction is the main cause of landscape fragmentation and degradation. Similar transformations were observed in adjacent landscapes of the Kara river basin (Badjana et al., 2014) and the upper Northern areas of Togo (Folega et al., 2014b; Folega et al., 2015), where agricultural practices associated with energy wood collection were targeted as prominent factor of natural vegetation loss. Though of small-scale in the study area, agricultural production is regarded as main reason of deforestation and forest degradation in rural areas (Lindstrom et al. (2012).
The reliability of these statistics and trends fundamentally depends on the accuracy of the classified maps measured via the Kappa indices of classification agreement. These Kappa indices (0.6-0.9) and overall accuracies (69%-92%) (Table 5) were high enough and satisfactory for modelling of LUCC (Aguirre-Gutiérrez et al., 2012; Monserud and Leemans, 1992; Leh et al., 2013). This resulted in reliable transition maps, which are the products of the individual LUC maps as suggested by Were et al. (2013). The overall accuracies of the transition maps were of 62.9%, 82.7%, 83.3% and 62.9%, respectively for the 3 and overall transition periods. Further, the producer and user accuracies not only ascertained the above accuracy statistics but also reported the classification errors, which are mostly due to spectral confusion between cover types (Were et al., 2013). In addition, other factors such as complex topography and vegetation patterns in the Mo basin (Diwediga et al., 2015) could drive a certain level of complexity in spectral responses rendering cumbersome the classification processes with less accurate outputs (Lu, 2006). Furthermore, misclassification errors could be introduced via data used for validation since the scale and the resolutions of the reference data for the past LUC maps were relatively poor. Though such data are of poor resolution for small-scale breakdown and might introduced biases (Verburg et al., 2013), they were of great interest for landscape monitoring in the data-scarce area of the Mo basin. Houet et al. (2010) illustrated that the ongoing challenge related to data availability compels to reliance of multisource data for assessing landscape dynamics.
Table 5 Accuacy assessment reports of the produced LUC maps from Landsat archives

5.2 Historical trends and processes of land use/cover change in the Mo basin

Through the analyses of LUCC in the contrasting landscapes (i.e. land protection status), large complex transformations occurred at the whole landscape level. All the LUC types experienced transformations at the early stages of the study period. This significant decrease in the quality of natural vegetation cover was mostly due to agricultural deforestation and wood product extraction (Dourma et al., 2009). In the neighbouring Basin of Kara River with higher human density, similar trends of natural vegetation were recorded with higher proportions of changes (Badjana et al., 2014). Combined with the large network of PA, human concentration on common lands of UPA induced much landscape degradation and fragmentation, expressed throughout the landscape metrics. The increase over time in the natural vegetation metrics (NP and PD) is an indicator of the decrease of landscape homogeneity across the river basin. Further, LPI, COH and AI decreased constantly over time indicating the increasing loss of landscape and habitat connectivity. Decreasing NP and PD of savannah expressed their increasing level of compactness (increasing savannah-specific LPI, AI and COH) and their dominance of the basin (Diwediga et al., 2015 Similar trends observed in LM for both PA and UPA demonstrate the weak level of protective law enforcement regarding the PA in the river basin (Wala et al., 2012; Diwediga et al., 2015), though natural processes had been shown as drivers of vegetation change in West Africa environments (Traore et al., 2015; Le et al., 2012). However, the Mo basin still exhibits high potential of wild landscapes attributable to the low populated compared to other regions of the country (DGSCN, 2010) associated with the high proportion of PA and landscape ruggedness.
Typically, LUCC processes were not only quantitative (i.e. amount or net change) but also qualitative (i.e. location or swapping) (Pontius et al., 2004; Schmitt-Harsh, 2013). Indeed, the swap changes of the natural vegetation categories were higher and greater than the absolute net changes, especially for woodlands and savannahs over time, suggesting that these LUC types experienced much more spatial transformation than perceived by the net change detection, during the transition periods. However, the constant swap decrease was an indicator of a decrease in recovery (due to human pressures) and loss (due especially to less availability of natural lands) of natural vegetation. Sole forest category exhibited quite stable values of swap changes when the extent of analyses is large (1972-2014). For the overall period, the fair swaps for savannahs and woodlands suggest that changes in their locations occurred lesser than as revealed by the transition periods. Thus, the length of transition period (number of years) is important factor affecting the detection of swapping processes, as longer periods tend to mask land transformations occurred in-between the observational dates. This indicates that the landscape is under perpetual dynamics which detection and acuity are consistent with high temporal resolutions (regular and closer observational dates). In the current study, the mapping approach exclusively relied on the landscape cover at the exact passing times of the satellite, which could not really reflect the LUC dynamics (Braimoh and Vlek, 2004). Although it has limits in the verification of the potential changes that occurred in-between the two successive dates (Braimoh and Vlek, 2005; Garedew, 2010), the approach was helpful for analyzing the historical land dynamics in the data scarce case of the Mo basin.

5.3 Landscape dynamics, land legislation and implications for sustainable land management

Regardless of the land protection status and the transition periods, the processes of change affected approximately 50% of the lands in the whole basin between 1972 and 2014. Persistent natural vegetation and processes of vegetation growth are becoming scarce in both PA and UPA. With a focus on the efficiency of law enforcement on biodiversity conservation, protection law played an important role for land conservation, though vegetation loss occurred in the nature reserves and parks. In fact, illegal incursions were noticed in PA, especially the Wildlife Reserve of Aledjo (Wala et al., 2012) and the national park of Fazao-Malfakassa, especially in the neighborhoods of Alombe and along Mo and Bouzalo rivers since 1978 (Aboudou, 2012). Despite the conservation measures, these types of wilderness loss were also reported in other PA of Northern Togo with more acute levels of degradation (Dimobe et al., 2014Folega et al., 2010; Folega et al., 2012). The fortunate situation of the PA in the Mo basin is attributable to the private status devolved to FMN Park especially during the period 1990-2015. Though it occurs at low rate compared to the situation outside PA, the decrease of natural vegetation should be of major concern to counter-act degradation scenarios similar to those of other PA of Togo (Folega et al., 2014; Wala et al., 2012). Nevertheless, the relative low rates of occurrence compared to the situation in UPA are indicators of the importance of PA in conserving land resources (Fan et al., 2013; Porter-Bolland et al., 2012). However, national government efforts are explicit towards the law reinforcement for PA. This is the case of FMN Park, which regained much more attention through two major concrete actions: (i) the assignment of the park to a private monitor and manager for 25 years (1990-2015) and (ii) redefinition of new boundaries that take somehow into account rural people needs.
The systematic transition of woodlands into savannahs and the expansion of agricultural land from savannah should be key indicators of land conservation challenges in the area. Without effective law enforcement within PA and conservation efforts and sensitisation in UPA, further decline in natural vegetation will continue affecting the Mo river landscapes. With regard to these trends, pro-active approaches need to be undertaken to counter-act the current nature and spatial patterns of degradation and address the underlying factors of these change (Lambin et al., 2003). This study showed that timely acute and spatial information could be drawn from satellite images at landscape level to detect hotspot areas where efforts should concentrate for conservation and restoration processes. Efforts should concentrate to reduce or impede the processes inducing land cover quality loss in both PA and UPA through effective institutional operationalisation at local and national level regarding land question to ensure a holistic framework for sustainable land use and conservation. This suggests a clear definition of land tenure regime that addresses major challenges regarding the management of common resources, especially outside PA where land resource allocation decision-making is uncontrolled and guided by individual household needs. Institutional and political settings towards land governance is a key cornerstone that should be clearly set as basis for sustainability (Paudel et al., 2015). The revision of current national policy regarding the redefinition of PA network in Togo needs careful consideration of the real needs of surrounding communities to avoid breaking the law (Vedeld et al., 2012; Tumusiime et al., 2011). Meanwhile, common lands outside PA might gain much more attention towards sustainable use in order to avoid issues related to common resources. Furthermore, focus should be given to the definition of a clear land information system and the role of all stakeholders from formal to informal as well as public-private partnership in rural communities. Reducing the dependence of households on forest products tend to be of positive effect in the process of landscape conservation. Therefore socio-economic conditions of rural households should be strengthened through diversification of income-generated activities. These options could be the promotion of orchards and cashew plantations, and agroforestry systems in current farmlands which could support social-ecological development (Mbow et al., 2014b; Mbow et al., 2014a). REDD+ implications of such strategies could be source of incentives for reducing land decline in the area (Mattsson et al., 2012; Nakakaawa et al., 2010).

6 Conclusion

Since there is a lack of information on the historical patterns of LUC in developing countries, the reliance on proxies such as the earth observation archives seems to be interesting to monitor landscape transformation and its factors. In this study, the use of historical Landsat data provided an assessment of the process of vegetation cover degradation in rural areas of the Mo river basin. The study provided more light on the extents, locations and rates of rural land transformation, mainly deforestation and forest degradation (DFD). The results revealed that the observed trend in the study area are the intensive decline of woodlands associated with an increase in savannah and forest extents as well as cultivated areas. Despite the important network of PA in the area, natural vegetation showed a decreasing trend from 99% in 1972 to 91% in 2014. This significant change in the natural land cover quality and extent were due to agricultural expansion and wood extraction in both PA and UPA. The trend in the landscape dynamics indicated a savannisation process throughout the river basin though there was an improvement of forest cover. The net balance of natural vegetation changes indicated an overall loss mostly due to woodland decline. However, there Mo basin still provides wide landscapes for wilderness protection and natural landscape integrity (Diwediga et al., 2015). This information may be of practical aspect in guiding managers and policy makers for reversing the loss of natural vegetation through the formulation and implementation of new strategies for the integrated land management. Restoring degraded areas and increasing awareness to maintain the quantity and quality of the natural vegetation are essential for supporting livelihood, biological conservation and global climate mitigation. With regard to the ecological and economic importance of the Mo river basin and its surrounding lands, attempting focus towards the reduction of vegetation decline, should be given through further analyses of the major trajectories of conversion, the underlying factors as well as the direct drivers, and the modelling of possible pathways of sustainable development of the basin.

The authors have declared that no competing interests exist.

References

Publishing order | Descend order by publishing year | Descend order by cited within
[1]
Aboudou M, 2012. Etat des lieux de l’occupation du sol dans et autour du parc national Fazao-Malfakassa. MERF-Togo/Fondation Franz Weber.

[2]
Afidégnon D, Guelly K A, Kokou Ket al., 2003. Carte de la végétation du Togo + Notice explicative. Projet Campus 96.348.110. Toulouse: L.B.E.V., Lome & L.E.T.

[3]
Aguirre-Gutiérrez J, Seijmonsbergen A C, Duivenvoorden J F, 2012. Optimizing land cover classification accuracy for change detection, a combined pixel-based and object-based approach in a mountainous area in Mexico.Applied Geography, 34: 29-37.

[4]
Badjana H M, Selsam P, Wala Ket al., 2014. Assessment of land-cover changes in a sub-catchment of the Oti basin (West Africa): A case study of the Kara River basin.Zbl. Geol. Paläont. Teil I, Jg, 1: 151-170.This study is presenting results from the analysis of land-cover dynamics in the Kara River basin from 1972 to 2000 using remote sensing and geographic information system (GIS) techniques. Multi-temporal satellite images - Landsat MSS (1972), TM (1987) and ETM+ (2000) were processed using object-oriented classification and postclassification comparison methods. The classification results showed a good agreement with overall accuracies of 77.1% for 1972, 85.34% for 1987 and 88.22% for 2000, the respective Kappa statistics being 0.7, 0.81 and 0.85. Change assessment showed that the basin has experienced important changes in land cover with significant decrease in natural vegetation. Agricultural extension and deforestation appear to be the dominant driving forces. In fact, agricultural land has doubled between 1972 and 2000 by increasing from 19.45% to 43.59% of the total basin area while savannah decreased from 63.41% to 45.19%. Forest land increased by 1.63% from 1972 to 1987 but showed a decrease of 5.87% from 1987 to 2000, while woody savannah decreased by 3.59% and human settlements increased during the same period. The analysis of land cover transition showed that important changes have occurred between 1987 and 2000. The results of this study will be useful for follow-up research such as hydrological and landscape processes modelling in the basin but also many applications including the biomass assessment for greenhouse gases inventories, the REDD+ (Reducing emissions from deforestation and forest degradationplus) project, the national programs for protected areas redefinition, vegetation resources conservation and sustainable land management, the vulnerability and adaptation studies etc. Nevertheless, further investigation integrating climatic, hydrological, spatial statistics and socio-economic models would help to better understand and quantify the relationships between different driving forces and their contributions to the changes.

[5]
Balthazar V, Vanacker V, Molina Aet al., 2015. Impacts of forest cover change on ecosystem services in high Andean mountains.Ecological Indicators, 48: 63-75.Land conversion affects the delivery of ecosystem goods and services. In this study, we used a 50 years time series of land cover maps to assess the potential impacts of forest cover changes on ecosystem services. A multi-source data integration strategy was followed to reduce inconsistencies in land cover change detection that result from the comparison of historical aerial photographs and satellite images. Our forest cover change analysis highlighted a shift from net deforestation to net reforestation in the early 1990s, consistent with the forest transition theory. When taking the nature of forest cover changes into account, our data show that the areal increase of the forested area was not associated with an improvement in ecological conditions. The overall capacity of the landscape to deliver ecosystem services dropped steadily by 16% over the 50 years- study period. Conversion of native forests to agricultural land was associated with the strongest decline in ecosystem services. Conversion of natural grasslands into pine plantations mostly led to negative and probably irreversible impacts on the delivery of ecosystem services. Conversion of degraded agricultural lands into pine plantations led to an improvement in ecological conditions. An effective spatial targeting of forestation programs has the potential to maximize the environmental benefits that forest plantations may offer while minimizing their environmental harm.

DOI

[6]
Bicheron P, Defourny P, Brockmann C et al., 2008. GLOBCOVER: Products Description and Validation Report. In: PROJECT G (ed.). MEDIAS-France/POSTEL.

[7]
Biro K, Pradhan B, Buchroithner Met al., 2013. Land use/land cover change analysis and its impact on soil properties in the northern part of Gadarif Region, Sudan.Land Degradation & Development, 24: 90-102.ABSTRACTSeveral decades of intensive dry land farming in the Gadarif region, located in the Eastern part of Sudan, has led to rapid land use/land cover (LULC) changes mainly due to agricultural expansion, government policies and environmental calamities such as drought. In this paper, an attempt has been made to analyse and monitor the LULC changes using multi-temporal Landsat data for the years 1979, 1989 and 1999 and ASTER data for the year 2009. In addition, efforts were made to discuss the impact of LULC changes on the selected soil properties. For this, a post-classification comparison technique was used to detect LULC changes from satellite images. Primarily, three main LULC types were selected to investigate the properties of soil, namely, cultivated land, fallow land and woodland. Moreover, soil samples were also collected at two depths of surface soil from ten sample plots for each of the LULC type. For these soil samples, various soil properties such as texture, bulk density, organic matter, soil pH, electrical conductivity, sodium adsorption ratio, phosphorous and potassium were analysed. The results showed that a significant and extensive change of LULC patterns has occurred in the last three decades in the study area. Further, laboratory tests revealed that soil properties were significantly affected by these LULC changes. The change of the physical and chemical properties of the soil may have attributed to the changes in the LULC resulting in land degradation, which in turn has led to a decline in soil productivity. Copyright 漏 2011 John Wiley & Sons, Ltd.

DOI

[8]
Braimoh A K, 2004. Modeling land-use change in the Volta Basin of Ghana.Abstract Land change studies in a 5,400km2 area within the Volta Basin of Ghana were carried out using satellite image and socio-economic analyses. The dominant change process was conversion of natural vegetation to cropland at an annual rate of 5%. Reversible land change trajectories involving accumulation of woody biomass indicate a certain level of rainfall-induced ecosystem resilience. Linear and logistic regressions identified agricultural land suitability, distance from main market and localities, child-dependency ratio and population density as the main drivers of change. Policy measures that would relieve human pressure on vegetation resources, guarantee food supply and promote commercialization of agriculture are suggested.

[9]
Braimoh A K, Vlek P L G, 2004. Land-cover change analyses in the Volta Basin of Ghana.Earth Interactions, 8: 1-17.Multitemporal Landsat Thematic Mapper (TM) images for 1984, 1992, and 1999 were used to map and detect land-cover changes in a 5400-km2 area within the Volta Lake basin of Ghana. The most dominant land-cover change was the conversion of natural vegetation to cropland, which occurred at an annual rate of 5%. While the data suggest an increase in human pressure, reversible change in woodland and grassland occurred in 4% and 2% of the landscape, respectively. A higher proportion of reversible land-cover changes relating to fallow agriculture occurred in about 14% of the landscape, whereas a higher overall increase in woody biomass (10%), compared to an overall decrease of 9%, indicates a certain level of rainfall-induced resilience in the ecosystem. Further research is needed to quantitatively evaluate the mechanisms enhancing vegetation recovery in dryland areas.

DOI

[10]
Braimoh A K, Vlek P L G, 2005. Land-cover change trajectories in Northern Ghana.Environmental Management, 36: 356-373.Land-cover change trajectories are an emergent property of complex human-nvironment systems such as the land-use system. An understanding of the factors responsible for land change trajectories is fundamental for land-use planning and the development of land-related policies. The aims of this study were to characterize and identify the spatial determinants of agricultural land-cover change trajectories in northern Ghana. Land-cover change trajectories were defined using land-cover maps prepared from Landsat Thematic Mapper dataset acquired in 1984, 1992, and 1999. Binary logistic regression was used to model the probability of observing the trajectories as a function of spatially explicit biophysical and socioeconomic independent variables. Population densities generally increased along the continuum of land-use intensity, whereas distance from market and roads generally decreased along this continuum. Apparently, roads and market serve as incentives for settlement and agricultural land use. An increase in population density is an important spatial determinant only for trajectories where the dominant change process is agricultural extensification. A major response to population growth is an increase in cultivation frequency around the main market. Agricultural intensification is highly sensitive to accessibility by roads. The increase in land-use intensity is also associated with low soil quality. These results suggest the need for policies to restore soil fertility for agricultural sustainability. The models also provide a means for identifying functional relationships for in-depth analyses of land-use change in Ghana.

DOI PMID

[11]
Carmona A, Nahuelhual L, 2012. Combining land transitions and trajectories in assessing forest cover change.Applied Geography, 32: 904-915.One restriction of landscape studies is that land use and cover change is often regarded as irreversible. A highly dynamic landscape in southern Chile was selected to show that forest cover change involves a series of complex transitions and trajectories. Using Landsat images from 1976, 1985, 1999 and 2007 an in-depth analysis of the transition matrix was conducted to separate random and systematic transitions which were grouped into trajectories using a pixel-history approach. Main trajectories were linked to fragmentation indices and farming systems through cluster analysis. Of the 247 trajectories identified, old growth forest persistence comprised 22% of the landscape, whereas deforestation trajectories comprised 20.9% and were mostly composed of changes from old growth forest to shrubland (13.9%). Trajectories of forest degradation from old growth to secondary forest comprised 19.7% of the landscape. The periods 1976–1985 and 1999–2007 concentrated the most systematic deforestation and degradation transitions. In turn, random transitions predominated between 1985 and 1999, probably in response to economic factors that acted suddenly on the landscape during the 80’s, such as the woodchip export and aquaculture booms. A close relationship between landscape fragmentation and the proportion of systematic transitions and farming systems was found; specifically, the highest entropy indices occurred in clusters which exhibited the lowest proportion of systematic transitions and the highest proportion (>70%) of peasant agricultural systems. Understanding the complexity of forest cover change trajectories is relevant for improving the prediction of possible landscape evolutions and establishing landscape management priorities.

DOI

[12]
Castro A J, Martín-López B, López Eet al., 2015. Do protected areas networks ensure the supply of ecosystem services? Spatial patterns of two nature reserve systems in semi-arid Spain.Applied Geography, 60: 1-9.Protected areas are essential for conserving biodiversity, and these lands have traditionally been set aside for this purpose alone. However, the increasing global demand for agricultural and forestry commodities creates conflict and tradeoffs between dedicating land for conservation versus food production. Efforts to set aside new lands for biodiversity conservation are compromised by the globally rising demand, creating trade-offs between lands dedicated to conservation versus food production. Ecosystem services are the benefits that humans obtain from ecosystems. Recent studies suggest that protected areas provide social and economic benefits that can be used to build political support and raise funds for conservation. We analyzed the capability of current protected area networks in the semi-arid region of Spain to provide intermediate regulating services (habitat preservation for threatened species, climate regulation, erosion control and water flow maintenance) to support the final provisioning service of cultivated crops to support local communities. We found that existing networks of protected lands supply considerable quantities of ecosystem services, in particular carbon stocks and groundwater recharge. Our results demonstrate that the integration of systematic analyses of ecosystem services gaps in protected area planning could contribute substantially to safeguarding ecosystem services and biodiversity jointly. However, our study also reveals substantial differences in intermediate ecosystem services supplied by different of protected areas networks, with category VI areas (Natura-2000 sites) generally showing the highest potential for ecosystem services supply. This demonstrates the important role of Natura-2000 sites for preserving regulating services in the European semi-arid region.

DOI

[13]
Damnyag L, Saastamoinen O, Blay Det al., 2013. Sustaining protected areas: Identifying and controlling deforestation and forest degradation drivers in the Ankasa Conservation Area, Ghana.Biological Conservation, 165, 86-94.Although protected areas in Africa contain possibly the highest repositories of carbon and thus can play a role in mitigating the effects of climate change through carbon sequestration, they are threatened due to increasing levels of deforestation and forest degradation (DFD). However, little information is available on the on-site causes of DFD in these areas. This paper estimates the levels of DFD and identifies the drivers in the Ankasa Conservation Area (ACA) in Ghana as a case study. A survey was used to identify both direct and underlying factors that promote the DFD. The extent of deforestation was estimated using satellite images. The survey data were analyzed using rankings and ordinal logistic regression techniques, while digital image classification and change detection were used to analyze land cover changes. The results show that DFD occurred at a higher rate in the periphery of the ACA compared to the core-protected and the farthest areas. Agricultural and wood harvesting activities were the main direct causes of DFD. Poverty and large in-migrations of people for cocoa farming were important underlying economic and population growth factors. To address these problems and enable ACA to contribute more to biodiversity conservation and climate change mitigation, the community resource management institutions should be fully adopted and strengthened and priority given to livelihood improvement and ecosystem services provision in the periphery of the ACA.

DOI

[14]
DGSCN2010. Recensement general de la population et de l'habitat. Resultats definitifs. Lome: Togo.

[15]
Diallo Y, Hu G, Wen X, 2010. Assessment of land use cover changes using NDVI and DEM in Puer and Simao counties, Yunnan Province, China.Report and Opinion, 2: 7-16.Abstract Change detection is the technique, which is used for the assessment of resources, where multi-date images are compared to find out the type and amount of change have occurred. The various applications of change detection are in agricultural, hydrological, forestry, environmental and ecological field. The rapid growth of Puer-Simao area and its agricultural area have probably resulted very rapidly and have had an unfavorable effect on the environment and therefore multi-temporal Landsat TM imagery for assessment land cover change has proven to be the best tool in this study. The performance of image classifiers that utilize only the remote sensing data may deteriorate, especially in mountainous regions, payable to the presence of shadows. In our study, a multisource classification approach to map land cover in Puer_Simao counties with high mountain peaks having elevations up to 2800 m above mean sea level has been adopted. Remote sensing data from Landsat TM image along with NDVI and DEM data layers have been used to perform multi-source classification using Maximum Likelihood Classifier. The change detection method used was NDVI differencing. From the results the forest or shrub land and Barren land cover types have decreased by about 6% and 23% from 1990 to 1999 respectively, while agricultural land, built-up and water areas have increased by about 19%, 4 % and 7% respectively. During the study period the change detected in reference with the location of LUC types in the different regions with up mentioned altitude and slope ranges was not significantly important. (World Rural Observations 2009;1(2):1-11). ISSN: 1944-6543 (print); ISSN: 1944-6551 (online).

[16]
Dourma M, Wala K, Bellefontaine Ret al., 2009. Comparaison de l’utilisation des ressources forestières et de la régénération entre deux types de forêts claires à Isoberlinia au Togo.Bois et Forets des Tropiques, 302: 5-19.

[17]
Ellis C E, 2011. Anthropogenic transformation of the terrestrial biosphere.Phil. Trans. R. Soc A, 369: 1010-1035.Human populations and their use of land have transformed most of the terrestrial biosphere into anthropogenic biomes (anthromes), causing a variety of novel ecological patterns and processes to emerge. To assess whether human populations and their use of land have directly altered the terrestrial biosphere sufficiently to indicate that the Earth system has entered a new geological epoch, spatially explicit global estimates of human populations and their use of land were analysed across the Holocene for their potential to induce irreversible novel transformation of the terrestrial biosphere. Human alteration of the terrestrial biosphere has been significant for more than 8000 years. However, only in the past century has the majority of the terrestrial biosphere been transformed into intensively used anthromes with predominantly novel anthropogenic ecological processes. At present, even were human populations to decline substantially or use of land become far more efficient, the current global extent, duration, type and intensity of human transformation of ecosystems have already irreversibly altered the terrestrial biosphere at levels sufficient to leave an unambiguous geological record differing substantially from that of the Holocene or any prior epoch. It remains to be seen whether the anthropogenic biosphere will be sustained and continue to evolve.

DOI PMID

[18]
Ellis C E, Fuller D Q, Kaplanet al., 2013. Dating the Anthropocene: Towards an empirical global history of human transformation of the terrestrial biosphere. Elementa:Science of the Anthropocene, 18(1): 1-6.Abstract Human use of land is a major cause of the global environmental changes that define the Anthropocene. Archaeological and paleoecological evidence confirm that human populations and their use of land transformed ecosystems at sites around the world by the late Pleistocene and historical models indicate this transformation may have reached globally significant levels more than 3000 years ago. Yet these data in themselves remain insufficient to conclusively date the emergence of land use as a global force transforming the biosphere, with plausible dates ranging from the late Pleistocene to AD 1800. Conclusive empirical dating of human transformation of the terrestrial biosphere will require unprecedented levels of investment in sustained interdisciplinary collaboration and the development of a geospatial cyberinfrastructure to collate and integrate the field observations of archaeologists, paleoecologists, paleoenvironmental scientists, environmental historians, geoscientists, geographers and other human and environmental scientists globally from the Pleistocene to the present. Existing field observations may yet prove insufficient in terms of their spatial and temporal coverage, but by assessing these observations within a spatially explicit statistically robust global framework, major observational gaps can be identified, stimulating data gathering in underrepresented regions and time periods. Like the Anthropocene itself, building scientific understanding of the human role in shaping the biosphere requires both sustained effort and leveraging the most powerful social systems and technologies ever developed on this planet.

DOI

[19]
Ellis E C, 2013. Sustaining biodiversity and people in the world's anthropogenic biomes.Current Opinion in Environmental Sustainability, 5: 368-372.Humans have reshaped more than three quarters of the terrestrial biosphere into anthropogenic biomes (anthromes), embedding substantial areas of remnant and recovering novel ecosystems within the agricultural and settled landscapes that sustain human populations. The need to conserve biodiversity in anthromes is increasingly recognized as critical, as anthromes have largely replaced wildlands in Earth's most biodiverse and productive regions, and novel ecosystems now cover nearly twice the global area of wildlands. Extinction rates may still be increasing. Nevertheless, recent studies indicate that under appropriate conditions, most native taxa may be sustainable within anthromes while at the same time increasing anthrome productivity in support of human populations.

DOI

[20]
Ern H, 1979. Die Vegetation Togos, Gliederung, Gefährdung, Erhaltung.Willdenowia, 9: 295-315.The territory of Togo (W. Africa) is subdivided, by means of biogeographical criteria, into five vegetational zones. On the base of field experience examples of the most characteristic and dominant plant communities of every zone are given. The actual, continued process of severe destruction and degradation of the plant cover is emphasized. The main factors responsible for this destruction are: production of charcoal and firewood, land clearing for agricultural purposes, and burning of savannah grassland during the dry season. The set-up of a network of conservation-orientated nature reserves is advocated, together with the foundation of a national Botanic Garden which could serve as an administrative and scientific conservational centre. /// Unter Berücksichtigung biogeographischer Faktoren wird eine Gliederung Togos in 5 02kologische Zonen vorgeschlagen. Für jede Zone werden die jeweils charakteristischen und dominierenden Vegetationsformationen beschrieben und Artenlisten aufgeführt. Im zweiten Abschnitt wird der augenblickliche Zustand der togoischen Vegetation geschildert, die im ganzen Lande einer fortschreitenden Degradierung unterliegt. Die Hauptfaktoren dieser Vegetationszerst02rung sind: Brennholzgewinnung und K02hlerei, Rodung für landwirtschaftliche Zwecke und absichtlich gelegte Grasbr01nde. Im Schlu08abschnitt wird die Notwendigkeit von Schutzma08nahmen für die stark gef01hrdete Pflanzendecke Togos betont und vorgeschlagen, gem0108 der Schlu08resolution der internationalen Konferenz "Survival or Extinction" vom September 1978 in Kew in Togo ein Netz von Vegetations-Schutzgebieten zu errichten, als deren Zentrum ein bei Lomé zu begründender Botanischer Garten zu fungieren h01tte. /// En tenant compte des données biogéographiques, le Togo est divisé en cinq zones écologiques. Pour chaque zone les formations végétales les plus caractéristiques et dominantes sont décrites en fournissant des listes d'espèces typiques. Puis nous présentons une vue d'ensemble sur l'état actuel de la végétation togolaise en soulignant que celle ci est soumise, dans tout le territoire, à une dégradation forte et continue. Les agents de cette dégradation sont: Production du bois de chauffage et du charbon, défrichements agricoles et mise en feu des savanes dans la saison sêche. Enfin nous proposons l'installation d'un réseau de réserves écologiques pour la végétation du Togo avec la création d'un Jardin Botanique près de Lomé qui doit servir comme centre administratif et scientifique de ce réseau de réserves.

[21]
Fan Z, Zhang X, Li Jet al., 2013. Land-cover changes of national nature reserves in China.Journal of Geographical Sciences, 23(2): 258-270.For preventing ecosystem degradation, protecting natural habitats and conserving biodiversity within the habitats, 2588 nature reserves have been established in China at the end of 2010. The total area is up to 149.44 million ha and covers over 15% of Chinese terrestrial surface. Land-cover change, as the primary driver of biodiversity change, directly impacts ecosystem structures and functions. In this paper, 180 National Nature Reserves (NNRs) are selected and their total area is 44.71 million ha, accounting for 29.9% of all NNRs in China. In terms of the ecosystem characteristics and their major protected object, all selected NNRs are classified into 7 types. A Positive and Negative Change Index of Land-cover (PNCIL) was developed to analyze the land-cover change of each NNRs type from the late 1980s to 2005. The results show that the land-cover of all selected NNRs types have degradated to a certain degree except the forest ecosystem reserves with a decreasing rate, but the rate of degradation alleviated gradually. The mean positive and negative change rates of land-cover in all core zones decreased by 0.69% and 0.16% respectively. The landscape pattern of land-cover in the core zones was more stable than that in the buffer zones and the experimental zones. Furthermore, the ecological diversity and patch connectivity of land-cover in selected NNRs increased generally. In short, the land-cover of 180 selected NNRs in China had a beneficial change trend after NNRs established, especially between 1995 and 2005.

DOI

[22]
FAO,1996. Forest Resources Assessment. Survey of Tropical Forest Cover and Study of Change Processes. Rome, Italy, 1990.

[23]
FAO, 2015. Global Forest Resource Assessment 2015. How are the world's forests changing? Rome, Italy.

[24]
Farooq A, 2012. Detection of change in vegetation cover using multi-spectral and multi-temporal information for District Sargodha, Pakistan.Soc. & Nat., Uberlândia, 24: 557-572.ABSTRACT Detection of change is the measure of the distinct data framework and thematic change information that can direct to more tangible insights into underlying process involving land cover and landuse changes. Monitoring the locations and distributions of land cover changes is important for establishing links between policy decisions, regulatory actions and subsequent landuse activities. Change detection is the process that helps in determining the changes associated with landuse and land cover properties with reference to geo-registered multi-temporal remote sensing information. It assists in identifying change between two or more dates that is uncharacterized of normal variation. After image to image registrations, the normalized difference vegetation index (NDVI), the transformed normalized difference vegetation index (TNDVI), the enhanced vegetation index (EVI) and the soil-adjusted vegetation index (SAVI) values were derived from Landsat ETM+ dataset and an image differencing algorithm was applied to detect changes. This paper presents an application of the use of multi-temporal Landsat ETM+ images and multi-spectral MODIS (Terra) EVI/NDVI time-series vegetation phenology metrics for the District Sargodha. The results can be utilized as a temporal land use change model for Punjab province of Pakistan to quantify the extent and nature of change and assist in future prediction studies. This will support environmental planning to develop sustainable landuse practices.

DOI

[25]
Folega F, Wala K, Zhang Cet al., 2011. Woody vegetation of protected areas in northern Togo. Cases of Barkoissi, Galangashi and Oti-Keran: Ecological and structure analyses of plant communities.For. Stud. China, 13: 23-35.

[26]
Folega F, Woegan Y A, Dourma Met al., 2015. Long term evaluation of green vegetation cover dynamic in the Atacora Mountain Chain (Togo) and its Relation to Carbon Sequuestration in West Africa.J. Mt. Sci., 12: 921-934.研究在趋于在一个 24 年的时期期间估计植被盖子的变化的 Togo 在 Atacora 山脉被做。它也试图评估生活的主要生产率(NPP ) 在一样的时期上种的网动态。盖住三个不同时期(1987， 2000，和 2011 ) 的 Landsat 形象被预先处理象 gapfilling 一样改正大气、公制辐射仪的参数 2011 SCL 离开想象。然后，象 NDVI (规范的差别植被索引) 那样的植被索引， SR (简单比率植被索引) ， SAVI (调整土壤的植被索引) ，和房屋(carnegie-ames- stanford 途径) 当模特儿因为 NPP 在掩盖学习区域以后在这些图象上被使用。结果在植被盖子显示出安静减少。植被损失比从 1987 ～ 2000 从 2000 ～ 2011 是更重要的，并且人为的活动能作为植被损失的主要原因被认为。由 NPP 计算的生物资源评价也在时间显示出减少。类似于植被盖子的变化，生态系统网生产率与 2000 和 1987 相比在 2011 是很低的。包括它在碳下沉的潜力，看起来，植被的一般健康条件否定地在这个区域被影响，它已经在下面威胁。借助于有效技术的这些生态系统的永久监视能提高持续管理工具在从采伐森林和森林降级(整顿) 减少排出物的框架。

DOI

[27]
Folega F, Zhang C, Woegan Y Aet al., 2014a. Structure and ecology of forest plant community in Togo.Journal of Tropical Forest Science, 26: 225-239.Vegetation was sampled in 86 plots of riparian forest and dry forests of northern Togo. The recorded data were subjected to floristic and quantitative ecological analyses in order to describe the main woody plant communities and their size structure. The plant groupings were determined by detrended correspondence analysis (DCA) and defined through the indicator value method. Alpha diversity index, and tree species structure features were computed for forest woody plant groupings. Results classified four woody communities. Two of them belonged to riparian forest while the other two, dry forest. A total of 77 plant species consisted of woody species. Pterocarpus santalino茂des and Anogeissus leiocarpus the most important species based on their indicator values. The diversity indices of P. santalino茂des, Eugenia kerstingii and A. leiocarpus and Vitellaria paradoxa plant communities were significant and indicated a wide distribution of species in the area, while their structural features reflected those of forests characteristic of the Sudanian climatic zones. In general, the natural regeneration rate among the plant groupings was satisfactory and quiet similar to those found in previous works in this region.

DOI

[28]
Folega F, Zhang C, Zhao Xet al., 2014b. Satellite monitoring of land-use and land-cover changes in northern Togo protected areas.Journal of Forestry Research, 25: 385-392.Remote-sensing data for protected areas in northern Togo, obtained in three different years (2007, 2000, and 1987), were used to assess and map changes in land cover and land use for this drought prone zone. The normalized difference vegetation index (NDVI) was applied to the images to map changes in vegetation. An unsupervised classification, followed by classes recoding, filtering, identifications, area computing and post-classification process were applied to the composite of the three years of NDVI images. Maximum likelihood classification was applied to the 2007 image (ETM+2007) using a supervised classification process. Seven vegetation classes were defined from training data sets. The seven classes included the following biomes: riparian forest, dry forest, flooded vegetation, wooded savanna, fallows, parkland, and water. For these classes, the overall accuracy and the overall kappa statistic for the classified map were 72.5% and 0.67, respectively. Data analyses indicated a great change in land resources; especially between 1987 and 2000 probably due to the impact of democratization process social, economic, and political disorder from 1990. Wide-scale loss of vegetation occurred during this period. However, areas of vegetation clearing and regrowth were more visible between 2000 and 2007. The main source of confusion in the contingency matrix was due to heterogeneity within certain classes. It could also be due to spectral homogeneity among the classes. This research provides a baseline for future ecological landscape research and for the next management program in the area.

DOI

[29]
Fontodji K J, Atsri H, Adjonou K et al., 2011. Impact of charcoal production on biodiversity in Togo (West Africa). In: (ED.), D. J. L.-P. (ed.).The Importance of Biological Interactions in the Study of Biodiversity. InTech.

[30]
Gaia V, 2011. An epoch debate.Science, 334: 32-37.

[31]
Garedew E, 2010. Land-use and land-cover dynamics and rural livelihood perspectives, in the semi-arid areas of Central Rift Valley of Ethiopia [D]. Swedish University of Agricultural Sciences.

[32]
Griffiths P, Kuemmerle T, Kennedy Ret al., 2012. Using annual time-series of Landsat images to assess the effects of forest restitution in post-socialist Romania.Remote Sensing of Environment, 118: 199-214.78 We quantified rapid and gradual forest changes since the mid-1980s. 78 The open Landsat archive enables dense time series for process-oriented analysis. 78 Collapse of socialism reduced annual timber exploitation in Romanian Carpathians. 78 During three restitution waves private owners increasingly exploited forests. 78 Annual time series allow assessing how institutional shocks affect land systems.

DOI

[33]
Gutiérrez Angonese J, Grau H R, 2014. Assessment of swaps and persistence in land cover changes in a subtropical periurban region, NW Argentina.Landscape and Urban Planning, 127: 83-93.A detailed spatial analysis of land cover changes was carried out in the periurban area of Great San Miguel de Tucum谩n and Sierra de San Javier, subtropical Argentina. Post-classification comparison of land cover maps of 1972 and 2010 was used to quantify the level of persistence, net gains, losses and swaps among urban, natural vegetation, and agriculture categories; framed in a hierarchical land use/cover classification. The spatial distribution of land cover changes was related to environmental and socio-economic variables. The overall land cover change pattern of -eriurban forest transition- was characterized by urban expansion, agriculture adjustment and associated forest recovery. Montane forests showed a net increase of 10%, expanding over mountain grasslands, which in turn lost 66% of their original area. Dry forests experienced high levels of swaps, being relocated into more humid areas and further away from access roads. Simultaneously, herbaceous agriculture was concentrated in flat areas more suitable for modern mechanized agriculture. In the foothills of the San Javier range, urban areas tripled their original extension replacing fertile agricultural lands (mainly sugar cane). Forest recovery and land-use intensification patterns are usually considered as an opportunity for conservation of biodiversity and ecosystem services. However, these new forests are characterized by the abundance of exotic species with little known ecological properties. Also, the replacement of highly productive agriculture by urban developments, and of natural montane grasslands by forests, imply negative changes in terms of agriculture production, the conservation of grassland biodiversity and landscape configuration with high recreational value.

DOI

[34]
He J, Lang R, Xu J, 2014. Local dynamics driving forest transition: Insights from upland villages in Southwest China.Forests, 5: 214-233.China has experienced extensive forest transition, from net deforestation to net forestation. Existing theories have highlighted economic growth, the intensification of agriculture and forest scarcity as the pathways of this transition, and studies, in particular from China, have also highlighted the contribution of a huge state afforestation program and the improved implementation and enforcement of forest protection policy and law. However, few studies have paid attention to local dynamics to provide a contextualized understanding of how forest transition has taken place at the local level and the significance of local factors in this change. This paper examines forest transition pathways in two villages in China. We consider the historical perspective and compare their local dynamics and variations to reach an understanding of the process of forest recovery at the local level. The results show that state forestry policies, including afforestation policy and tenure reform, arguably contribute to forest increase, while local processes including livelihood change and institutional factors play a key role in driving and shaping forest transition. We argue that there is a need for local-level studies and recommend including local institutions in forest transition analysis, contextualizing the socio-ecological interactions within the broader concept of political economy.

DOI

[35]
Houet T, Verburg P H, Loveland T R, 2010. Monitoring and modelling landscape dynamics.Landscape Ecology, 25: 163-167.Introduction Changes in land cover and land use are among the most pervasive and important sources of recent alterations of the Earth’s land surface. Land changes significantly affect key aspects of Earth System functioning, for example in contributing to local and regional climate change as well as to global climate warming, impacting biodiversity and water quality, or increasing soil degradation (Vitousek et al. 1997 ; Stohlgren et al. 1998 ; Houghton et al. 1999 ). Landscape dynamics studies integrating human–environment interactions and related to environmental issues have become increasingly important. Over the years these studies moved away from a focus on detecting and identifying land use and land cover changes (Loveland et al. 1999 , 2002 ; Lambin et al. 2001 ) and understanding driving forces of landscape changes (Bürgi et al. 2004 ; Antrop 2005 ) to modelling present land systems for predicting land cover changes (Veldkamp and Lambin 2001 ; Corgne et al. 2004 ; Hepinstall et al. 2008 ) ...

DOI

[36]
IGN, 1986. Cartes topographiques 1/200 000, 1 res Editions, Feuilles Sokode et Kara. France and D.C. N. C. & M.E.M.P.T., Togo.

[37]
Jiang J, Zhou J, Wu Het al., 2005. Land cover changes in the rural-urban interaction of Xi’an region using Landsat TM/ETM data.Journal of Geographical Sciences, 15(5): 423-430.

[38]
Kang N, Sakamoto T, Imanishi Jet al., 2013. Characterizing the historical changes in land use and landscape spatial pattern on the oguraike floodplain after the Meiji Period.Intercultural Understanding, 3: 11-16.Research on change in land use and landscape pattern is the foundation for studies exploring natural and culturallandscape of a region. This study used GIS software and utilized topographic maps to examine the changes that occurred in theOguraike floodplain, during the time points of 1888, 1909, 1961, and 2002. The Oguraike floodplain, which was dominated by thelandscape of Oguraike Pond and paddy fields in 1888, was dominated by the landscape of urban areas and paddy fields in 2002.Moreover, urban areas, cropland, paddy fields, and grasslands have become concentrated into larger patches, whilst the water bodieshave become more fragmented. Overall, there has been a reduction in landscape diversity on the floodplain.

[39]
Kennedy R E, Townsend P A, Gross J Eet al., 2009. Remote sensing change detection tools for natural resource managers: Understanding concepts and tradeoffs in the design of landscape monitoring projects.Remote Sensing of Environment, 113: 1382-1396.Remote sensing provides a broad view of landscapes and can be consistent through time, making it an important tool for monitoring and managing protected areas. An impediment to broader use of remote sensing science for monitoring has been the need for resource managers to understand the specialized capabilities of an ever-expanding array of image sources and analysis techniques. Here, we provide guidelines that will enable land managers to more effectively collaborate with remote sensing scientists to develop and apply remote sensing science to achieve monitoring objectives. We first describe fundamental characteristics of remotely sensed data and change detection analysis that affect the types and range of phenomena that can be tracked. Using that background, we describe four general steps in natural resource remote sensing projects: image and reference data acquisition, pre-processing, analysis, and evaluation. We emphasize the practical considerations that arise in each of these steps. We articulate a four-phase process that guides natural resource and remote sensing specialists through a collaborative process to articulate goals, evaluate data and options for image processing, refine or eliminate unrealistic paths, and assess the cost and utility of different options.

DOI

[40]
Kim I, Le Q B, Park Set al., 2014. Driving forces in archetypical land-use changes in a mountainous watershed in East Asia.Land, 3: 957-980.Identifying patterns and drivers of regional land use changes is crucial for supporting land management and planning. Doing so for mountain ecosystems in East Asia, such as the So-yang River Basin in South Korea, has until now been a challenge because of extreme social and ecological complexities. Applying the techniques of geographic information systems (GIS) and statistical modeling via multinomial logistic regression (MNL), we attempted to examine various hypothesized drivers of land use changes, over the period 1980 to 2000. The hypothesized drivers included variables of topography, accessibility, spatial zoning policies and neighboring land use. Before the inferential statistic analyses, we identified the optimal neighborhood extents for each land use type. The two archetypical sub-periods, i.e., 1980–1990 with agricultural expansions and 1990–2000 with reforestation, have similar causal drivers, such as topographic factors, which are related to characteristics of mountainous areas, neighborhood land use, and02spatial zoning policies, of land use changes. Since the statistical models robustly capture the mutual effects of biophysical heterogeneity, neighborhood characteristics and spatial zoning regulation on long-term land use changes, they are valuable for developing coupled models of social-ecological systems to simulate land use and dependent ecosystem services, and to support sustainable land management.

DOI

[41]
Kokou K, Nuto Y, Atsri H, 2009. Impact of charcoal production on woody plant species in West Africa: A case study in Togo.Scientific Research and Essay, 4: 881-893.In Togo, fuel wood and charcoal account for more than 80% of total householdenergy requirements. Charcoal production results in a high pressure on the commonly used woody species. This study was carried out to assess the impact of charcoal production on the vegetation. Surveys involving 310 charcoal producers from 4 production areas were carried out. In each production area, density, diameter, height and basal area of woody species were measured both in unexploited plots and exploited plots. Inside the 4 charcoal production areas, 158 woody species were identified including 34 regularly exploited for charcoal production, that is, 15 preferred and 19 by default. Diversity indexes (Species Richness, Shannon and Evenness) and variability of densities are significantly higher in unexploited plots than those in exploited plots. The average heights vary between 4 and 7 m inside the exploited plots and from 6 to 9 m inside the unexploited plots. Average diameter and basal area are statistically not different inside unexploited plots and exploited plots. The most common regeneration methods inside the 4 charcoal production areas are seedlings and coppices. The preferred species regenerate better in exploited plots than those exploited by default. The study concludes that the negative effect of charcoal production on natural ecosystems has resulted in the depletion of the biodiversity, density, height, diameter of the stands and basal area of the woody species. Key words: Charcoal production, plant diversity, woody population structure, regeneration.

DOI

[42]
Lambin E F, Geist H J, Lepers E, 2003. Dynamics of land use and land cover change in tropical regions.Annu. Rev. Environ. Resour., 28: 205-241.

[43]
Le Q B, Tamene L, Vlek P L G, 2012. Multi-pronged assessment of land degradation in West Africa to assess the importance of atmospheric fertilization in masking the processes involved.Global and Planetary Change, 92/93: 71-81.Separating human-induced land degradation from that caused by natural processes in the world of global climate and atmospheric change is a challenging task, but important for developing mitigation strategies. Current remote-sensing data and spatio-temporal analyses allow the distinction of climate and human-induced land degradation on a sub-continental scale, but the underlying processes cannot be discerned at this scale. This study is conducted at a river-basin scale to (1) identify land degradation hotspots in a basin or sub-basin, and (2) assess the correspondence and divergence of land degradation assessed by NDVI shifts with and without accounting for atmospheric fertilization with that based on soil erosion assessment at a sub-basin scale. Long-term remote sensing (NDVI) and rainfall data were used to identify human-induced land degradation hotspot areas in the Volta basin. The results were compared with the critical zone of soil loss in the White Volta sub-basin derived from a spatially distributed soil erosion model, validated by field-measured data. A spatial comparison of the above studies revealed that the biomass productivity (NDVI)-based land degradation assessment grossly underestimated the extent to which soil is being lost, unless a correction was included to account for atmospheric fertilization. Based on inter-annual NDVI signals land degradation was evident in about 8% of the Volta basin's landmass, but when accounting for atmospheric fertilization, as much as 65% of the land is losing some of its vital attributes such as soil quality or vegetation productivity. The study demonstrates the need for using a multi-pronged assessment strategy in land degradation assessment that offers an insight of the processes involved in land degradation.

DOI

[44]
Leh M, Bajwa S, Chaubey I, 2013. Impact of land use change on erosion risk: An integrated remote sensing, geographic information system and modeling methodology.Land Degradation & Development, 24: 409-421.The objective of this study was to evaluate the impact of rapidly changing land use on erosion and sedimentation in a mixed land use watershed in the Ozark Highlands of the USA. The research combines a geographic information system-based soil erosion modeling approach with land use change detection to quantify the influence of changing land use on erosion risk. Five land use/land cover maps were generated or acquired for a 20-year period (1986 through 2006) at approximately 5-year intervals to assess land use change and to predict a projected (2030) land use scenario for the West Fork White River watershed in Northwest Arkansas. The Unit Stream Power based Erosion/Deposition model was applied to the observed and predicted land use to assess the impact on erosion. Total erosion from urban areas was predicted to increase by a factor of six between 1986 and 2030 based on the projected 2030 land use. Results support previous reports of increased urbanization leading to increased soil erosion risk. This study highlights the interaction of changes in land use with soil erosion potential. Soil erosion risk on a landscape can be quantified by incorporating commonly available biophysical data with geographic information system and remote sensing, which could serve as a land/watershed management tool for the rapid assessment of the effects of environmental change on erosion risk. Copyright (c) 2011 John Wiley & Sons, Ltd.

DOI

[45]
Lindstrom S, Mattson E, Nissanka S P, 2012. Forest cover change in Sri Lanka: The role of small scale farmers.Applied Geography, 34: 680-692.Forest cover in Sri Lanka has decreased rapidly during the last century and only fragments of the once widespread natural forest cover remain. This paper analyzes forest cover change and small scale farmers' relation to natural forests around two protected forest areas in Sri Lanka; Kanneliya Forest Reserve and Knuckles Conservation Forest. Methods used are spatial analysis to observe changes in forest cover from the 1980s until 2010, interviews with small scale farmers and key informants as well as field observations. In Kanneliya Forest Reserve, a decrease in forest cover is observed, particularly due to population increase and expanding tea plantations. In Knuckles Conservation Forest on the other hand, we find an overall increase in forest cover due to expansion of tree plantations, a ban on shifting cultivation and emigration from the area followed by natural forest regeneration. Agriculture is the most common source of income in both study areas and there is a clear link between conversions of forests to agricultural expansion. The profits from agricultural activities are in general insufficient to sustain small scale farmers' needs and the most common alternative source of income is achieved through resources extracted from the forest. Since 2001, demarcation of forest boundaries around the two forest reserves has reduced encroachment and illegal felling of timber. However, this policy has simultaneously threatened the livelihoods of peripheral communities in the forest buffer zones, especially in the investigated villages around Knuckles Conservation Forest. Despite successful attempts to reduce deforestation rates through governmental interventions, further incorporation of local people into the management of forests as stipulated in the current forest policy should be continued.

DOI

[46]
Lu D, 2006. The potential and challenge of remote sensing-based biomass estimation.International Journal of Remote Sensing, 27: 1297-1328.Remotely sensed data have become the primary source for biomass estimation. A summary of previous research on remote sensing‐based biomass estimation approaches and a discussion of existing issues influencing biomass estimation are valuable for further improving biomass estimation performance. The literature review has demonstrated that biomass estimation remains a challenging task, especially in those study areas with complex forest stand structures and environmental conditions. Either optical sensor data or radar data are more suitable for forest sites with relatively simple forest stand structure than the sites with complex biophysical environments. A combination of spectral responses and image textures improves biomass estimation performance. More research is needed to focus on the integration of optical and radar data, the use of multi‐source data, and the selection of suitable variables and algorithms for biomass estimation at different scales. Understanding and identifying major uncertainties caused by different stages of the biomass estimation procedure and devoting efforts to reduce these uncertainties are critical.

DOI

[47]
Lung T, Schaab G, 2010. A comparative assessment of land cover dynamics of three protected forest areas in tropical eastern Africa.Environmental Monitoring and Assessment, 61: 531-548.Processes of deforestation, known to threaten tropical forest biodiversity, have not yet been studied sufficiently in East Africa. To shed light on the patterns and causes of influences on protected forest ecosystems, comparisons of different study areas regarding land cover dynamics and potential drivers are needed. We analyze the of land cover since the early 1970s for three protected East African rainforests and their surrounding farmlands and assess the relationship between the observed changes in the context of the protection status of the forests. Processing of Landsat satellite imagery of eight or seven time steps in regular intervals results in 12 land cover classes for the -Nandi forests (Kenya) and Budongo Forest (Uganda) whereas ten are distinguished for Mabira Forest (Uganda). The overall classification accuracy assessed for the year 2001 or 2003 is similarly high for all three study areas (81% to 85%). The time series reveal that, despite their protection status, -Nandi forests and Mabira Forest experienced major forest decrease, the first a continuous forest loss of 31% between 1972/1973 and 2001, the latter an abrupt loss of 24% in the late 1970s/early 1980s. For both forests, the temporally dense time series show short-term fluctuations in forest classes (e.g., areas of forest regrowth since the 1980s or exotic secondary bushland species from the 1990s onwards). Although selectively logged, Budongo Forest shows a much more stable forest cover extent. A visual overlay with population distribution for all three regions clearly indicates a relationship between forest loss and areas of high population density, suggesting population pressure as a main driver of deforestation. The revealed forest losses due to local and commercial exploitation further demonstrate that weak management impedes effective forest protection in East Africa.

DOI PMID

[48]
Mander Ü, Uuemaa E, 2010. Landscape assessment for sustainable planning.Ecological Indicators, 10: 1-3.The 1st International Conference on Impact Assessment of Land Use Changes (IALUC) was held from the 6th to 9th of April 2008 at Humboldt University in Berlin, Germany. At the meeting, 160 oral presentations (including five plenary speeches) and 69 posters by representatives from 39 countries were presented. Papers and posters in four general sessions (Tools and Models, Policies and Scenarios, Scale and Indicators, and Participation and Socio-Economy) considered various aspects of the impact assessment of land use and landscape changes. This special section of the journal Ecological Indicators consists of nine selected presentations from different sub-sessions of the IALUC conference considering the use of indicators in landscape assessment.

DOI

[49]
Matsushita B, Yang W, Chen Jet al., 2007. Sensitivity of the Enhanced Vegetation Index (EVI) and Normalized Difference Vegetation Index (NDVI) to topographic effects: A case study in high-density cypress forest.Sensors, 7: 2636-2651.Vegetation indices play an important role in monitoring variations in vegetation.The Enhanced Vegetation Index (EVI) proposed by the MODIS Land Discipline Groupand the Normalized Difference Vegetation Index (NDVI) are both global-based vegetationindices aimed at providing consistent spatial and temporal information regarding globalvegetation. However, many environmental factors such as atmospheric conditions and soilbackground may produce errors in these indices. The topographic effect is another veryimportant factor, especially when the indices are used in areas of rough terrain. In thispaper, we theoretically analyzed differences in the topographic effect on the EVI and theNDVI based on a non-Lambertian model and two airborne-based images acquired from amountainous area covered by high-density Japanese cypress plantation were used as a casestudy. The results indicate that the soil adjustment factor 0104000004L01040000 in the EVI makes it moresensitive to topographic conditions than is the NDVI. Based on these results, we stronglyrecommend that the topographic effect should be removed in the reflectance data beforethe EVI was calculated01040000”as well as from other vegetation indices that similarly include a term without a band ratio format (e.g., the PVI and SAVI)01040000”when these indices are used in the area of rough terrain, where the topographic effect on the vegetation indices having only a band ratio format (e.g., the NDVI) can usually be ignored.

DOI PMID

[50]
Mattsson E, Persson U M, Ostwald Met al., 2012. REDD+ readiness implications for Sri Lanka in terms of reducing deforestation.Journal of Environmental Management, 100, 29-40.Any system to compensate countries for reduced emissions from deforestation and forest degradation (REDD+) requires a historical reference level against which future performance can be measured. Here we examine the possibilities Sri Lanka, a small forest country with limited data on forest carbon stocks, has to get ready for REDD+. We construct a historical reference level using available forest inventory data combined with updated 2008 and 2009 in situ carbon density data for Sri Lankan forests. Furthermore, we use a combination of qualitative and quantitative data to attribute the clearing of Sri Lankan forests in the latest years for which national forest inventory data are available, 1992–1996, to various proximate drivers and to estimate the opportunity cost of forest conservation. We estimate that baseline deforestation emissions in Sri Lanka amounted to 17MtCO2yr611 in the 1992–1996 period, but conclude that it is challenging for Sri Lanka to produce a robust and accurate reference level due to the lack of nationally based inventories. We find that the majority of forest clearing (87%) is due to small-scale, rainfed farming, with the two other major drivers being rice and tea cultivation. Further, Sri Lankan revenues from REDD+ participation could be substantial, but they are sensitive to REDD+ policy transaction cost, highly uncertain timber revenues, and particularly the carbon price paid for emission reductions. The latter needs to be higher than $5–10/tCO2 if there are to be substantial incentives for Sri Lanka to participate in REDD+. There is, however, a large gap in the knowledge of deforestation drivers that needs to be filled if Sri Lanka is to formulate an effective policy response to forest degradation in REDD+. For successful REDD+ implementation in Sri Lanka to happen, technological assistance, readiness assistance, and continued political momentum are crucial.

DOI PMID

[51]
Mbow C, Smith P, Skole Det al., 2014a. Achieving mitigation and adaptation to climate change through sustainable agroforestry practices in Africa.Current Opinion in Environmental Sustainability, 6: 8-14.Agroforestry is one of the most conspicuous land use systems across landscapes and agroecological zones in Africa. With food shortages and increased threats of climate change, interest in agroforestry is gathering for its potential to address various on-farm adaptation needs, and fulfill many roles in AFOLU-related mitigation pathways. Agroforestry provides assets and income from carbon, wood energy, improved soil fertility and enhancement of local climate conditions; it provides ecosystem services and reduces human impacts on natural forests. Most of these benefits have direct benefits for local adaptation while contributing to global efforts to control atmospheric greenhouse gas concentrations. This paper presents recent findings on how agroforestry as a sustainable practice helps to achieve both mitigation and adaptation objectives while remaining relevant to the livelihoods of the poor smallholder farmers in Africa.

DOI

[52]
Mbow C, Van Noodwijk M, Prabhu Ret al., 2014b. Knowledge gaps and research needs concerning agroforestry’s contribution to Sustainable Development Goals in Africa.Current Opinion in Environmental Sustainability, 6: 162-170.This review addresses the role of agroforestry in the links between food security and agricultural sustainability in Africa. We illustrate that the products and services flowing from the integration of trees within farming systems can contribute to food security, farmer livelihoods and environmental resilience. However, for agroforestry to be adopted it should not be constrained by policies which hinder the integration of trees, with crops and livestock. This policy scenario can best be met when the governance of food production at local to global scales is multi-sectoral and based on a ‘Systems Approach’. Nevertheless, the adoption of agroforestry has recently been greatly supported by the international agenda on the mitigation of climate change and to achieve sustainable food production. In conclusion we pose the hypothesis that “Agroforestry concepts and practices can form an effective, efficient and fair pathway towards the achievement of many Sustainable Development Goals”, and discuss the main messages and research questions emerging from the papers presented in this special issue.

DOI

[53]
McGarigal K, Cushman S A, Neel M Cet al., 2002. Fragstats: Spatial pattern analysis program for categorical maps, Amherst, USA, University of Massachusetts.

[54]
McGarigal K, Marks B J, 1995. FRAGSTATS: Spatial pattern analysis program for quantifying landscape structure, Corvallis, Oregon, Forest Science Department Oregon State University.

[55]
MEA 2005. Ecosystems and human well-being: Synthesis. Washington, DC: Island Press.

[56]
Monserud R A, Leemans R, 1992. Comparing global vegetation maps with the Kappa statistic.Ecological Modelling, 62: 275-293.ABSTRACT The Kappa statistic is presented as an objective tool for comparing global vegetation maps. Such maps can result from either compilations of observed spatial patterns or from simulations from models that are global in scope. The method is illustrated by comparing global maps resulting from applying a modified Holdridge Life Zone Classification to current climate and several climate change scenarios (CO2 doubling). These scenarios were based on the results of several different general circulation models (GCMs). The direction of change in simulated vegetation patterns between different GCMs was found to be quite similar for all future projections. Although there were differences in magnitude and extent, all simulations indicate potential for enormous ecological change. The Kappa statistic proved to be a useful and straightforward measure of agreement between the different global vegetation maps. Furthermore, Kappa statistics for individual vegetation zones clearly indicated differences and similarities between those maps. The Kappa statistic was found to be most useful for rank ordering of agreement, both across a series of maps and across the various vegetation zones within a map.

DOI

[57]
Nakakaawa C A, Vedeld P O, Aune J B, 2010. Spatial and temporal land use and carbon stock changes in Uganda: Implications for a future REDD strategy.Mitigation and Adaptation Strategies for Global Change, 16: 25-62.Using a map overlay procedure in a Geographical Information System environment, we quantify and map major land use and land cover (LULC) change patterns in Uganda period 1990–2005 and determine whether the transitions were random or systematic. The analysis reveals that the most dominant systematic land use change processes were deforestation (woodland to subsistence farmland—3.32%); forest degradation (woodland to bushland (4.01%) and grassland (4.08%) and bush/grassland conversion to cropland (5.5%) all resulting in a net reduction in forests (6.1%). Applying an inductive approach based on logistic regression and trend analyses of observed changes we analyzed key drivers of LULC change. Significant predictors of forest land use change included protection status, market access, poverty, slope, soil quality and presence/absence of a stream network. Market access, poverty and population all decreased the log odds of retaining forests. In addition, poverty also increased the likelihood of degradation. An increase in slope decreased the likelihood of deforestation. Using the stock change and gain/loss approaches we estimated the change in forest carbon stocks and emissions from deforestation and forest degradation. Results indicate a negligible increase in forest carbon stocks (3,260 t C yr-1) in the period 1990–2005 when compared to the emissions due to deforestation and forest degradation (2.67 million t C yr-1). In light of the dominant forest land use change patterns, the drivers and change in carbon stocks, we discuss options which could be pursued to implement a future national REDD plus strategy which considers livelihood, biodiversity and climate change mitigation objectives.

DOI

[58]
Nolte C, Agrawal A, Silvius K Met al., 2013. Governance regime and location influence avoided deforestation success of protected areas in the Brazilian Amazon.Proc. Natl. Acad. Sci. USA, 110: 4956-4961.Protected areas in tropical countries are managed under different governance regimes, the relative effectiveness of which in avoiding deforestation has been the subject of recent debates. Participants in these debates answer appeals for more strict protection with the argument that sustainable use areas and indigenous lands can balance deforestation pressures by leveraging local support to create and enforce protective regulations. Which protection strategy is more effective can also depend on (i) the level of deforestation pressures to which an area is exposed and (ii) the intensity of government enforcement. We examine this relationship empirically, using data from 292 protected areas in the Brazilian Amazon. We show that, for any given level of deforestation pressure, strictly protected areas consistently avoided more deforestation than sustainable use areas. Indigenous lands were particularly effective at avoiding deforestation in locations with high deforestation pressure. Findings were stable across two time periods featuring major shifts in the intensity of government enforcement. We also observed shifting trends in the location of protected areas, documenting that between 2000 and 2005 strictly protected areas were more likely to be established in high-pressure locations than in sustainable use areas and indigenous lands. Our findings confirm that all protection regimes helped reduce deforestation in the Brazilian Amazon.

DOI PMID

[59]
Ouedraogo I, 2010. Land use dynamics and demographic change in southern Burkina Faso.Acta Universitatis Agriculturae Sueciae.With the increasing world population, coupled with the technology improvement, man has emerged as the major, most powerful and universal instrument of environmental change in the biosphere today. To understand and predict the impacts of this change in the future, long-term reconstruction of land use and cover changes at global, regional and local scales is a prerequisite. The objective of this study was to assess the impacts of population growth on land cover change and generate knowledge that supports sound and informed decision-making on sustainable resource management. The study was done in Sissili Province, southern Burkina Faso, West Africa, where favorable rainfall and availability of arable land have contributed to attract farmers from the arid, crowded and unproductive zones of the north and centre of the country. The methodologies used were a combination of land cover change detection through time-series image processing (1976 - 2006), assessment of population dynamics, measurement of selected landscape metrics, and detection of systematic and random cover transition underlying the processes of change. The results showed that since the 1970s, cultivated areas have been expanded to the detriment of forest, and the expansion of cropland and the decline in forest cover are associated with population growth. Measurements of landscape metrics (Normalized Landscape Shape Index, Interspersion and Juxtaposition Index, and Area Weighted Fractal Dimension Index) highlighted the prevalence of environmental-unfriendly shifting cultivation practices and continual forest degradation. Land cover transition analyses showed that most changes were driven by systematic processes, such as changes induced by population growth, which underpin random changes that bring rapid and abrupt change temporarily with a potential to recover or not, depending on resilience and feedback mechanisms of the land cover type. To sustain the resource base, appropriate land management policy should be issued. The strategies that aim at minimizing the side-effects of the growing population on the environment in southern Burkina Faso might include population control, application of the national land tenure system, promotion of agricultural intensification related policies, promotion of fast-growing trees in plantations, and diversification of sources of income generation for rural people.

[60]
Pang A, Li C, Wang X., 2010. Land use/cover change in response to driving forces of Zoige County, China.Procedia Environmental Sciences, 2: 1074-1082.The changing characteristics of landscape patterns in Zoige County, from 1986 to 2005 were investigated using time-series high-quality Land sat Thematic Mapper(TM) images. The article first carried out the images data processing using the eCognition software, and then used the GIS and FRAGSTAT software to calculate the Land Use/Cover Change (LUCC) in Zoige County, and the changes and trend of wetland during the study area. The results showed that the landscape become more complicated and heterogeneous due mainly to fragmentation; NP and PD of the wetland increased 95.36% and 95.24% respectively, along with other metrics, this paper concluded that the wetland in Zoige County has been extensively disturbed; there was a net area of 410.93 km2 turned into grassland from wetland, which indicated that the ditching for grassland enlargement result in the degradation of the wetland in Zoige County. Through analysis of LUCC driving forces, finally got the conclusions: the climate change and human disturbance factors, including increasing temperature, over-grazing, drainage of water systems, were both responsible for the wetland degradation in Zoige County.

DOI

[61]
Pang C, Yu H, He J, 2013. Deforestation and changes in landscape patterns from 1979 to 2006 in Suan County, DPR Korea.Forests, 4: 968-983.The Democratic People Republic of Korea (DPR Korea) suffered considerable upland deforestation during the 1990s, yet its consequences remain relatively unknown. This paper examines this deforestation and resulting land-use change patterns by analysis of Landsat satellite images from 1979, 1992, 2001 and 2006 in Suan County, Hwanghae Province, DPR Korea. Results show that there has been significant closed canopy forest loss and a dramatic expansion of agricultural land during this period. Most forestlands were converted to farmland during 1992 and 2001. Food shortages, along with fuelwood and timber extraction, are considered to be the main drivers of deforestation. Landscape analysis also showed that closed canopy forests have been severely fragmented and degraded. These research findings make a contribution to an insufficient body of literature on environmental issues in DPR Korea and helps to establish a baseline for monitoring land-use and land-cover changes in the country.

DOI

[62]
Pattanaik C, Reddy C S, Reddy P M, 2011. Assessment of spatial and temporal dynamics of tropical forest cover: A case study in Malkangiri district of Orissa, India.Journal of Geographical Sciences, 21(1): 176-192.Tropical forests have been recognized as having global conservation importance. However,they are being rapidly destroyed in many regions of the world. Regular monitoring of forests is necessary for an adaptive management approach and the successful implementation of ecosystem management. The present study analyses the temporal changes in forest ecosystem structure in tribal dominated Malkangiri district of Orissa,India,during 1973-2004 period based on digitized forest cover maps using geographic information system (GIS) and interpretation of satellite data. Three satellite images Landsat MSS (1973),Landsat TM (1990) and IRS P6 LISS III (2004) were used to determine changes. Six land cover types were delineated which includes dense forest,open forest,scrub land,agriculture,barren land and water body. Different forest types were also demarcated within forest class for better understanding the degradation pattern in each forest types. The results showed that there was a net decrease of 475.7 km2 forest cover (rate of deforestation = 2.34) from 1973 to 1990 and 402.3 km2 (rate of deforestation = 2.27) from 1990 to 2004. Forest cover has changed over time depending on a few factors such as large-scale deforestation,shifting cultivation,dam and road construction,unregulated management actions,and social pressure. A significant increase of 1222.8 km2 agriculture area (1973-2004) clearly indicated the conversion of forest cover to agricultural land. These alterations had resulted in significant environmental consequences,including decline in forest cover,soil erosion,and loss of biodiversity. There is an urgent need for rational management of the remaining forest for it to be able to survive beyond next decades. Particular attention must be paid to tropical forests,which are rapidly being deforested.

DOI

[63]
Paudel N S, Vedeld P O, Khatri D B, 2015. Prospects and challenges of tenure and forest governance reform in the context of REDD+ initiatives in Nepal.Forest Policy and Economics, 52: 1-8.61This paper examines the REDD+ implementation in Nepal – where several REDD+ readiness initiatives and piloting are being implemented.61The paper reveals how key drivers of deforestation – contested forest tenure and weak governance, are underplayed in the readiness stage risking the future of successful REDD+ implementation.61The authors are a mix of practitioner and academic researchers.61The paper brings the most updated information and analysis of Nepal’s REDD+ readiness and informs policy makers in countries with similar process.

DOI

[64]
Peng J, Wang Y, Zhang Yet al., 2010. Evaluating the effectiveness of landscape metrics in quantifying spatial patterns.Ecological Indicators, 10: 217-223.The effectiveness of landscape metrics in quantifying spatial patterns is fundamental to metrics assessment. Setting 36 simulated landscapes as sample space and focusing on 23 widely used landscape metrics, their effectiveness in quantifying the complexity of such spatial pattern components as number of patch types, area ratio of patch types and patch aggregation level, were analyzed with the application of the multivariate linear regression analysis method. The results showed that all the metrics were effective in quantifying a certain component of spatial patterns, and proved that what the metrics quantified were not a single component but the complexity of several components of spatial patterns. The study also showed a distinct inconsistency between the performances of landscape metrics in simulated landscapes and the real urban landscape of Shenzhen, China. It was suggested that the inconsistency resulted from the difference of the correlation among spatial pattern components between simulated and real landscapes. After considering the very difference, the changes of all 23 landscape metrics against changing of number of patch types in simulated landscapes were consistent with those in the real landscape. The phenomenon was deduced as the sign effect of spatial pattern components on landscape metrics, which was of great significance to the proper use of landscape metrics.

DOI

[65]
Petursson J G, Vedeld P, Sassen M, 2013. An institutional analysis of deforestation processes in protected areas: The case of the transboundary Mt. Elgon, Uganda and Kenya.Forest Policy and Economics, 26: 22-33.Protected areas (PAs) are a country's key strategy to conserve and manage forest resources. In sub-Saharan Africa, the effectiveness and efficiency of PA institutions in delivering sustainable outcomes is debated, however, and deforestation has not been avoided within such formal regimes. This paper analyzes the processes that led to deforestation within the PAs on the transboundary Mt. Elgon, Uganda-Kenya, employing institutional theory. Landsat satellite imagery helped identify and quantify forest loss over time. The study showed how, since 1973, about a third of all forests within the PAs on Elgon have been cleared in successive processes. Within formal protected area regimes, complex political and institutional factors drive forest loss. We argue, therefore, that policies to counter deforestation using a PA model have to be considered and understood against the broader background of these factors, originating both inside and outside the PA regimes. (C) 2012 Elsevier B.V. All rights reserved.

DOI

[66]
Pontius J R G, Shusas E, Mceachern M, 2004. Detecting important categorical land changes while accounting for persistence.Agriculture, Ecosystems and Environment, 101: 251-268.The cross-tabulation matrix is a fundamental starting point in the analysis of land change, but many scientists fail to analyze the matrix according to its various components and thus fail to gain as much insight as possible concerning the potential processes that determine a pattern of land change. This paper examines the cross-tabulation matrix to assess the total change of land categories according to two pairs of components: net change and swap, as well as gross gains and gross losses. Analysis of these components can distinguish between a clearly systematic landscape transition and a seemingly random landscape transition. Multiple resolution analysis provides additional information concerning the distances over which land change occurs. An example of change among four land categories in central Massachusetts illustrates the methods. These methods enable scientists to focus on the strongest signals of systematic landscape transitions, which is necessary ultimately to link pattern to process.

DOI

[67]
Porter-Bolland L, Ellis E A, Guariguata M Ret al., 2012. Community managed forests and forest protected areas: An assessment of their conservation effectiveness across the tropics.Forest Ecology and Management, 268: 6-17.This paper assesses the role of protected and community managed forests for the long term maintenance of forest cover in the tropics. Through a meta-analysis of published case-studies, we compare land use/cover change data for these two broad types of forest management and assess their performance in maintaining forest cover. Case studies included 40 protected areas and 33 community managed forests from the peer reviewed literature. A statistical comparison of annual deforestation rates and a Qualitative Comparative Analysis were conducted. We found that as a whole, community managed forests presented lower and less variable annual deforestation rates than protected forests. We consider that a more resilient and robust forest conservation strategy should encompass a regional vision with different land use types in which social and economic needs of local inhabitants, as well as tenure rights and local capacities, are recognized. Further research for understanding institutional arrangements that derive from local governance in favor of tropical forest conservation is recommended.

DOI

[68]
Portillo-Quintero C A, Sanchez A M, Valbuena C Aet al., 2012. Forest cover and deforestation patterns in the Northern Andes (Lake Maracaibo Basin): A synoptic assessment using MODIS and Landsat imagery.Applied Geography, 35: 152-163.This work synthesizes results from the application of land cover classification techniques and probability sampling of satellite imagery for estimating forest extent and deforestation in Lake Maracaibo Basin (Venezuela and Colombia). A forest map was produced using a semi-automated supervised classification routine on MODIS 8-day 500-m imagery acquired in January 2010. Results show that forests occupy 29,710km 2 which represents 38% of the basin's total terrestrial landmass. From this extent, 61% belongs to Venezuela and 39% falls within the Colombian region. Findings indicate a drastic decrease in forest cover as a result of anthropogenic agricultural and urban expansion, especially when compared to its potential extent within the ‘Maracaibo dry forests’ and the ‘Venezuelan Andean montane forests’ ecoregions. Using time series of Landsat imagery, deforestation rates for the 1985–2010 time period were calculated. The analysis was performed on 24 samples blocks of 10×10km 2 randomly allocated within previously defined change probability strata. The general spatial distribution of deforestation rates was predicted by a simple regression model between sample blocks and prior change probabilities at the basin scale. Our results indicate that deforestation was low (<0.5%/y) in 85% of the basin, with highly focalized deforestation fronts (intermediate-to-high rates, <2.5%/y) in three regions: a) the Motatán river sub-basin in the Eastern Cordillera, b) the lower slopes of the Catatumbo river sub-basin and c) the submontane regions of the Apón and Santa Ana river sub-basins. The results of this paper lead the way for understanding current patterns in socioeconomic drivers of forest clearing in Lake Maracaibo Basin. The study also demonstrates the feasibility of using alternatives methods to the time-consuming and financially unsustainable methods traditionally used at national and sub-national scale in Venezuela and other Latin American countries.

DOI

[69]
Renetzeder C, Schindler S Peterseil Jet al., 2010. Can we measure ecological sustainability? Landscape pattern as an indicator for naturalness and land use intensity at regional, national and European level.Ecological Indicators, 10: 39-48.European landscapes have been shaped over the centuries by processes related to human land use, which are reflected in regionally distinct landscape patterns. Since landscape pattern has been linked to biodiversity and other ecological values of the landscapes, this paper explores landscape pattern as a tool for ecological sustainability assessments at the regional (Austrian Cultural Landscapes), national (Austria) and European (European Union + Norway, Switzerland) level with focus on agricultural landscapes. A set of landscape metrics served as a basis to assess naturalness and geometrisation of Austrian and European landscapes as a proxy for their sustainability. To achieve an accurate spatially explicit assessment, we applied a spatial reference framework consisting in units that are homogeneous in biophysical and socio-economic contexts, adapted the regional approach for its application at European level, and developed relative sustainability thresholds for the landscape metrics. The analyses revealed that several landscape metrics, particularly the "Number of Shape Characterising Points" showed a high correlation with the degree of naturalness. The sustainability map of Austria based on an ordinal regression model revealed well-known problem regions of ecological sustainabiiity. At the European level, the relative deviation from the average pattern showed clearly the simplification processes in the landscapes. However, a better spatial resolution of land cover data would add to the refinement of pattern analysis in regions and therefore the assessment of sustainability. We recommend the combination of information of different scales for the formulation and implementation of sustainability policies.

DOI

[70]
Rogan J, Chen D, 2004. Remote sensing technology for mapping and monitoring land-cover and land-use change.Progress in Planning, 61: 301-321.Information about land cover and land use is a very important component of the planning process as it can contribute to the debate on the current arrangements and patterns and the need to modify land use as part of a regional plan, a resource development or management project, an environmental planning exercise, or as a baseline study of a region. Planners may seek to suggest modifications to land-use patterns to achieve some social or economic outcomes, or as part of an environmental conservation or sustainability project, or to avoid some predicted future unwanted consequences. Access to accurate land-use maps can assist planners and the enterprise of planning. It is in this context that remote sensing is able to contribute. The purpose of this monograph is to present an overview and critique of the growing field of remote sensing as it applies to the mapping and monitoring of land-cover and land-use at a range of spatial and temporal scales. The ability of remote sensing to contribute to the mandate of planners and managers has changed significantly over the last decade. Satellite data are now available that can be used to map and monitor change from continental to local scales and over daily to weekly temporal scales. With the recent launch of satellites capable of collecting data that is comparable to aerial platforms, there is an enhanced capability of identifying change at small spatial scales. Similarly, advances in image processing, database management and spatial analysis tools have enhanced our ability to analyse these data for depicting land-cover and land-use change. Here, remote sensing technologies are described along with methods of analysing remote sensing data for detecting change at local, regional and continental scales. It is this diverse range of scales of observation and analysis that are now key to mapping and monitoring both anthropogenic and natural, and dramatic and incremental change. These aspects are demonstrated using case studies with different objectives and applied at different scales.

DOI

[71]
Rojas C, Pino J, Basnou Cet al., 2013. Assessing land-use and -cover changes in relation to geographic factors and urban planning in the metropolitan area of Concepción (Chile). Implications for biodiversity conservation.Applied Geography, 39: 93-103.

[72]
Romero-Ruiz M H, Flantua S G A, Tansey Ket al., 2012. Landscape transformations in savannas of northern South America: Land use/cover changes since 1987 in the Llanos Orientales of Colombia.Applied Geography, 32: 766-776.This study presents a detailed spatial, quantitative assessment of the land use/cover changes (LUCC) in the savanna region of Llanos Orientales in Colombia. LUCC was determined from multitemporal satellite imagery (Landsat and CBERS) from 1987 to 2007. Systematic landscape transitions were identified and put in the context of population change and economic activity. The results showed that during the period 1987 to 2007, 14% of the study area underwent some kind of land use/cover change, with most change occurring in the last decade. Systematic transitions were observed from flooded savannas to crops and exotic pastures. An important land cover change was linked to the expansion of palm oil plantations from 31km 2 in 1987 to 162km 2 in 2007. The observed changes are shown to be related to the economic and market-oriented-development from before 1970 to the present day. Based on the future economic development plans, the Llanos Orientales will continue to undergo significant change as an estimated 70% of the 17,000km 2 have been identified for conversion to plantation, or for petroleum and mining purposes. We provide recommendations for future economy integrated conservation, by proposing the implementation of a Llanos ecological network.

DOI

[73]
Ruelland D, Levavasseur F, Tribotté A, 2010. Patterns and dynamics of land-cover changes since the 1960s over three experimental areas in Mali.International Journal of Applied Earth Observation and Geoinformation, 12: 11-17.This paper addresses a critical need to provide a better quantitative understanding of how the Sudano-Sahelian environments actually been changing under the combined impacts of climate variability and the increasing pressure of human activity. Using Corona, Landsat and SPOT satellite images of three areas (90-250 km 2) along the climatic gradient of a large catchment in Mali, significant land-cover changes since the 1960s were identified through visual interpretation of images following a common classification scheme. The pattern and trajectory of changes differed markedly between the three areas studied. Overall, the 40-year trends indicate: (i) in the Sahelian area, a steady increase in croplands and erosional surfaces with sparse vegetation and a corresponding drastic reduction in woody covers; (ii) in the Sudano-Sahelian area, a large increase in croplands and a moderate reduction in woody covers; (iii) in the Sudanian area, agricultural extension, deforestation, but also reforestation and land rehabilitation, due to alternating periods of exploitation and recolonization by natural vegetation. These patterns and dynamics can be partially explained by the differences in demographic pressure between the three areas. They also highlight differences in response to anthropogenic and climate forcings depending on the areas- respective climatic and environmental endowments. This study is a first step towards an in-depth analysis of the various forces and processes driving these changes and the formulation of prospective environmental scenarios for the catchment in line with hydrological studies.

DOI

[74]
Schindler S, Von Wehrden H, Poirazidis Ket al., 2013. Multiscale performance of landscape metrics as indicators of species richness of plants, insects and vertebrates.Ecological Indicators, 31: 41-48.Landscape metrics are widely used to investigate the spatial structure of landscapes. Numerous metrics are currently available, yet only little empirical research has comparatively examined their indicator value for species richness for several taxa at several scales. Taking a Mediterranean forest landscape - Dadia National Park (Greece) - as a case study area, we explored the performance of 52 landscape level landscape metrics as indicators of species richness for six taxa (woody plants, orchids, orthopterans, amphibians, reptiles, and small terrestrial birds) and for overall species richness. We computed the landscape metrics for circular areas of five different extents around each of 30 sampling plots. We applied linear mixed models to evaluate significant relations between metrics and species richness and to assess the effects of the extent of the considered landscape on the performance of the metrics. Our results showed that landscape metrics were good indicators for overall species richness, woody plants, orthopterans and reptiles. Metrics quantifying patch shape, proximity, texture and landscape diversity resulted often in well-fitted models, while those describing patch area, similarity and edge contrast rarely contributed to significant models. Spatial scale affected the performance of the metrics, since woody plants, orthopterans and small terrestrial birds were usually better predicted at smaller extents of surrounding landscape, and reptiles frequently at larger ones. The revealed pattern of relations and performances will be useful to understand landscape structure as a driver and indicator of biodiversity, and to improve forest and landscape management decisions in Mediterranean and other forest mosaics.

DOI

[75]
Schleuning M, Farwig N, Peters M Ket al., 2011. Forest fragmentation and selective logging have inconsistent effects on multiple animal-mediated ecosystem processes in a tropical forest.PLoS One, 6: e27785.Forest fragmentation and selective logging are two main drivers of global environmental change and modify biodiversity and environmental conditions in many tropical forests. The consequences of these changes for the functioning of tropical forest ecosystems have rarely been explored in a comprehensive approach. In a Kenyan rainforest, we studied six animal-mediated ecosystem processes and recorded species richness and community composition of all animal taxa involved in these processes. We used linear models and a formal meta-analysis to test whether forest fragmentation and selective logging affected ecosystem processes and biodiversity and used structural equation models to disentangle direct from biodiversity-related indirect effects of human disturbance on multiple ecosystem processes. Fragmentation increased decomposition and reduced antbird predation, while selective logging consistently increased pollination, seed dispersal and army-ant raiding. Fragmentation modified species richness or community composition of five taxa, whereas selective logging did not affect any component of biodiversity. Changes in the abundance of functionally important species were related to lower predation by antbirds and higher decomposition rates in small forest fragments. The positive effects of selective logging on bee pollination, bird seed dispersal and army-ant raiding were direct, i.e. not related to changes in biodiversity, and were probably due to behavioural changes of these highly mobile animal taxa. We conclude that animal-mediated ecosystem processes respond in distinct ways to different types of human disturbance in Kakamega Forest. Our findings suggest that forest fragmentation affects ecosystem processes indirectly by changes in biodiversity, whereas selective logging influences processes directly by modifying local environmental conditions and resource distributions. The positive to neutral effects of selective logging on ecosystem processes show that the functionality of tropical forests can be maintained in moderately disturbed forest fragments. Conservation concepts for tropical forests should thus include not only remaining pristine forests but also functionally viable forest remnants.

DOI PMID

[76]
Schmitt-Harsh M, 2013. Landscape change in Guatemala: Driving forces of forest and coffee agroforest expansion and contraction from 1990 to 2010.Applied Geography, 40: 40-50.This study examines the land-use/cover change (LUCC) dynamics and drivers for two prominent land-use/cover systems in Guatemala: natural forests (FOR) and coffee agroforests (CAF). To-date, very little research has examined the LUCC dynamics of CAF, in large part due to the high degree of spectral similarity that exists between agroforests and other forest-cover types. Given the ecosystem and livelihood services provided by shade-grown coffee production, it is increasingly necessary to map and identify the dynamics and drivers of CAF changes over space and time. This research uses remote sensing analysis, land transition matrices, and multinomial regression models to examine LUCC dynamics over two ten-year intervals (1990-2000; 2000-2010) in Guatemala. Spatially explicit biophysical (e.g. slope, elevation) and accessibility (e.g. distance to roads) factors are used to model and compare drivers of change for CAF and FOR. Results demonstrate LUCC dynamics and drivers for the two land-use/cover systems to be complex over space and time. For example, FOR losses are evident for both time intervals, largely associated with conversion to CAF and croplands (CPL) in low slope, low altitude areas, and in areas close to existing croplands, respectively. CAF losses are also evident in the 1990s, but are outpaced by expansion in the 2000s. Losses are associated with conversion to CPL, particularly near roads and existing croplands, while expansion and/or persistence of CAF occurs near cities. These results suggest that conservation programs aimed at tree cover preservation and expansion should consider natural forests and managed agroforests separately. Further, such programs should be tailored to specific locations and institutional settings given the influence of topography and accessibility factors in determining localized patterns of landscape transformations over space and time. (c) 2013 Elsevier Ltd. All rights reserved.

DOI

[77]
Specht M J, Pinto S R R, Albuqueque U Pet al., 2015. Burning biodiversity: Fuelwood harvesting causes forest degradation in human-dominated tropical landscapes.Global Ecology and Conservation, 3: 200-209.This study provides an approximation of the potential impact of fuelwood harvesting in one of the most threatened tropical biodiversity conservation hotspots, the northern portion of the Brazilian Atlantic Forest. We test the relationship between fuelwood consumption and per capita income for 270 households distributed over 7 rural settlements. In general 76% of the households use fuelwood regularly and consume on average 686 kg/person/year of tree biomass, poorer people, however, consume 961 kg/person/year. Harvesting is concentrated to a few early successional species. Yet, annual rural population demand from 210 municipalities may reach 303,793 tons, equivalent to 1.2 to 2.1 thousand hectares of tropical forest. Fuelwood harvesting cannot be ignored as a major and chronic source of forest degradation in highly fragmented and densely populated landscapes and conciliating biodiversity conservation with poverty amelioration is an urgent task.

DOI

[78]
Teferi E, Bewket W, Uhlenbrook Set al., 2013. Understanding recent land use and land cover dynamics in the source region of the Upper Blue Nile, Ethiopia: Spatially explicit statistical modeling of systematic transitions.Agriculture, Ecosystems & Environment, 165: 98-117.The objective of this paper was to quantify long-term land use and land cover changes (LULCC) and to identify the spatial determinants of locations of most systematic transitions for the period 1957-2009 in the Jedeb watershed, Upper Blue Nile Basin. Black and white aerial photographs of 1957 and Landsat imageries of 1972 (MSS), 1986 (TM), 1994 (TM) and 2009 (TM) were used to derive ten land use and land cover classes by integrated use of Remote Sensing (RS) and Geographic Information System (GIS). Post-classification change detection analysis based on enhanced transition matrix was applied to detect the changes and identify systematic transitions. The results showed that 46% of the study area experienced a transition over the past 52 years, out of which 20% was due to a net change while 26% was attributable to swap change (i.e. simultaneous gain and loss of a given category during a certain period). The most systematic transitions are conversion of grassland to cultivated land (14.8%) followed by the degradation of natural woody vegetation and marshland to grassland (3.9%). Spatially explicit logistic regression modeling revealed that the location of these systematic transitions can be explained by a combination of accessibility, biophysical and demographic factors. The modeling approach allowed improved understanding of the processes of LULCC and for identifying explanatory factors for further in-depth analysis as well as for practical interventions for watershed planning and management.

DOI

[79]
Tfwala S S, Manyatsi A M, Wang Y, 2012. Assessment of land degradation at Velezizweni, Swaziland.Research Journal of Environmental and Earth Sciences, 4: 878-883.The objective of the research was to determine the status of land degradation and its implications for Velezizweni community, an area situated in the Highveld of Swaziland. Landsat ETM digital data taken on December 2009 was used to classify the land into different land cover categories. A questionnaire was developed and administered to 190 homesteads to solicit information on perception of land degradation from land users. About 19% of the area was classified as eroded, with about 45% classified as having poor vegetation cover. Land degradation was perceived as a serious problem by 81 and 90% of the respondents on cropping land and communal grazing land respectively. Invasion of alien plant species (Chromoleana odorata and Acacia pycnatha) was a serious problem, reducing the carrying capacity of grazing land. The results demonstrated that degradation was a serious problem despite the community being involved in some sustainable land management practises which included application of kraal manure to crops, ploughing across the slopes and keeping grass filter strips between cultivated lands.

[80]
Traoré L, Sop T K, Dayamba S Det al., 2012. Do protected areas really work to conserve species? A case study of three vulnerable woody species in the Sudanian zone of Burkina Faso.Environment, Development and Sustainability, 15: 663-686.Natural vegetation and native plant species contribute significantly to the daily needs of local people especially in developing countries. This exerts a high pressure on local species and jeopardizes the conservation of the most vulnerable plants. In Burkina Faso, conservation measures, such as the creation of protected forests, have been taken to safeguard the remaining indigenous vegetation. However, little is known about the effectiveness of these protected areas in conserving biodiversity. This study assessed and compared the population structures and regeneration potential of three vulnerable woody species- Diospyros mespiliformis Hochst., Prosopis africana (Guill. & Perr.) Taub. and Sterculia setigera Del.-攊n protected and unprotected areas in the Sudanian zone of Burkina Faso. The population structure and regeneration pattern of each species were compared between the North and South Sudanian sectors of Burkina Faso. The populations of all three species were unstable in both protected and unprotected areas. D. mespiliformis and P. africana displayed relatively good regeneration while P. africana lacked regeneration in unprotected areas. Regeneration was poor for S. setigera, regardless of protection status. The results suggest that the populations of the targeted species are unstable, regardless of the protection status of the area considered. This is probably due to the high anthropogenic pressure facing natural resources and raises serious concerns about the effectiveness of the protected areas in conserving biodiversity. Urgent measures are needed to ensure effective and efficient management and conservation of biodiversity in the protected areas of Burkina Faso.

DOI

[81]
Traore S S, Forkuo E K, Traore C S Pet al., 2015. Assessing the inter-relationship between vegetation productivity, rainfall, population and land cover over the Bani River Basin in Mali (West Africa).IOSR Journal of Engineering, 5: 10-18.This research investigated the inter-relationship between vegetation productivity, measured using the Normalized Difference Vegetation Index (NDVI), change in rainfall and population density in the context of perceived greening and degradation trends over the Bani River Basin (BRB). A 30-year (1982-2011), 8-km gridded rainfall data sets was produced by inverse distance weighted (IDW) interpolation of monthly data from 40 meteorological stations contained within the basin. Population data were retrieved from the National Population Statistic data base for 1987, 1997, and 2009. Rainfall and NDVI time-series trends were computed for the 30-year period and analysed. The relationship between rainfall and NDVI at pixel level, and NDVI and population densities was analysed using a Pearson correlation. Land Use and Land Cover (LULC) conversion rates were computed for the same period using multi-temporal 30-meter Landsat imagery; ground surveys for selected areas within the basin were used for further cross-verification. The computed NDVI trends revealed that, vegetation 'greening' trends are mostly associated with areas where natural vegetation is still well represented. Concurrent with increases in rainfall over the period analysed, this finding supports the hypothesis that re-greening observed in that area is the result of multi-decadal fluctuations in climate, rather than improved land management.

[82]
Traore S S, Landmann T, Forkuo E Ket al., 2014. Assessing long term trends in vegetation productivity change over the Bani River Basin in Mali (West Africa).Journal of Geography and Earth Sciences, 2: 21-34.Using time series of Normalized Difference Vegetation Index (NDVI) and rainfall data, we investigated historical vegetation productivity trends from 1982 to 2011 over the Bani River Basin in Mali. Statistical agreements between long-term trends in vegetation productivty, corresponding rainfall and rate of land cover change from Landsat time-series imagery was used to discern climate versus human-induced vegetation cover change. Spearman correlation was used to investigate the relationship between metrics of vegetation, rainfall trends and land cover change categories. The results show there is a positive correlation between increases in rainfall and some land cover classes, while some classes such as settlements were negatively correlated with vegetation productivity trends. Croplands and Natural Vegetation were positively correlated (r=0.89) with rainfall while settlements have a negative correlation with NDVI time series trend (r=-057). Despite the fact that rainfall is the major determinant of vegetation cover dynamics in the study area, it appears that other human-induced factors such as urbanization have negatively influenced the change in vegetation cover in the study area. The results show that a combined analysis of NDVI, rainfall and spatially explicit land cover change provides a comprehensive insight into the drivers of vegetation cover change in semi-arid Africa.

DOI

[83]
Tumusiime D M, Vedeld P, Gombya-Ssembajjwe W, 2011. Breaking the law? Illegal livelihoods from a protected area in Uganda.Forest Policy and Economics, 13: 273-283.Forests are important to local livelihoods and regulating access to forests will have consequences to those livelihoods and may promote illegal harvesting. This study analyses how local people make a living, focusing on the illegal collection of a forest's resources following its declaration as a Protected Area (PA). A household survey was conducted between October and December 2005, combining semi-structured individual household interviews and village level focus group discussions. Six sub-counties bordering the Rwenzori Mountain National Park in Uganda were chosen at random and two sample villages randomly selected from each. Through a participatory wealth ranking exercise, all the individual households in each sample village were assigned to one of three categories: rich, medium or poor. From this stratified list five individual households were randomly selected from each category for semi-structured interviews. Household livelihood outcomes were assessed and a fractional logit regression was used to estimate factors influencing dependency on forest income. Households with less access to assets exhibited greater dependence on forest resources. The average household was poor with a per adult equivalent unit income of 0.5 USD/day, with 18.6% of their income being derived from environmental resources. Based on income per adult equivalent unit, households were divided into poor and less poor. Both categories reported illegal collection of forest products. The poor households derived 32% of their environmental income and 12% of their total income from the park compared to the less poor at 18% and 4.5% respectively. The park resources reduced income inequality, as well as the incidence and depth of poverty by 2.8, 3.0, and 5.0 percentage points, respectively. Small reductions in the incidence of poverty suggest that forest resources may not be reliable as a pathway out of poverty, but the poverty depth measure shows that forest resources have a significant impact on helping to make the poor less poor. Under such circumstances, our observation is that increased law enforcement alone is unlikely to protect the park. Interventions that allow managed access to these resources in the short term, whilst creating operational opportunities outside the areas to cater for local peoples' rights and needs in the longer term may be more suitable.

DOI

[84]
Uuemaa E, Mander Ü, Marja R, 2013. Trends in the use of landscape spatial metrics as landscape indicators: A review.Ecological Indicators, 28: 100-106.The paper gives an overview on the trends in the usage of landscape metrics as indicators for: land use changes, habitat functions (biodiversity, habitats), landscape regulating functions (fire control, microclimate control, etc.), and information functions (landscape aesthetics). We reviewed papers published in international peer-reviewed journals that are indexed by the Institute of Science Information (ISI) Web of Knowledge from 2000 to 2010. The terms “landscape metrics”, “landscape indexes” and “landscape indices” were searched. Our analysis showed that application of the landscape metrics to characterize various ecosystem services and landscape functions has broadened during the last 10 years. Number of studies related to regulating and information functions of landscapes is increasing. However, the main exploitation field of the metrics is evaluation the change in land use/land cover.

DOI

[85]
Vedeld P, Jumane A, Wapalila Get al., 2012. Protected areas, poverty and conflicts.Forest Policy and Economics, 21: 20-31.This paper investigates livelihoods of communities around Mikumi, Tanzania's fourth largest national park, and impacts of living close to the park. People are very poor in the area, also beyond the areas close to the park. The average income is around 0.45 USD per person per day. People report food shortages in two out of the last five years. Even “the least poor group” earns no more than an average of 2 USD/cap and day.

[86]
Verburg P H, Van Asselen S, Van Der Zandenet al., 2013. The representation of landscapes in global scale assessments of environmental change.Landscape Ecology, 28: 1067-1080.Landscape ecology has provided valuable insights in the relations between spatial structure and the functioning of landscapes. However, in most global scale environmental assessments the representation of landscapes is reduced to the dominant land cover within a 0.5 degree pixel, disregarding the insights about the role of structure, pattern and composition for the functioning of the landscape. This paper discusses the contributions landscape ecology can make to global scale environmental assessments. It proposes new directions for representing landscape characteristics at broad spatial scales. A contribution of landscape ecologists to the representation of landscape characteristics in global scale assessments will foster improved information and assessments for the design of sustainable earth system governance strategies.

DOI

[87]
Vu Q M, Le Q B, Vlek P L G, 2014. Hotspots of human-induced biomass productivity decline and their social-ecological types toward supporting national policy and local studies on combating land degradation.Global and Planetary Change, 121: 64-77.61We delineated hotspots of human-induced biomass productivity decline in Vietnam.61We classified and characterized socio-ecological types of the detected hotspots.6119% of national land experienced biomass productivity decline for the last 25years.61The findings provided new information guiding further studies at lower scales.

DOI

[88]
Waiswa D, 2011. Dynamics of forest cover extent, forest fragmentation and their drivers in the Lake Victoria crescent, Uganda from 1989 to 2009 [D]. Virginia Polytechnic Institute and State University.

[89]
Wala K, Woegan Y A, Borozi Wet al., 2012. Assessment of vegetation structure and human impacts in the protected area of Aledjo (Togo).African Journal of Ecology, 50: 355-366.Protected areas constitute strategy for biodiversity conservation. Unfortunately, these sanctuaries of biodiversity are submitted to a high human pressure in Togo. This study carried out in the Alédjo protected area, aimed to make an analysis of various forms of human footprints and their impact on its plant resources. Methodological approach was based on forest inventory completed by inquiries. Ninety-four wooded species belonging to 35 families were counted. Floristic data analysis showed that seven species: Isoberlinia doka Craib & Stapf, Daniellia oliveri (Rolfe) Hutch. & Dalz., Berlinia grandiflora (Vahl) Hutch. & Dalz., Pterocarpus erinaceus Poir., Zanha golungensis Hiern, Khaya senegalensis (Desv.) A. Juss., Pentadesma butyracea Sabine was prominent. Five vegetation types were identified: riparian forests, dry forests, open forests, savanna woodlands, tree/shrub savannas with variable structural characteristics. The diversity indices in these plant communities are well significant and indicate a good distribution of species in the area. Several human activities such as fuel wood, fruits and medicinal plants gathering, carbonization, pasture were found within the protected area. Local authorities and associations are involved in the management of the protected area, but the participation of local populations needs to be improved.RésuméLes aires protégées constituent une stratégie de conservation de la biodiversité. Malheureusement, au Togo, ces sanctuaires de biodiversité sont soumis à de fortes pressions humaines. Cette étude, réalisée dans l'aire protégée d'Alédjo, visait à faire une analyse des diverses formes d'empreintes humaines et de leurs impacts sur ses ressources végétales. L'approche méthodologique s'est basée sur un inventaire forestier, complété par des enquêtes. Nous avons dénombré 94 espèces ligneuses, appartenant à 35 familles. L'analyse des données floristiques a mis en évidence sept espèces dominantes, à savoir Isoberlinia doka Craib & Stapf, Daniellia oliveri (Rolfe) Hutch. & Dalz., Berlinia grandiflora (Vahl) Hutch. & Dalz., Pterocarpus erinaceus Poir., Zanha golungensis Hiern, Khaya senegalensis (Desv.) A. Juss. et Pentadesma butyracea Sabine. Nous avons identifié cinq types de végétation: les forêts galeries, les forêt sèches, les forêts claires, les savanes boisées et des savanes arborées/arbustives avec des caractéristiques structurales variables. Les indices de diversité dans les communautés végétales sont élévés et indiquent une bonne distribution des espèces dans l'aire. On a relevé dans l'aire protégée plusieurs activités humaines, comme la collecte de bois de feu, de fruits et de plantes médicinales, la carbonisation et le p09turage. Des autorités et des associations locales sont impliquées dans la gestion de l'aire protégée, la participation des populations locales être améliorée.

DOI

[90]
Walz U, 2011. Landscape structure, landscape metrics and biodiversity.Living Rev. Landscape Res., 5: 5-16.

[91]
Wang D, Gong J, Chen Let al., 2013. Comparative analysis of land use/cover change trajectories and their driving forces in two small watersheds in the western Loess Plateau of China.International Journal of Applied Earth Observation and Geoinformation, 21: 241-252.To prevent soil loss and achieve better ecological environments, soil conservation measures have been taken during the past decades in the western Loess Plateau of China. In this paper, a case study was taken in Luoyu valley and Lver valley, two sub-watersheds of Xihe watershed and comparison was carried out between them. The main object of this study is to monitor land use/cover changes in the two similar small watersheds utilizing SPOT5 imageries by object-oriented human-揷omputer interactive classification method, further develop the method of spatio-temporal analysis of land use/cover change by using pattern metrics of change trajectories and relative land use suitability index ( R ) in smaller watersheds, and make comparisons between the two similar small watersheds, taking water and soil conservation measures into consideration. Results show that combining GIS and RS, this method can be perfectly applied to make comparisons between different small watersheds with similar geographical backgrounds. And land use/cover spatiotemporal dynamic change characteristics can be preferably expressed by pattern metrics of change trajectories and R values based on topographical data. Different emphases have been laid according to their own geological backgrounds in the two watersheds and human activities have different effects on the landscapes of the two watersheds. The main change pattern is from slope farmland to terrace (322, the largest in Luoyu valley) or to economic fruit forest (344, the largest in Lver valley). R value of every slope grade in both of the two watersheds drops with the rising of slope degree on the whole and it shows that there is still much to do for people in the two watersheds in consideration that all the R values are still lower than 0.7.

DOI

[92]
Weng Y-C, 2007. Spatiotemporal changes of landscape pattern in response to urbanization.Landscape and Urban Planning, 81: 341-353.The combined method of urban gradient analysis and landscape metrics in analyzing the changes of landscape pattern has been widely applied since its introduction by Luck and Wu (2002). In order to address the temporal dynamics of landscape change, this study integrated transect analysis with temporal trend analysis and specifically discussed how changes of residential pattern are related to forms of urban growth. Using Dane County, Wisconsin, USA as an example, a 60 km transect passing through the City of Madison was set up to represent a continuum of rural-urban-rural landscapes. Changes of landscape pattern from 1968 to 2000 were analyzed by FRAGSTATS with four metrics-ercentage of landscape (PLAND), Shannon's evenness index (SHEI), patch density (PD), and mean patch size (MPS). Findings from metric analyses revealed that the degree of land-use diversity and landscape fragmentation is positively related to the degree of urbanization. Specifically, at the class-level, residential land-use type shows the strongest positive relationship to the degree of urbanization in all of the class-level metrics adopted. Changes in residential land-use pattern were further analyzed with the number of housing units. The analyses revealed that there are different patterns of residential development along the transect in the study area-攚ith the core urban area expanding outward in a contiguous manner while the rural areas have scattered development. This study demonstrated the additional insights into landscape change by integrating the spatial and the temporal perspectives and by targeting the forms of residential developments.

DOI

[93]
Were K O, Dick Q B, Singh B R, 2013. Remotely sensing the spatial and temporal land cover changes in Eastern Mau forest reserve and Lake Nakuru drainage basin, Kenya.Applied Geography, 41: 75-86.61We examine land cover changes in Eastern Mau forest and Lake Nakuru basin, Kenya for 38 years.61The method is based on partitioning, hybrid classification and spatial reclassification of Landsat data.61The highest losses are for the forests-shrublands followed by grasslands.61The highest gains are for the croplands followed by built-up lands.61We recommend integrated environmental and agricultural policy formulation process.

DOI

[94]
White J, Shao Y, Kennedy Let al., 2013. Landscape Dynamics on the Island of La Gonave, Haiti, 1990-2010.Land, 2: 493-507.The island of La Gonave lies northwest of Port-au-Prince and is representative of the subsistence Haitian lifestyle. Little is known about the land cover changes and conversion rates on La Gonave. Using Landsat images from 1990 to 2010, this research investigates landscape dynamics through image classification, change detection, and landscape pattern analysis. Five land cover classes were considered: Agriculture, Forest/Dense Vegetation (DV), Shrub, Barren/Eroded, and Nonforested Wetlands. Overall image classification accuracy was 87%. Results of land cover change analysis show that all major land cover types experienced substantial changes from 1990 to 2010. The area percent change was 6139.7, 6122.7, 87.4, and 617.0 for Agriculture, Forest/Dense Vegetation, Shrub, and Barren/Eroded. Landscape pattern analysis illustrated the encroachment of Shrub cover in core Forest/DV patches and the decline of Agricultural patch integrity. Agricultural abandonment, deforestation, and forest regrowth combined to generate a dynamic island landscape, resulting in higher levels of land cover fragmentation.

DOI

[95]
Wilson H E, Sader S A, 2002. Detection of forest harvest type using multiple dates of Landsat TM imagery.Remote Sensing of Environment, 80: 385-396.A simple and relatively accurate technique for classifying time-series Landsat Thematic Mapper (TM) imagery to detect levels of forest harvest is the topic of this research. The accuracy of multidate classification of the normalized difference vegetation index (NDVI) and the normalized difference moisture index (NDMI) were compared and the effect of the number of years (1–3, 3–4, 5–6 years) between image acquisition on forest change accuracy was evaluated. When Landsat image acquisitions were only 1–3 years apart, forest clearcuts were detected with producer's accuracy ranging from 79% to 96% using the RGB-NDMI classification method. Partial harvests were detected with lower producer's accuracy (55–80%) accuracy. The accuracy of both clearcut and partial harvests decreased as time between image acquisition increased. In all classification trials, the RGB-NDMI method produced significantly higher accuracies, compared to the RGB-NDVI. These results are interesting because the less common NDMI (using the reflected middle infrared band) outperformed the more popular NDVI. In northern Maine, industrial forest landowners have shifted from clearcutting to partial harvest systems in recent years. The RGB-NDMI change detection classification applied to Landsat TM imagery collected every 2–3 years appears to be a promising technique for monitoring forest harvesting and other disturbances that do not remove the entire overstory canopy.

DOI

[96]
Wittig R, Konig K, Schmidt Met al., 2007. A study of climate change and anthropogenic impacts in West Africa.Env. Sci. Pollut. Res., 14: 182-189.Background, Aim and Scope During the last decades ecological conditions in West Africa have dramatically changed. Very evident is the climate change, which has resulted in a southward shift of the climate zones, e.g. a spread of the desert (Sahara) into the Sahelian zone. After the drought period of the early 1970s and 1980s, livestock density increased resulting in an intensification of grazing pressure. This anthropogenous phenomenon leads to similar landscape changes as those caused by the climate. Only very few investigations exist on vegetation dynamics, climate changes and land use changes for the Sudanian zone. The paper presents data on changes of precipitation, of land use, of the geographical range of species, and of the composition of the flora, which have to be regarded as proofs of the sahelisation of large areas of the Sudanian zone. Materials and Methods Area of investigation: Burkina Faso. Precipitation data analysis: precipitation data from 67 stations; time series analysis and geo-statistical spatial interpolation. Analysis of land use change: Landsat satellite MSS and ETM+ data, acquired for two different dates between 1972 and 2001 analyzed by the software ERDAS/IMAGINE version 8.6 and ArcView 3.2 with the Spatial Analyst extension. Intensive ground truthing (160 training areas). Inventory of the flora: based on the data of the Herbarium Senckenbergianum (FR) in Frankfurt, Germany, and of the herbarium of the university of Ouagadougou (OUA), Burkina Faso, as well as on various investigations on the vegetation of Burkina Faso carried out in the years 1990 to 2005 by the team of the senior author. Life form analysis of the flora: based on the inventory of permanent plots. Results and Discussion Precipitation: Remarkable latitudinal shift of isohyets towards the South translates to a general reduction of average rainfall in great parts of the country. The last decade (1990-1999) shows some improvement, however, the more humid conditions of the 1950 and 1960 are not yet established again. Landcover change: In the study region the extent of arable fields and young fallows increased during the last 30 years from 580 km 2 in 1972 to 2870 km 2 in 2001. This means an average land cover conversion rate of 0.9% per year for the 6 departments considered. Change of the distribution of Sahelian and Sulanian species: Several species, mentioned in older literature as strictly Sahelian, today also occur in the Sudanian zone. Parallel to the spread of former strictly Sahelian species into the Sudanian zone, some former Sahelo-Sudanian species have withdrawn from the Sahel . Changes of the life form spectra of the flora: Considering their life form spectra, the flora of heavily grazed and of protected areas in the Sudanian zone show great differences. On areas intensively grazed the percentage of therophytes is evidently higher than on protected areas. Just the opposite is true for the phanerophytes. Their percentage is higher on the protected area than on the grazed zones. At the first glance, it is obvious to link the changes in flora and vegetation with the climate changes that have occurred during the last five decades (decrease of annual precipitation). However, not only climatic conditions have changed, but also population has increased, the percentage of land intensively used for agriculture and pasturing has increased and the time for soil regeneration today is much shorter than it was some decades ago. Thus, the landscape of the Sudanian zone has become a more Sahelian character. A comparison of the flora of an intensively used area of the Sudanian zone with that of a protected area shows a remarkable change in the life form spectra. The spectrum of the intensively used area is almost identical with that of the typical Sahelian flora. This comparison shows that the anthropogenic influence plays a greater role in the sahelisation of the Sudanian zone than the climate change. Conclusion Climate change and anthropogenic influence both, lead to a sahelisation of landscape and flora. Thus in many parts of the Sudanian zone of West Africa sahelisation phenomena will remain and even increase independently from the reestablishment of the more humid climate conditions of the 1950ies. Recommendations and Perspectives In order to maintain some parts of the characteristic Sudanian landscape with its characteristic flora and vegetation, the number and size of protected areas should be augmented. For all protected areas it has to be ensured, that protection is reality, i.e. respected an understood by local people, not only fiction. As long as the enlargement of intensively used areas continues the sahelisation of flora, vegetation and landscape will continue too.

DOI PMID

[97]
Woegan Y A, 2007. Diversite des formations végétales de deux aires protégées de l’Atakora Nord : la réserve de faune d’Alédjo et Malfakassa [D]. University of Lome.

[98]
Wu C-F, Lin Y-P, Chiang L-Cet al., 2014. Assessing highway's impacts on landscape patterns and ecosystem services: A case study in Puli Township, Taiwan.Landscape and Urban Planning, 128: 60-71.Highway construction facilitates urban growth in Taiwan. However, the long-term effects of transportation infrastructure are not well understood; these include land-use changes, changes in landscape patterns, and the alteration of ecosystem services. To assess the effects of different land-use scenarios under various agricultural and environmental conservation policy regimes, this study applies an integrated approach to analyze the effects of Highway 6 construction on Puli Township. Interviews with neighborhood leaders of Puli Township, along with remote sensing analysis, reveal that both biophysical and socioeconomic factors are the major forces driving land-use change. The effects of these land-use changes are varied. An example is the road-effect zone, which for Puli Township extends 400m perpendicular to the length of the highway; however, due to differing spatial patterns it is highly asymmetric; indirect effects include the spatial restructuring of certain landscapes, which can drastically influence habitat dynamics. Land-use simulation results indicate that agricultural and environmental conservation policies have significant effects on projected land-use patterns in the southern part of Puli's downtown area and in areas along major roads. Specifically, highway construction and subsequent urbanization under various land-use policies result in varying degrees of isolation and fragmentation in the overall landscape pattern. A habitat quality assessment using the InVEST model indicates that the conservation of agricultural and forested lands improves habitat quality and preserves rare habitats. In summary, appropriate environmental policies will mediate both the direct and indirect impacts of Highway 6 on landscape patterns and ecosystem services in Puli Township.

DOI

[99]
Zhai D L, Cannon C H, Dai Z Cet al., 2015. Deforestation and fragmentation of natural forests in the upper Changhua watershed, Hainan, China: Implications for biodiversity conservation.Environ. Monit. Assess., 187: 4137.Hainan, the largest tropical island in China, belongs to the Indo-Burma biodiversity hotspot. The Changhua watershed is a center of endemism for plants and birds and the cradle of Hainan’s main rivers

DOI PMID

[100]
Zhang Z, Van Coillie F, De Clercq E Met al., 2013. Mountain vegetation change quantification using surface landscape metrics in Lancang watershed, China.Ecological Indicators, 31: 49-58.Land cover and vegetation change are among the most important aspects of environmental change. Vegetation change can be quantified by landscape pattern indices (LPI). Landscape indices are routinely calculated using planar land use/land cover (LU/LC) maps, obtained by the projection of a non-flat landscape surface into a two-dimensional Cartesian space. Especially in mountainous areas, quantification on planar maps can lead to underestimation of vegetation and land cover changes. Hoechstetter et al. (2008) developed a method to compute LPIs in a surface structure by calculating landscape patch surface area and surface perimeter from digital elevation models (DEM). As yet there have been no applications of these surface landscape indices on land use/land cover and vegetation change quantification. The objectives of this study are to (1) choose a LPI method (surface metrics pattern analysis or common planimetric metrics pattern analysis) for vegetation change quantification; and (2) employ the selected surface LPI method to assess vegetation pattern change in two mountainous areas of the Lancang watershed, Yunnan Province, China. The results show that the surface approach to estimate changes of class area (CA), mean patch area (MPA), and mean Euclidean Near-Neighbor distance (MENN) may obtain more accurate results for quantifying vegetation change in steep mountain areas. Forest fragmentation increased significantly over time in the two different mountainous study areas. The patches of two land cover classes, (i) agricultural land and (ii) low density forest and tall shrubs, became more aggregated in the northern (temperate) study area. In the southern (tropical) study area, rubber plantations increased considerably in size and became more aggregated.

DOI

[101]
Zheng B, Myint S W, Fan C, 2014. Spatial configuration of anthropogenic land cover impacts on urban warming.Landscape and Urban Planning, 130: 104-111.Anthropogenic land cover types greatly influence the urban heat island (UHI) effects. This study examined effects of composition and spatial pattern of anthropogenic land cover features on land surface temperature (LST) in Phoenix, Arizona, USA, using a land cover map derived from high resolution satellite data and ASTER LST data. The spatial pattern of land cover features was measured by local Moran's I —a continuous spatial autocorrelation index, which can effectively describe dispersed or clustered patterns of land cover features. Our results showed that the composition and spatial pattern of buildings have minimal impacts on LST, while those of paved surfaces alter LST more drastically. The local Moran's I of paved surfaces have a stronger positive correlation with nighttime ( r 2 02=020.38) than daytime ( r 2 02=020.17) temperatures, suggesting that clustered paved surfaces create stronger warming effects at night. We further controlled for land cover compositions to minimize their effects on LST, and found that the magnitude of warming effects caused by clustered paved surfaces differed among landscapes of varying land cover compositions. Correlations between local Moran's I of paved surfaces and LST becomes stronger when paved surface fraction exceeds 50%. These results illustrated aggregate warming effects of clustered paved surfaces, and provide insights on UHI mitigation for land managers and urban planners.

DOI

[102]
Zhou Q, Li B, Kurban A, 2008. Trajectory analysis of land cover change in arid environment of China.Int. J. Remote Sensing, 29: 1093-1107.Remotely sensed data have been utilized for environmental change study over the past 30 years. Large collections of remote sensing imagery have made it possible for spatio‐temporal analyses of the environment and the impact of human activities. This research attempts to develop both conceptual framework and methodological implementation for land cover change detection based on medium and high spatial resolution imagery and temporal trajectory analysis. Multi‐temporal and multi‐scale remotely sensed data have been integrated from various sources with a monitoring time frame of 30 years, including historical and state‐of‐the‐art high‐resolution satellite imagery. Based on this, spatio‐temporal patterns of environmental change, which is largely represented by changes in land cover (e.g., vegetation and water), were analysed for the given timeframe. Multi‐scale and multi‐temporal remotely sensed data, including Landsat MSS, TM, ETM and SPOT HRV, were used to detect changes in land cover in the past 30 years in Tarim River, Xinjiang, China. The study shows that by using the auto‐classification approach an overall accuracy of 85–90% with a Kappa coefficient of 0.66–0.78 was achieved for the classification of individual images. The temporal trajectory of land‐use change was established and its spatial pattern was analysed to gain a better understanding of the human impact on the fragile ecosystem of China's arid environment.

DOI

Options
Outlines

/

〈 〉