INTRODUCTION

Vegetation and land use in the Sahel region are strongly shaped by human activities and variation in rainfall. The relative importance of these two factors has sparked persistent debates over what drives change in the Sahel. Following a series of severe droughts and famines in the 1970s and 1980s, researchers and policy makers revived a narrative of continuous human induced degradation and desertification, which was linked to population growth (Swift 1996, Reenberg 2012). Tree monitoring was previously focused on areas defined as forest, ignoring most of the trees present in the Sahel (Brandt et al. 2020). In the mid-2000s, observations of increases in vegetation since the early 1980s lead to a narrative of “re-greening” of large parts of the Sahel (Fensholt et al. 2009, Herrmann et al. 2005, Olsson et al. 2005, Souverijns et al. 2020). Over the period 2000–2015, there was no significant trend in total vegetation in most parts of the Sahel, but an increase in 16% of the area (Leroux et al. 2017), whereas woody vegetation was in general increasing, primarily in areas with low anthropogenic pressure (Brandt et al. 2016). The spatial variability was high, but there was no decrease in woody vegetation in established croplands. In fact, woody cover in farmlands in semi-arid Sahel was higher than in surrounding savannas (Brandt et al. 2018). Increase in vegetation has generally been seen as beneficial for peoples’ livelihoods (Reij et al. 2005, Tougiani et al. 2009, Sendzimir et al. 2011). However, some studies have revealed that along with increased tree and shrub cover, there has been a shift in species composition toward more drought tolerant and/or exotic species, as well as a loss of large trees (Herrmann and Tappan 2013, Hänke et al. 2016), which change the value of the vegetation increase to people.

In this context, estimating the dynamics of ecosystem services, the benefits people obtain from nature (Millennium Ecosystem Assessment 2005), can reveal how these multiple changes in vegetation have been connected to multiple aspects of people’s livelihoods. However, there has been limited success in analyzing temporal changes in ecosystem services (Nicholson et al. 2009, Bürgi et al. 2015, Berbés-Blázquez et al. 2016). Existing approaches to analyzing dynamics of ecosystem services are based on biophysical features and secondary data (Renard et al. 2015, Tomscha and Gergel 2016) or historic documents (Wilkinson et al. 2013, Bürgi et al. 2015), but in the Sahel region, secondary data and historic documents for the same area over time are absent or difficult to locate or access. For example, there is no data on collected wild foods beyond one or a few villages at a single point in time. Yield data for dominant crops is available since the 1980s, but not on finer scale than administrative province. Livestock data is also available for provinces but extrapolated from one census. Land cover maps, particularly historical, are limited and often only available at very low spatial resolution (Souverijns et al. 2020).

In the Sahel, rural people are highly dependent on the direct benefits produced by their local landscapes. For example, in Burkina Faso 63% of the total population and 80% of the rural population are occupied in different forms of agriculture, predominantly small-scale (INSD 2014). Using the framework of ecosystem services allowed us to move beyond the focus on a single resource (e.g., crops) to study change in multiple provisioning services across the village landscapes. The provisioning ecosystem services that we studied co-exist with cultural ecosystem services and depend on regulating and supporting ecosystem services across these landscapes. For example, water is essential to life in the Sahel, and while all the ecosystem services we assessed require water, the ecosystem services that ensure water (moisture recycling, storage, and purification) are exceedingly difficult to estimate as they emerge from a temporally and spatially variable combination of tree density, landscape topography, soil hardening, and infrastructure built to support percolation and storage, such as planting pits (zaï), stone bunds, and ponds. A fuller understanding on how people depend on landscapes is needed as livelihoods in the Sahel region still face enormous challenges, not the least from effects of climate change (Salack et al. 2016, Sultan and Gaetani 2016), widespread poverty (UNDP 2015), and a rapidly growing population (FAO 2017).

We estimated historical changes in ecosystem services using a simple model to estimate changes in key provisioning ecosystem services through a combination of analysis of historic imagery, and results from previous fieldwork. We view ecosystem services as being co-produced by ecosystems and people (see, e.g., Reyers et al. 2013, Palomo et al. 2016). This co-production includes both human input of different capitals to enhance desired services, and the needs, values, and aspirations that shape what is appreciated as an ecosystem service (Daw et al. 2011, 2016, Chan et al. 2012, Ernstson 2013, Bennett et al. 2015).

Our assessment of changes in ecosystem services is based on what we term social-ecological patches: landscape units that correspond with the words that local people use when describing their landscapes, characterized by a combination of land use, land cover, and topography (Sinare et al. 2016), and the provisioning ecosystem services associated with each of these patches today. We explore the range of plausible change using (1) the assumption that social-ecological patches produced similar sets of ecosystem services in the past to what they do today, and (2) high and low estimates for plausible change based on literature and field observations of changes in the combination of availability and use of ecosystem services over time. This allows us to estimate ecosystem services across time despite having only partial knowledge or data of past changes in relationships between villagers and their landscapes.

METHODS

This study was a part of a research project investigating ecosystem services and livelihoods in relation to the observed re-greening in Burkina Faso. It builds in particular on the assessment of current ecosystem services from different social-ecological patches made by Sinare et al. (2016), but is focused on how this information can be combined with interpretation of historical images to estimate change in ecosystem services over time.

Study sites

The study was conducted in six villages in Burkina Faso with similar biophysical and social characteristics, three in the administrative region Nord and three in the region Centre-Nord (Fig. 1). These regions have had, respectively, higher and lower vegetation recovery since the droughts that ended in the early 1980s than expected from increases in rainfall over the same period. This was shown in a large-scale study of residual trends in normalized difference vegetation index (NDVI) 1982–2008 that cannot be explained by change in 3-month cumulative rainfall (Herrmann et al. 2005). The villages are located around the 600 mm yearly rainfall isohyet 1968–2000 (OECD/SWAC 2014). The population in the villages range from 600 to 1500 inhabitants (Government of Burkina Faso 2005, see Sinare et al. 2016 for details), a size that is representative of villages in the region.

Mapping of past social-ecological patches

We mapped the change in spatial extent of social-ecological patches using visual interpretation of aerial photographs and satellite images from six times between 1952 and 2016: Aerial photographs from Institut Géographique du Burkina (IGB) and Institut National de l’Information Géographique et forestière (IGN) 1952/1955 (resolution - pixel size in meters at interpretation 0.9/0.7–0.8); Corona satellite images 1967/1968 (2.8/2.2–2.3); aerial photographs from IGB 1983/1984 (1.5–1.6/1.6); aerial photographs from IGB 1996 (1.5–1.8); satellite images CNES Astrium/Google Earth Pro 2006/2010 (0.9/0.6–0.7) and satellite images CNES Astrium/Google Earth Pro 2013/2016 (1.7/1.2–1.3). For more detail see Table 1. We imported all images to ArcGIS 10.3 software (ESRI 2013) and converted to UTM projection to allow for distance measurements. CNES Astrium/Google Earth Pro images from 2006/2010 were used as base layer for georeferencing all other images. We selected easily identifiable ground control points, for example sacred forest patches, for the georeferencing, resulting in 0.15–1 ground control point per square kilometer. This is comparable or exceeding the density of ground control points used in a similar landscape (Ruelland et al. 2011). Each image was transformed with first order polynomial transformation.

The images were interpreted manually on-screen. Manual interpretation has previously been shown to give more reliable results than automated techniques in the Sahel context (Tappan et al. 2004, Tappan and McGahuey 2007) because of the possibility of combining different image sources including film-based aerial photographs, and integrating field knowledge in the interpretation. For this study, the detailed field knowledge from participatory research activities including walks across all villages with regular stops to document landscape use (methods described in Sinare et al. 2016), was important for the image interpretation.

We used the village borders identified in 2011 (Sinare et al. 2016) for all image analysis because there are no official maps of village borders available. We applied a grid on the time series of images for each village, and classified each cell as a type of social-ecological patch. We did this by looking at each 100 x 100 meters cell potentially covering multiple social-ecological patches, and classifying it by the social-ecological patch at its center (within a circle with 5 m radius around the cell center). The social-ecological patch class had to cover at least 0.25 hectares connected land for the cell to be classified to this patch class. We selected this method as compared to a method mapping polygons because it was faster and as accurate. The method was tested against polygons mapped for 2006 (Sinare et al. 2016) and yielded the same percentage cover of the different social-ecological patches.

We classified nine types of social-ecological patches (term in local language Mooré in parenthesis): Fields (Pùtâ), Shrubland (Weoogo), Homesteads (Kamanse), Depressions (Baongho), Forest (Kaongo), Bare soil (Zipellé), Woodland (Kangré), Irrigated land for vegetables (French term Jardin maraîcher used), and Public buildings (see Table 2 for definitions). As compared to the social-ecological patch types identified in the villages 2011 (Sinare et al. 2016), we added woodland, land around public buildings, and irrigated land for vegetables, and combined fallows into shrubland or fields. Fallows were not possible to detect, as young fallow looks like fields and old fallow resembles shrubland on the images, and therefore were classified into these categories. The attribution of values for ecosystem services to each social-ecological patch builds on modified scores from previous in-depth fieldwork (Sinare et al. 2016; see Appendix 1 for detailed description).

Woody vegetation has been documented to provide a wide range of ecosystem services such as food, fodder, firewood, and regulation of water, carbon nutrients, and micro climate in these landscapes (see, e.g., Bayala et al. 2014, Sinare and Gordon 2015). To examine the role of woody vegetation on cultivated land for the generation of ecosystem services, we removed the scores for leaf vegetables from trees, fruits, medicine, firewood, and construction material for the social-ecological patches, fields, homesteads (cultivated areas around houses), and depressions (cultivated lower lying areas) to create a “without trees” scenario.

Change in ecosystem services

We estimated the historical generation of ecosystem services in two ways. The first method assumed a constant relationship between ecosystem services and social-ecological patches, while the second estimated a range of ecosystem service values based on how relationships with social-ecological patches have changed over time. Comparing these methods allowed us to assess which historical estimates are likely robust or uncertain.

For the first method, we made the naïve assumption that each social-ecological patch type generated the same set and relative yield of ecosystem services over time. This assumption allowed us to extrapolate our field data on current ecosystem services produced by village landscapes into historical landscapes. We multiplied the production of ecosystem services in each patch type by the frequency of that patch type to estimate the total availability of each ecosystem service for each historical time period and village (Fig. 2). However, ecosystem service generation has changed over time because of changes in farming and land use practices, as well as in peoples’ values and needs.

For the second method, we estimated a historic high and low production of each ecosystem service from each social-ecological patch type to establish a plausible interval for the generation of ecosystem services in 1952 relative to 2016. The co-produced nature of ecosystem services implies that change over time depends (a) on the generation of the potential ecosystem service, i.e., the material that can become an ecosystem service if valued, which is influenced, for example, by change in vegetation, or in technology and inputs used in agriculture; and (b) on people’s demand for that material, which makes it a provisioning ecosystem service. The demand depends both on the population density (which we address below), and on the degree to which the population is dependent on the local landscape. To estimate a plausible interval for change in ecosystem services, we combined (a) and (b). The assumptions for historical high and low values for ecosystem services were based upon previous fieldwork (Sinare et al. 2016), further discussion with villagers about changes over time, and the previous studies from the region that we could find, which give an indication of changes (Table 3). We expected that most estimates of change would be between these values. We linearly extrapolated this range between 1952 and 2016 to estimate how the production of ecosystem services has changed over time. By combining this change in production with the change in the social-ecological patches in each village we were able to estimate the change in available ecosystem services in each village (Fig. 2). We used available population data for each village and time period (Table 4), to estimate how ecosystem service availability has change on a per person basis. We compared per person availability of each ecosystem service in the past to its current availability to identify whether this availability has changed over time. These calculations were done using R (R Core Team 2019), with the packages raster (Hijmans 2020), rasterVis (Perpinan Lamigueiro and Hijmans 2019), rgdal (Bivand et al. 2019), ggplot2 (Wickham 2016), and ggpubr (Kassambara 2020).

RESULTS

Change in social-ecological patches over time

The dominant trend in all villages is a sharp increase in fields with trees (on average 60% increase between 1952 and 2016), and decrease in shrubland area (on average 64% decrease; Fig. 3). The villages have different rates of change, with less increase in villages that were predominantly cultivated already in the 1950s. For example, fields covered almost 70% of Boursouma in 1952, and that area increased by only 17% over the whole period, being relatively stable since the 1980s. Lebda, on the other extreme, was covered with less than 30% fields in 1955, and the field area increased by 70% over the whole period. Zarin had the largest increase in fields, where it almost doubled. The percentage increase in fields is higher in all villages in region Centre-Nord (70–99%) as compared to region Nord (17–56%). Fields with trees covers on average 64% of the village area and shrubland covers on average 18%.

There were substantial changes in the extent of less common social-ecological patches (Fig. 3b). On average, homesteads have quadrupled. Bare soil has also increased with larger increase in the region Centre-Nord than in the region Nord. Lebda had the largest presence of bare soil with 8% of the landscape covered in 2010, and a nine-fold increase in bare soil surface versus its minimum extent (1983). All the villages shared a relatively recent peak in the area of bare soil, in 1996 or 2006/2010. The area of bare soil has been very limited in Boursouma and Oula over the whole period. Woodland was present in smaller areas in most villages in the past, but has now vanished. The existence of this social-ecological patch was confirmed by the villagers in the focus group discussions. The area of depressions has been relatively stable through time, as they are mainly topographically defined. Public land is mainly included in the maps to avoid overestimation of the homestead areas. In Lebda, a dam was constructed in 2009, transforming the northwest corner of the village into a small reservoir. In 2016, irrigated vegetables had emerged as a distinct social-ecological patch in Lebda, but this was not included in the assessment of change in ecosystem services because of its limited spatial extent.

Change in ecosystem services over time

The overall trends of change in social-ecological patches were surprisingly similar in all villages (Fig. 3b), therefore we present the results on change in ecosystem services over time as averages for all villages. We describe the results of our plausible interval model, and then contrast these results to those from the linear model.

Since 1952 cultivated ecosystem services from cereals, legumes, and vegetables have increased. Vegetables increased by a factor of 4–23 times, while cereals and legumes increased 1.5–7 times (Fig. 4a). For leaf vegetables from herbs and resources for livestock it is not possible to discern whether there has been an increase or decrease. Medicine, construction material, fruits, and leaf vegetables from trees decreased. Leaf vegetables from trees and fruit decreased less, maintaining 90–50% and 80–40% of their respective 1952 levels. Medicine, firewood, and construction material declined to 66–20% of their previous values. Only three estimates from the linear model, for legumes, cereals, and livestock, were within the intervals estimated by our plausible interval model. The linear model estimated less change in ecosystem services than the plausible interval model for all services except leaf vegetables from herbs, where the linear model estimated an increase and the plausible interval estimated a decrease.

Population density has almost doubled over the period 1960–2006 (Table 4). The generation of the cultivated ecosystem services cereals, legumes, and vegetables have almost kept up with this population increase, or increased slightly per capita (Fig. 4b). Leaf vegetables from herbs has kept up with population growth or declined somewhat per capita. All other ecosystem services have declined per capita. Firewood declined to about a third of its level in 1952 even in the linear estimate, and has, together with medicine, fruits, construction material, and leaf vegetables from trees decreased to 11–54% of their 1952 levels. Informants described how they nowadays need to walk long distances to collect the firewood they need.

By comparing our estimates of ecosystem services with a scenario in which there were no trees and shrubs in croplands, we demonstrate that cropland trees substantially contribute to availability of leaf vegetables from trees, fruits, medicine, firewood, and construction material (Fig. 5). Thus, the decline in ecosystem services generation over time would have been much higher if trees and shrubs were not conserved on fields in the farming system in the study area.

DISCUSSION

We first reflect on the accuracy of our mapping of social-ecological patches, estimates of population density, and simple model for historical extrapolation of ecosystem services. We discuss its value for studies of change in ecosystem services, and how the model could be improved and extended. Second, we discuss the challenges for landscape-based livelihoods in Burkina Faso that our historic analysis indicates, and their policy implications.

Image interpretation and population data

Interpretation of aerial photographs and satellite images is affected by seasonal differences between photo dates. The elements that we used for the identification of social-ecological patches such as trees and shrubs, field borders, permanent soil and water conservation structures, and buildings are visible all seasons. We are therefore confident that seasonal differences between images taken in the beginning of the dry season (November) and in the end of the dry season (April) did not have a major impact on the results. The aim of our approach was to map trends in change in social-ecological patches and ecosystem services over time. We were therefore less interested in exact map accuracies for exact points. When working with historical images it is difficult to assess classification accuracy as other reference material is seldom complete (Manies and Mladenoff 2000, Schulte et al. 2002). For the study sites, neither ground truthing points nor historical land use maps are available. However, the interpretation of the images in this study was facilitated by the knowledge of the landscape from previous in-depth fieldwork. To study change in larger areas, some automation to identify social-ecological patches in images would be necessary. For recent remote sensing images, Malmborg et al. (2018) developed a hybrid classification approach to address this. Even though we were unable to map fallows, the set of provisioning ecosystem services from fallows can be represented by the combination of shrublands and fields that the actual fallows are classified as in our maps. However, excluding fallows as a separate social-ecological patch would be problematic in an assessment of regulating ecosystem services because fallows help restore soil fertility through accumulation of nutrients and organic material.

We used data on population densities for the smallest unit where reliable data were available. Data at village level were available for some years, but have been proven unreliable at least for some villages from fieldwork experience. However, this village level data suggest 2–3 times higher population densities in Boursouma, Oula, and Koalma, and 1.25–2 times higher population densities in Zarin, as compared to the regional densities used in our analysis (Marchal 1983, Government of Burkina Faso 1988, 2000, 2005). This indicates that the decrease in ecosystem services per capita might actually be larger than our results show (Fig. 4b). However, in Reko and Lebda, the population densities might be up to 30% lower than the regional densities used. Actual population in the villages varies within a year, which is not captured in the population data. Young men in particular migrate to cities or neighboring countries to work during the dry season and return for the harvest season. This means that there are seasonal differences in beneficiaries of ecosystem services, a lack of labor force for investment in the landscape, e.g., when fields with soil and water conservation structures are prepared, and an inflow of financial capital.

Landscape variability

Landscapes vary over both time and space in ways that are poorly known. Key issues within patches include variation in crop yield and composition of woody vegetation. Key issues within the landscape include configuration and interactions between patches, and the emergence of new types of social-ecological patches. Soil and water conservation techniques are widespread investments to increase water infiltration and storage and thereby yields in the study area, particularly in villages in region Nord. After the droughts in the 1970s and 1980s, village committees led the construction of stone bunds empowered by customary chiefs and elders and supported by development projects (Batterbury 2010). A farmer in the region frequently mentioned by informants, Yacouba Sawadogo, started to improve traditional planting pits, zaï, in the 1980s by making them larger and adding manure, a practice that spread over large areas through farmer-to-farmer study visits (Pasiecznik and Reij 2020). Sawadogo was a Right Livelihood Award laureate in 2018 and UNEP Champion of the Earth 2020. The construction of stone bunds and zaï is labor intensive, and it is more likely that these techniques are adopted by farmers who are trusted and can mobilize mutual assistance work, or head larger or wealthier households (Yaméogo et al. 2018). Yields have also been enhanced by increased use of manure and fertilizers, animal traction, and improved crop varieties adapted to a shorter rainy season. However, yield data is available at a relatively coarse scale while yields are highly variable, even within fields. In addition, there is seasonal variation in generation of ecosystem services, where crops are grown during one rainy season, and many wild foods harvested during a single period each year, as well as interannual variation in generation due to rainfall variability.

A second key source of variation among patches are differences in the density and species composition of woody vegetation. There is only partial data on these changes. In the villages in region Nord, woody vegetation decreased after the droughts in the 1970s and 1980s, but then recovered until 2006 (Hänke et al 2016). However, the species composition has shifted to more drought tolerant and introduced species, which do not generate the same set of ecosystem services. Informants in our study reported a decrease in species diversity in shrublands, which has made it more difficult to find fruits, leaves, and products for traditional medicinal use. On the contrary, the introduced fast-growing species may increase availability of construction material and firewood (Hänke et al. 2016). Regional differences in the quality of woody vegetation in social-ecological patches is a potential explanation of the lack of difference between the two study regions, despite their different NDVI trends that motivated the site selection (Fig. 1). Vegetation greenness measured as NDVI is potentially more affected by differences in vegetation density and species composition within social-ecological patches, than by changes between social-ecological patch classes. Studies from Senegal have shown that it is difficult to relate differences in vegetation greenness between pixels to changes in landscape composition and livelihoods that are measured and perceived on the ground (Herrmann and Tappan 2013, Herrmann et al. 2014). Including better assessment of woody resources quality in patches (see, e.g., Hänke et al. 2016 for methods) would improve estimates of ecosystem services, but is difficult to apply to this long-time series of historical images because of the lack of data of historical species composition.

Along with changes within patches, changes in the landscape composition and configuration can also alter the supply of ecosystem services. This study used total patch area, but not relationships among patches or patch configuration to estimate the supply of ecosystem services. Better understanding of how people use different patches during different seasons would potentially allow configuration of patches to be used to better estimate changes in ecosystem services. For example, if shrubland patches decrease below a certain size, they become too small to keep livestock and to maintain diversity of medicinal plants and fruits (Malmborg et al. 2018).

Despite these aspects, we argue that this approach is valuable for the literature on change in ecosystem services over time because of its simplicity and transparency, which allows estimates to be easily refined when additional observations of changes in availability or value to people at different points in time become available for these landscapes. The linear extrapolation, the high estimate of change, and the low estimate of change all have a consistent direction of change for almost all ecosystem services. This indicates a robust estimation of change in ecosystem services. For some services, the plausible range of change is large, indicating that the underlying assumptions are important, and that a more precise understanding of how relationships between people and nature have changed is needed to reduce uncertainty. The approach of analyzing a range of plausible change in ecosystem services over time can also be used in other settings, particularly where data availability is low.

Challenges for future livelihoods in rural Burkina Faso

In interviews, villagers stressed that while they diversify their activities to earn cash income, they depend on the local landscape for their livelihoods. This is consistent with occupation and income data from a national survey (INSD 2014), showing that despite the general importance of migrant remittances (Batterbury 2010), more than half of the households do not report remittances. Our results indicate that field expansion in the studied village landscapes in northern Burkina Faso is approaching limits, both because of lack of additional land that can be meaningfully cultivated, and because of lack and fragmentation of shrubland for livestock grazing. Farmers in Boursouma and Oula, for example, set aside fields for livestock to prevent their livestock from grazing on other farmers’ land, because shrubland is too scarce. Farmers do not have enough crop residues or other fodder sources to replace grazing with fodder during the cropping season. Specialized pastoralists with larger herds are present in the area, and field expansion can exacerbate already occurring conflicts between crop producers and livestock producers (Sanfo et al. 2015). Land is no longer sufficient to leave fields fallow according to informants, and if they do leave fields uncultivated, the high pressure of livestock grazing is expected to degrade rather than increase soil fertility. The decline of the fallow practice increases the need for external input of nutrients on fields to maintain or increase yields. Manure from farmers’ own cattle is insufficient, which means that access to manure depends on social capital to maintain or establish good relationships with livestock herders. Mineral fertilizers can be purchased, but farmers can often afford only limited quantities.

Concern that conversion of shrubland to fields is approaching saturation has been raised in other parts of Burkina Faso (Hansen and Reenberg 1998, Knauer et al. 2017, Jahel et al. 2018). In parts of the Sahel, where cropland area has been stable, this stability has been explained by access to improved agricultural practices (Reij et al. 2005, Tappan and McGahuey 2007) and national or international markets (van Vliet et al. 2013). If that is true in our study area, it means that not enough farmers have had access to improved agricultural practices and/or local, national, or international markets. Our study shows that although changed agricultural practices such as soil and water conservation methods have improved yields, the increase has neither been enough to hinder expansion of cultivated land, nor to lift people out of poverty. In the regions Centre-Nord and Nord, 47% and 70% of the population, respectively, lives in poverty, according to nationally defined income levels (corresponding to approximately US$0.75 per day). These poverty rates are higher than the national average of 40% (Government of Burkina Faso 2016).

Avoiding trade-offs in agricultural landscapes with trees

To further intensify crop production, multiple challenges must be navigated. Farmers’ capacity to apply manure and/or mineral fertilizer has become central to gain a good or even acceptable yield. Farmers do apply compost or manure, but the availability of manure is generally insufficient for the area cultivated (Kambire 2016), and mineral fertilizers are often too expensive for farmers to apply to the whole field area. Increased temperatures and more irregular rains are expected in the area (Salack et al. 2016, Sultan and Gaetani 2016). These changes would make it more difficult for farmers to increase yields of rainfed crops and increase the need for water management, which can reduce insecurity and increase incomes for farmers. In this context, access to and management of small-scale reservoirs is important. The construction of a dam in Lebda in 2009 was a major change for the population and informants stressed the much higher income per hectare from vegetable cultivation as compared to millet production. Because of the rapid increase in dams in Burkina Faso in general (Cecchi et al. 2009, Ofosu et al. 2010), irrigated land for dry season vegetable production that we identified as a new patch type, should be considered in regional analyses or projections of future changes. This patch type may be limited in surface area, but has the potential to contribute substantially to the diversification of production and increase of farmers’ income. Life in the study area, particularly the region Centre-Nord, is today challenged by numerous terrorist attacks targeting civilians. This has resulted in more than 1 million internally displaced persons since the beginning of 2019 (UNHCR 2021). This violence is an immediate threat to people, and hinders investments in land and livestock.

All non-cultivated ecosystem services have decreased per capita over the study period (Fig. 4b). Population is rapidly growing in Burkina Faso; in 1960 the population was under 5 million people, whereas in 2019 it was over 20 million (World Bank 2019). Currently population growth rates remain high, at 3.1% per year (Government of Burkina Faso 2016). These high rates of population growth are likely to persist, as the population is forecast to increase for the rest of the century (FAO 2017). This rapid population growth presents opportunities and risks for Burkina Faso, and means per capita availability of ecosystem services is likely to continue to decline without unprecedented changes to land management. The vast majority of people depend on firewood for cooking (more than 95% according to Government of Burkina Faso 2006a, 2006b), and there is already a deficit in firewood and charcoal in relation to the demand in the regions Nord and Centre-Nord (Kambire et al. 2015). This decline in fuel availability highlights the urgent need to develop rural access to more sustainable energy sources. Decline in wild foods is likely to have negative effects on nutrition, as many have much higher nutritional values than cultivated vegetables, and may not be available or affordable for people to purchase. For example, frequently cited food items from trees such as leaves of Adansonia digitata, nuts of Sclerocarya birrea, and calyx of Bombax costatum, have high concentrations of different minerals (Smith et al. 1996), and food leaves from herbs such as Corchorus tridens and Amaranthus hybridus have high values of proteins and mineral nutrients (Humphry et al. 1993). Decreased availability of firewood in particular, but also wild foods affect the time available for women, because mainly women are responsible for their collection (see, e.g., Reij et al. 2005, Perez et al. 2015) and need to spend more time collecting when resources are scarce.

Improved productivity in agriculture is a main goal of Burkina Faso‛s government (Government of Burkina Faso 2016) and international development donors, who also focus on increasing yields in African smallholder agriculture, including in Burkina Faso (see Moseley 2017 and Gengenbach et al. 2018 for critical perspectives). Our analysis shows that non-crop provisioning ecosystem services from croplands are essential to people’s livelihoods in these multifunctional village landscapes. If the croplands in this study (fields with trees, depressions, and homesteads) had only been assigned values for crops, we would have seen much lower levels of leaf vegetables from trees, fruits, medicine, firewood, and construction material across all time steps and estimates (Fig. 5). The importance of trees and shrubs for generation of provisioning as well as regulating ecosystem services in the Sahel has also been previously documented (see, e.g., Bayala et al. 2014, Sinare and Gordon 2015). Maintenance of trees and shrubs on fields is closely linked to development of social capital in a context where colonial agricultural extension services focused on removal of trees from fields, a practice that persisted for decades (Reij and Garrity 2016). Building of trust between farmers and extension services, transfer of authority to village committees for land and vegetation management, and capacity development are examples of this development of social capital (Hilhorst 2008, Batterbury 2010, Pasiecznik and Reij 2020). Our study illustrates the importance to consider the multiple ecosystem services generated from trees and shrubs in any landscape intervention to avoid trade-offs, for example, when designing strategies to improve crop productivity, or in the establishment of mining activities, which is an important income source in the regions and for Burkina Faso. We hope that the findings and methods developed in this research can help guide future landscape management decisions.

CONCLUSION

We estimated historical ecosystem services in village landscapes in Burkina Faso, using a simple model that estimates the range over which the social-ecological production and use of ecosystem services has changed. These estimates are based on a combination of image interpretation, field data, and literature data. The approach has the strengths of being simple, transparent, and easy to refine if additional data on change in generation or value to people become available. We believe that this approach will be useful to estimate historical ecosystem services in other regions in which there is limited historical data.

Our analysis revealed that although the generation of cultivated ecosystem services has almost kept up with population growth or increased somewhat per capita, the potential to generate other provisioning ecosystem services has substantially declined. Fields cover between 46% and 78% of village area and grazing areas are increasingly scarce. Historical analysis indicates that there is little potential for further expansion of cultivated area or grazed area. Although yields have increased, future intensification is limited by lack of access to reliable and sustained access to water from percolation and storage of surface water through natural or artificial structures, lack of access to manure and fertilizers, and yield reductions due to climate change. Alternatives for some non-cultivated ecosystem services like medicine and construction material are becoming increasingly available. However, there are currently limited alternatives for other ecosystem services. Today, the majority of Burkina Faso’s population depends on wood for cooking. Furthermore, there are no alternatives to compensate for the declining availability of nutritious wild foods. The ecosystem services provided by trees and shrubs on fields do not have accessible alternatives, therefore trees and shrubs should be preserved in landscape interventions.

LITERATURE CITED

Batterbury, S. 2010. Land, environmental management and the new governance in Burkina Faso. Pages 241-262 in L. Godden and M. Tehan, editors. Comparative perspectives on communal lands and individual ownership: sustainable futures. Routledge, London, UK.

Bayala, J., J. Sanou, Z. Teklehaimanot, A. Kalinganire, and S. J. Ouédraogo. 2014. Parklands for buffering climate risk and sustaining agricultural production in the Sahel of West Africa. Current Opinion in Environmental Sustainability 6(1):28-34. https://doi.org/10.1016/j.cosust.2013.10.004

Bennett, E. M., W. Cramer, A. Begossi, G. Cundill, S. Díaz, B. N. Egoh, I. R. Geijzendorffer, C. B. Krug, S. Lavorel, E. Lazos, L. Lebel, B. Martín-López, P. Meyfroidt, H. A. Mooney, J. L. Nel, U. Pascual, K. Payet, N. P. Harguindeguy, G. D. Peterson, A. H. Prieur-Richard, B. Reyers, P. Roebeling, R. Seppelt, M. Solan, P. Tschakert, T. Tscharntke, B. L. Turner, P. H. Verburg, E. F. Viglizzo, P. C. L. White, and G. Woodward. 2015. Linking biodiversity, ecosystem services, and human well-being: three challenges for designing research for sustainability. Current Opinion in Environmental Sustainability 14:76-85. https://doi.org/10.1016/j.cosust.2015.03.007

Berbés-Blázquez, M., J. A. González, and U. Pascual. 2016. Towards an ecosystem services approach that addresses social power relations. Current Opinion in Environmental Sustainability 19:134-143. https://doi.org/10.1016/j.cosust.2016.02.003

Bivand, R., T. Keitt, B. Rowlingson. 2019. rgdal: Bindings for the ‘geospatial.’ Data Abstraction Library. R package version 1.4-8. https://CRAN.R-project.org/package=rgdal

Brandt, M., P. Hiernaux, K. Rasmussen, C. Mbow, L. Kergoat, T. Tagesson, Y. Z. Ibrahim, A. Wélé, C. J. Tucker, and R. Fensholt. 2016. Assessing woody vegetation trends in Sahelian drylands using MODIS based seasonal metrics. Remote Sensing of Environment 183:215-225. https://doi.org/10.1016/j.rse.2016.05.027

Brandt, M., K. Rasmussen, P. Hiernaux, S. Herrmann, C. J. Tucker, X. Tong, F. Tian, O. Mertz, L. Kergoat, C. Mbow, J. L. David, K. A. Melocik, M. Dendoncker, C. Vincke, and R. Fensholt. 2018. Reduction of tree cover in West African woodlands and promotion in semi-arid farmlands. Nature Geoscience 11(5):328-333. https://doi.org/10.1038/s41561-018-0092-x

Brandt, M., C. J. Tucker, A. Kariryaa, K. Rasmussen, C. Abel, J. Small, J. Chave, L. V. Rasmussen, P. Hiernaux, A. A. Diouf, L. Kergoat, O. Mertz, C. Igel, F. Gieseke, J. Schöning, S. Li, K. Melocik, J. Meyer, S. Sinno, E. Romero, E. Glennie, A. Montagu, M. Dendoncker, and R. Fensholt. 2020. An unexpectedly large count of trees in the West African Sahara and Sahel. Nature 587(7832):78-82. https://doi.org/10.1038/s41586-020-2824-5

Breusers, M. 2001. Pathways to negotiate climate variability: land use and institutional change in the Kaya region, Burkina Faso. African Studies Centre, Leiden, The Netherlands.

Bürgi, M., J. Silbernagel, J. Wu, and F. Kienast. 2015. Linking ecosystem services with landscape history. Landscape Ecology 30:11-20. https://doi.org/10.1007/s10980-014-0102-3

Cecchi, P., A. Meunier-Nikiema, N. Moiroux, and B. Sanou. 2009. Towards an atlas of lakes and reservoirs in Burkina Faso. Pages 1-23 in M. Andreini, editor. Small Reservoirs Toolkit. International Water Management Institute, Colombo, Sri Lanka.

Chan, K. M. A., T. Satter, and J. Goldstein. 2012. Rethinking ecosystem services to better address and navigate cultural values. Ecological Economics 74:8-18. https://doi.org/10.1016/j.ecolecon.2011.11.011

Daw, T., K. Brown, S. Rosendo, and R. Pomeroy. 2011. Applying the ecosystem services concept to poverty alleviation: the need to disaggregate human well-being. Environmental Conservation 38(4):370-379. https://doi.org/10.1017/S0376892911000506

Daw, T. M., C. C. Hicks, K. Brown, T. Chaigneau, F. A. Januchowski-Hartley, W. W. L. Cheung, S. Rosendo, B. Crona, S. Coulthard, C. Sandbrook, C. Perry, S. Bandeira, N. A. Muthiga, B. Schulte-Herbrüggen, J. Bosire, and T. R. McClanahan. 2016. Elasticity in ecosystem services: exploring the variable relationship between ecosystems and human well-being. Ecology and Society 21(2):11. https://doi.org/10.5751/ES-08173-210211

Environmental Systems Research Institute (ESRI). 2013. ArcGIS Desktop: Release 10.3. Environmental Systems Research Institute, Redlands, California, USA.

Ernstson, H. 2013. The social production of ecosystem services: a framework for studying environmental justice and ecological complexity in urbanized landscapes. Landscape and Urban Planning 109(1):7-17. https://doi.org/10.1016/j.landurbplan.2012.10.005

Fensholt, R., K. Rasmussen, T. Theis Nielsen, and C. Mbow. 2009. Evaluation of earth observation based long term vegetation trends - intercomparing NDVI time series trend analysis consistency of Sahel from AVHRR GIMMS, Terra MODIS and SPOT VGT data. Remote Sensing of Environment 113:1886-1898. https://doi.org/10.1016/j.rse.2009.04.004

Food and Agriculture Organization (FAO). 2017. The future of food and agriculture: trends and challenges. FAO, Rome, Italy.

Gengenbach, H., R. A. Schurman, T. J. Bassett, W. A. Munro, and W. G. Moseley. 2018. Limits of the New Green Revolution for Africa: reconceptualising gendered agricultural value chains. Geographical Journal 184:208-214. https://doi.org/10.1111/geoj.12233

Government of Burkina Faso. 1988. Recensement general de la population 1985: structure par age et sexe des villages du Burkina Faso. Ouagadougou, Burkina Faso.

Government of Burkina Faso. 2000. Analyse des resultats du recensement général de la population et de l‛habitation de 1996. Ouagadougou, Burkina Faso.

Government of Burkina Faso. 2005. DECRET N°2005___/PRES/PM/MATD portant répertoire des villages administratifs et des secteurs de communes du Burkina Faso. Ouagadougou, Burkina Faso.

Government of Burkina Faso. 2006a. Recensement général de la population et de l‛habitation de 2006 - monographie de la region du Centre Nord. Ouagadougou, Burkina Faso.

Government of Burkina Faso. 2006b. Recensement général de la population et de l‛habitation de 2006 - Monographie de la region du Nord. Ouagadougou, Burkina Faso.

Government of Burkina Faso. 2016. PNDES Plan National de Développement Économique et Social. Ouagadougou, Burkina Faso.

Hänke, H., L. Börjeson, K. Hylander, and E. Enfors-Kautsky. 2016. Drought tolerant species dominate as rainfall and tree cover returns in the West African Sahel. Land Use Policy 59:111-120. https://doi.org/10.1016/j.landusepol.2016.08.023

Hansen, T. S., and A. Reenberg. 1998. Approaching local limits to field expansion - land use pattern dynamics in semi-arid Burkina Faso. Danish Journal of Geography 98(1):56-70. https://doi.org/10.1080/00167223.1998.10649411

Herrmann, S. M., A. Anyamba, and C. J. Tucker. 2005. Recent trends in vegetation dynamics in the African Sahel and their relationship to climate. Global Environmental Change 15(4):394-404. https://doi.org/10.1016/j.gloenvcha.2005.08.004

Herrmann, S. M., I. Sall, and O. Sy. 2014. People and pixels in the Sahel: a study linking coarse-resolution remote sensing observations to land users’ perceptions of their changing environment in Senegal. Ecology and Society 19(3):29. https://doi.org/10.5751/ES-06710-190329

Herrmann, S. M., and G. G. Tappan. 2013. Vegetation impoverishment despite greening: a case study from central Senegal. Journal of Arid Environments 90:55-66. https://doi.org/10.1016/j.jaridenv.2012.10.020

Hijmans, R. J. 2020. raster: Geographic data analysis and modeling. R package version 3.0-12. https://CRAN.R-project.org/package=raster

Hilhorst, T. 2008. Local governance institutions for sustainable natural resource management in Mali, Burkina Faso and Niger. Royal Tropical Institute, Amsterdam, The Netherlands.

Humphry, C. M., M. S. Clegg, C. L. Keen, and L. E. Grivetti. 1993. Food diversity and drought survival. The Hausa example. International Journal of Food Sciences and Nutrition 44(1):1-16. https://doi.org/10.3109/09637489309017417

Institut National de la Statistique et de la Démographie (INSD). 2009. Recensement general de la population et de l‛habitation (RGPH) de 2006. Analyse des resultats definitifs. Theme 2: Etat et structure de la population. INSD, Ouagadougou, Burkina Faso.

Institut National de la Statistique et de la Démographie (INSD). 2014. Enquête Multisectorielle Continue 2014. INSD, Ouagadougou, Burkina Faso. https://microdata.worldbank.org/index.php/catalog/2538/get-microdata

Jahel, C., E. Vall, Z. Rodriguez, A. Bégué, C. Baron, X. Augusseau, and D. Lo Seen. 2018. Analysing plausible futures from past patterns of land change in West Burkina Faso. Land Use Policy 71:60-74. https://doi.org/10.1016/j.landusepol.2017.11.025

Kambire, F. C. 2016. Effet combiné du travail du sol et de la gestion de la fumure organique dans l‛agrosystème cotonnier au Burkina Faso. Dissertation. Université Catholique de Louvain, Louvain-la-Neuve, Belgium.

Kambire, H. W., I. N. S. Djenontin, A. Kabore, H. Djoudi, M. P. Balinga, M. Zida, and S. Assembe-Mvondo. 2015. La REDD+ et l‛adaptation aux changements climatiques au Burkina Faso: causes, agents et institutions. Document Occasionnel 123. CIFOR, Bogor, Indonesia.

Kassambara, A. 2020. ggpubr: ggplot2 Based publication ready plots. R package version 0.2.5. https://CRAN.R-project.org/package=ggpubr

Knauer, K., U. Gessner, R. Fensholt, G. Forkuor, and C. Kuenzer. 2017. Monitoring agricultural expansion in Burkina Faso over 14 years with 30 m resolution time series: the role of population growth and implications for the environment. Remote Sensing 9:132. https://doi.org/10.3390/rs9020132

Leroux, L., A. Bégué, D. Lo Seen, A. Jolivot, and F. Kayitakire. 2017. Driving forces of recent vegetation changes in the Sahel: lessons learned from regional and local level analyses. Remote Sensing of Environment 191:38-54. https://doi.org/10.1016/j.rse.2017.01.014

Malmborg, K., H. Sinare, E. E. Kautsky, I. Ouedraogo, and L. J. Gordon. 2018. Mapping regional livelihood benefits from local ecosystem services assessments in rural Sahel. PLoS ONE 13(2):e0192019. https://doi.org/10.1371/journal.pone.0192019

Manies, K. L., and D. J. Mladenoff. 2000. Testing methods to produce landscape-scale presettlement vegetation maps from the U.S. public land survey records. Landscape Ecology 15:741-754. https://doi.org/10.1023/A:1008115200471

Marchal, J.-Y. 1983. Yatenga, Nord Haute-Volta: La dynamique d‛un espace rural Soudano-Sahelien. ORSTOM, Paris, France.

Millennium Ecosystem Assessment. 2005. Ecosystems and human well-being: synthesis. Island Press, Washington, D.C., USA.

Ministère de l‛agriculture et de securité alimentaire. 2013. Agristat 2013. Ouagadougou, Burkina Faso.

Moseley, W. G. 2017. The New Green Revolution for Africa: a political ecology critique. Brown Journal of World Affairs XXIII(II):177-190.

Nicholson, E., G. M. Mace, P. R. Armsworth, G. Atkinson, S. Buckle, T. Clements, R. M. Ewers, J. E. Fa, T. A. Gardner, J. Gibbons, R. Grenyer, R. Metcalfe, S. Mourato, M. Muûls, D. Osborn, D. C. Reuman, C. Watson, and E. J. Milner-Gulland. 2009. Priority research areas for ecosystem services in a changing world. Journal of Applied Ecology 46:1139-1144. https://doi.org/10.1111/j.1365-2664.2009.01716.x

OECD/SWAC. 2014. An atlas of the Sahara-Sahel: Geography, economics and security. West African studies, Organisation for Economic Cooperation and Development Publishing, Paris, France.

Ofosu, E. A., P. van Der Zaag, N. C. van De Giesen, and S. N. Odai. 2010. Productivity of irrigation technologies in the White Volta basin. Physics and Chemistry of the Earth 35(13-14):706-716. https://doi.org/10.1016/j.pce.2010.07.005

Olsson, L., L. Eklundh, and J. Ardö. 2005. A recent greening of the Sahel - trends, patterns and potential causes. Journal of Arid Environments 63(3):556-566. https://doi.org/10.1016/j.jaridenv.2005.03.008

Ozer, P. 2004. Bois de feu et déboisement au Sahel: mise au point. Secheresse 15(3):243-251.

Palomo, I., M. R. Felipe-Lucia, E. M. Bennett, B. Martín-López, and U. Pascual. 2016. Disentangling the pathways and effects of ecosystem service co-production. Advances in Ecological Research 54:245-283. https://doi.org/10.1016/bs.aecr.2015.09.003

Pasiecznik, N. and C. Reij, editors. 2020. Restoring African Drylands. Tropenbos International, Wageningen, The Netherlands.

Perez, C., E. M. Jones, P. Kristjanson, L. Cramer, P. K. Thornton, W. Förch, and C. Barahona. 2015. How resilient are farming households and communities to a changing climate in Africa? A gender-based perspective. Global Environmental Change 34:95-107. https://doi.org/10.1016/j.gloenvcha.2015.06.003

Perpinan Lamigueiro, O., and R. Hijmans. 2019. rasterVis. R package version 0.47.

Powell, J. M., R. A. Pearson, and P. H. Hiernaux. 2004. Crop-livestock interactions in the West African drylands. Agronomy Journal 96(2):469-483. https://doi.org/10.2134/agronj2004.4690

R Core Team 2019. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/

Reenberg, A. 2012. Insistent dryland narratives: portraits of knowledge about human-environmental interactions in Sahelian environment policy documents. West African Journal of Applied Ecology 20(1):97-111.

Reij, C., and D. Garrity. 2016. Scaling up farmer-managed natural regeneration in Africa to restore degraded landscapes. Biotropica 48(6):834-843. https://doi.org/10.1111/btp.12390

Reij, C., G. Tappan, and A. Belemvire. 2005. Changing land management practices and vegetation on the Central Plateau of Burkina Faso (1968-2002). Journal of Arid Environments 63(3):642-659. https://doi.org/10.1016/j.jaridenv.2005.03.010

Renard, D., J. M. Rhemtulla, and E. M. Bennett. 2015. Historical dynamics in ecosystem service bundles. Proceedings of the National Academy of Sciences of the United States of America 112(43):13411-13416. https://doi.org/10.1073/pnas.1502565112

Republique de Haute-Volta. 1962. La situation démographique en Haute-Volta: résultats partiels de l‛énquête démographique 1960-1961. Institut national de la statistique et des études économiques, Paris, France.

Republique de Haute-Volta. 1978. Recensement général de la population Décembre 1975. Institut national de la statistique et de la démographie, Ouagadougou, Republique de Haute-Volta.

Reyers, B., R. Biggs, G. S. Cumming, T. Elmqvist, A. P. Hejnowicz, and S. Polasky. 2013. Getting the measure of ecosystem services: a social-ecological approach. Frontiers in Ecology and the Environment 11(5):268-273. https://doi.org/10.1890/120144

Ruelland, D., A. Tribotte, C. Puech, and C. Dieulin. 2011. Comparison of methods for LUCC monitoring over 50 years from aerial photographs and satellite images in a Sahelian catchment. International Journal of Remote Sensing 32(6):1747-1777. https://doi.org/10.1080/01431161003623433

Salack, S., C. Klein, A. Giannini, B. Sarr, O. N. Worou, N. Belko, J. Bliefernicht, and H. Kunstman. 2016. Global warming induced hybrid rainy seasons in the Sahel. Environmental Research Letters 11:104008. https://doi.org/10.1088/1748-9326/11/10/104008

Sanfo, A., I. Savadogo, K. E. Abalo, and Z. Nouhoun. 2015. Climate change: a driver of crop farmers - agro pastoralists conflicts in Burkina Faso. International Journal of Applied Science and Technology 5(3):92-104.

Schulte, L. A., D. J. Mladenoff, and E. V. Nordheim. 2002. Quantitative classification of a historic northern Wisconsin (U.S.A.) landscape: mapping forests at regional scales. Canadian Journal of Forest Research 32:1616-1638. https://doi.org/10.1139/x02-082

Sendzimir, J., C. P. Reij, and P. Magnuszewski. 2011. Rebuilding resilience in the Sahel: regreening in the Maradi and Zinder Regions of Niger. Ecology and Society 16(3):1. https://doi.org/10.5751/ES-04198-160301

Sinare, H., and L. J. Gordon. 2015. Ecosystem services from woody vegetation on agricultural lands in Sudano-Sahelian West Africa. Agriculture, Ecosystems and Environment 200:186-199. https://doi.org/10.1016/j.agee.2014.11.009

Sinare, H., L. J. Gordon, and E. E. Kautsky. 2016. Assessment of ecosystem services and benefits in village landscapes - a case study from Burkina Faso. Ecosystem Services 21:141-152. https://doi.org/10.1016/j.ecoser.2016.08.004

Smith, G. C., M. S. Clegg, C. L. Keen, and L. E. Grivetti. 1996. Mineral values of selected plant foods common to southern Burkina Faso and to Niamey, Niger, West Africa. International Journal of Food Sciences and Nutrition 47(1):41-53. https://doi.org/10.3109/09637489609028560

Souverijns, N., M. Buchhorn, S. Horion, R. Fensholt, H. Verbeeck, J. Verbesselt, M. Herold, N. E. Tsendbazar, P. N. Bernardino, B. Somers, and R. Van De Kerchove. 2020. Thirty years of land cover and fraction cover changes over the Sudano-Sahel using landsat time series. Remote Sensing 12:3817. https://doi.org/10.3390/rs12223817

Sultan, B., and M. Gaetani. 2016. Agriculture in West Africa in the twenty-first century: climate change and impacts scenarios, and potential for adaptation. Frontiers in Plant Science 7:1262. https://doi.org/10.3389/fpls.2016.01262

Swift, J. 1996. Desertification: narratives, winners & losers. Pages 73-90 in M. Leach and R. Mearns, editors. The lie of the land. International African Institute, London, UK.

Tappan, G., and M. McGahuey. 2007. Tracking environmental dynamics and agricultural intensification in southern Mali. Agricultural Systems 94:38-51. https://doi.org/10.1016/j.agsy.2005.07.011

Tappan, G. G., M. Sall, E. C. Wood, and M. Cushing. 2004. Ecoregions and land cover trends in Senegal. Journal of Arid Environments 59:427-462. https://doi.org/10.1016/j.jaridenv.2004.03.018

Tomscha, S. A., and S. E. Gergel. 2016. Ecosystem service trade-offs and synergies misunderstood without landscape history. Ecology and Society 21(1):43. https://doi.org/10.5751/ES-08345-210143

Tougiani, A., C. Guero, and T. Rinaudo. 2009. Community mobilisation for improved livelihoods through tree crop management in Niger. GeoJournal 74:377. https://doi.org/10.1007/s10708-008-9228-7

United Nations Development Programme (UNDP). 2015. Human development report 2015. UNDP, New York, New York, USA.

United Nations High Commissioner for Refugees (UNHCR). 2021. Burkina Faso Operational Update. UNHCR, Geneva, Switzerland.

van Vliet, N., A. Reenberg, and L. V. Rasmussen. 2013. Scientific documentation of crop land changes in the Sahel: a half empty box of knowledge to support policy? Journal of Arid Environments 95:1-13. https://doi.org/10.1016/j.jaridenv.2013.03.010

Wickham, H. 2016. ggplot2: Elegant graphics for sata Analysis. Springer-Verlag, New York, New York, USA.

Wilkinson, C., T. Saarne, G. D. Peterson, and J. Colding. 2013. Strategic spatial planning and the ecosystem services concept - an historical exploration. Ecology And Society 18(1):37. https://doi.org/10.5751/ES-05368-180137

World Bank. 2019. Population, total - Burkina Faso. World Bank, Washington, D.C., USA. https://data.worldbank.org/indicator/SP.POP.TOTL?end=2019&locations=BF&start=1960&view=chart

Yaméogo, T. B., W. M. Fonta, and T. Wünscher. 2018. Can social capital influence smallholder farmers’ climate-change adaptation decisions? Evidence from three semi-arid communities in Burkina Faso, West Africa. Social Sciences 7(3):33. https://doi.org/10.3390/socsci7030033