Trees in agricultural landscapes enhance provision of ecosystem services in Sub-Saharan Africa

ABSTRACT

The recent interest in multi-functional agricultural landscapes has not been matched with formal assessment of the roles that trees play across the spectrum of ecosystem services (ESs) provided in Sub-Saharan Africa (SSA). A structured literature review (1995–2014) assessed 350 journal articles about provision of one or more ESs by trees on farms and in agricultural landscapes in SSA. This revealed information on 15 ESs from studies in 23 countries covering arid (1% of studies), semi-arid (49%), sub-humid (26%) and humid (24%) agro-ecological zones. The majority of the studies reported provisioning (39%) and supporting (35%) followed by regulating (26%) ESs while studies on cultural services were scarce. Beneficial impacts of trees were dominating (58%), in particular in semi-arid zones where they were associated with enhancing water and nutrient cycling. A decline in some ESs was reported in 15% of the studies, while 28% found no effect of trees. Although the effects of trees were mainly positive, a decline in crop production was noted as a key trade-off against the provisions of ESs, such as modification of microclimate. This highlights the need to manage trade-offs among impacts of trees on ES provision to reduce competition and increase complementarity between trees and crops.

KEYWORDS:

• Agroforestry
• agricultural productivity
• biodiversity
• multi-functionality
• trade-offs

EDITED BY Christine Fürst

Introduction

Expansion and intensification of agricultural land is a major driver of the recent decline of biodiversity and ecosystem services (ESs) (Foley et al. 2005). This underscores the emerging value placed on multi-functional agricultural landscapes and their potential to supply multiple ESs (Tscharntke et al. 2011). For example, in addition to providing food and fibre, agricultural landscapes offer habitats for insects that deliver pollination and pest control services (Chaplin-Kramer & Kremen 2012; Morandin & Kremen 2013), and soil organisms that influence soil structure and nutrient cycling thus altering water and nutrient flows within the landscape (Barrios et al. 2012). A majority of the ESs in these landscapes are mediated by species diversity, including trees (Van Noordwijk et al. 2011).

ESs have particular relevance in Sub-Saharan Africa (SSA), where the majority of the human population dwell in rural areas and rely on ESs for their living through smallholder farming, pastoralism and fisheries (Egoh et al. 2012). Sub-Saharan Africa has high levels of poverty and diminishing ESs provided by natural and semi-natural habitats, e.g. due to deforestation. With increasing population, communities have been moving to marginal and fragile areas, and are especially vulnerable where regulating and supporting services get diminished, resulting in greater susceptibility to drought, pest and disease outbreaks or soil degradation (Boko et al. 2007). Enhanced ES provision is, therefore, critical for building resilience and improving food and nutrition security for smallholder farmers in SSA.

Trees are an integral part of many farming landscapes, where they provide a range of ESs that support livelihoods (Sinclair 1999). These trees are introduced into farms for particular reason(s), or are selectively retained after land use change. Farmers raise and manage a mix of exotic and indigenous trees in different ways to provide a variety of ESs (Nyaga et al. 2015). For example, leguminous trees grown for green manure or fodder can be managed in hedgerows to also stabilize soil against erosion and provide firewood (Angima et al. 2002; Mutegi et al. 2008). ESs provided by trees may be involved in both trade-offs (where some ESs increase while others decrease) and synergies (where multiple ESs are enhanced simultaneously) (Rodríguez et al. 2006; Raudsepp-Hearne et al. 2010). Trade-offs and synergies are caused by interactions among ESs (e.g. competition or complementarity) or simultaneous responses to the same driver(s) of ecosystem change (Bennett et al. 2009).

The recent surge of interest in ESs within agricultural landscapes requires formal assessment of the roles that trees play across the spectrum of ES provision now considered important in SSA. This review focuses on two related questions: (1) what is the scientific evidence base that integration of trees on farms and in agricultural landscapes contributes to enhanced ES provision for smallholder farmers in SSA, and which are the trade-offs?, (2) what quantitative studies are available, in which context and at which scale were they carried out, and what do they show? Whereas the importance of trees in provision of individual ES in SSA farms is widely studied; the question of which ESs increase or decrease when trees are incorporated in SSA agricultural landscapes has received less attention so far. Previous reviews and meta-analyses on the role of trees have focused on only one or a few selected ES at a time, for example impact of fertilizer trees on maize production (Sileshi et al. 2009; Akinnifesi et al. 2010) and the effect of agroforestry on pest control (Pumariño et al. 2015). This review aimed to (1) identify provisioning, supporting, regulating and cultural ESs of trees in agricultural landscapes in SSA; (2) determine the impacts of trees on ES provision and discuss how potential trade-offs can be managed; and (3) identify knowledge gaps that need to be further addressed by research.

A structured review with vote count was used to review, analyse and synthesize published data on the roles of trees in agricultural landscapes in order to quantify different ESs at field, farm and landscape scale in different agro-ecological (arid, semi-arid, sub-humid and humid) zones of SSA. This enabled us to capture the widest possible range of ESs.

Methods

Literature search and retrieval

The review process involved electronic search of material indexed in Web of Science and Scopus. A comprehensive list of ESs provided by trees was prepared from literature. The services were grouped into four categories following the millennium ecosystem assessment classification (MA 2005). Four search strings were set up retrieving documents indexed in Web of Science for ESs in each category (Table 1). This was supplemented by search on Google and Google scholar to identify other published articles and find web pages that might provide references. Published articles, theses, reports, conference proceedings and working papers were examined to identify potential peer reviewed publications that had arisen from the work, or presence of relevant citations.

Table 1. Search strategy used to retrieve publications from databases.

Ecosystem service (ES) categories^a	ES^b	Search strings^c
Provisioning services – products obtained from the ecosystem	Biochemical, e.g. dyes, perfumes, medicine Fibre/wood Food Fodder Fresh water Fuelwood Manure Genetic resources Gum Organic fertilizer Ornamental products	TS = [(charcoal OR compost OR fibre OR dye OR fodder OR fresh water OR fruit* OR fuel OR genetic diversity OR genetic material OR green manure OR gum OR medicin* OR mulch* OR nut* OR oil* OR perfume* OR spice OR tannin OR timber OR wood) AND (tree* OR agroforest* OR parkland* OR farm) AND Africa]
Regulating services – benefits obtained from the regulation of ecosystem processes	Air quality control Biodiversity Bio-control (pests and diseases) Carbon sequestration and storage Erosion control Flood control Habitat connectivity Microclimate Pollination Water purification (quality) Water regulation Wind breaks	TS = [(air quality OR biodiversity OR carbon* OR crop damage OR disease OR erosion OR evaporat* OR infiltrat* OR insect OR intercept* OR microclimate OR soil moisture OR natural enemi* OR parasit* OR pathogen OR pest OR pollinat* OR runoff OR shelterbelt OR transpiration OR water regulation OR weed OR wind) AND (tree* OR agroforest* OR parkland* OR farm) AND Africa]
Supporting services – natural processes that maintain other ESs	Biological nitrogen fixation Nutrient cycling Photosynthesis Soil fertility Water cycling	TS = [(carbon OR fertiliz* OR mineral OR nitrogen fix* OR nutrient OR nutrient cycl* OR soil fertility OR soil organic matter) AND (tree* OR agroforest* OR parkland* OR farm) AND Africa]
Cultural services – the non-material benefits obtained from ecosystems	Ecotourism Ornamental (live tree) Scent and sounds from birds and trees Shade/meeting places Sport (game hunting) Traditional, religious values	TS = [(aesthetic OR beaut* OR cultural OR tourism OR educat* OR ethnic OR heritage OR knowledge OR landscap* OR ornament* OR recreation OR spirit* OR tradition) AND (tree* OR agroforest* OR parkland* OR farm) AND Africa]

^a Classification according to Millennium Ecosystem Assessment (MA 2005).

^b ESs determined from literature as provided by trees within each category.

^c Search string used to retrieve document indexed in Web of Science for each category.

Identification of relevant studies

Studies retrieved were assessed for relevance using the title followed by the abstract and finally full text appraisal. Publications included in the study were those that satisfied the following criteria: (1) published peer reviewed journal article (for practical reasons limited to 20 years; 1995–2014), (2) study conducted on farms or in agricultural landscapes in SSA and (3) study reporting the ES provided or affected by trees (as compared to a non-tree treatment/situation).

Reviewer bias was reduced by having each of the 12 authors appraising a random sample of six publications. Consequently, about 20% of the studies were assessed by two authors to determine the repeatability of the selection criteria, the ES identified and the metadata obtained.

Summarizing the data

Each investigation that had been analysed separately was considered as an independent study. For example, if one paper presented results from the same experiment conducted at different locations and analysed separately, they were considered as two independent studies. A total of 361 peer-reviewed journal articles that reported 438 studies were reviewed (Table A1). Each study was categorized according to the FAO classification of agro-ecological zones (arid, semi-arid, sub-humid and humid).

Different types of information were extracted from each study reviewed: (1) where the study was conducted (region within SSA, country and name of the site and the geographic coordinates, agro-ecological zone); (2) how the study was conducted, including the design (experimental, observational or modelling), method of investigation (on-station or on-farm) and scale of investigation (field, farm, landscape); (3) number of ES reported, ES indicators and effect of trees on ES; (4) agroforestry practice, control treatment and the agroforestry components, such as tree species, crops and livestock. In addition, general information about the study (author, the year of publication and journal where the study was published), and biophysical characteristics of the site (climatic details, edaphic conditions, slope and altitude) were recorded.

A vote-count approach was used to determine the impact of trees on the ES, with results from each study being classified as either positive – meaning that the trees led to a significant increase in an ES; negative – that trees led to a significant decline in an ES; and neutral – where the influence of trees on an ES was not significant. The frequencies and proportions of studies with different effects of trees on a particular ES were calculated based on the overall number of studies reviewed.

Results

Distribution, context, design and spatial scale of studies

The review yielded 438 studies dispersed in 23 countries across SSA. The studies were carried out in arid (1%), semi-arid (49%), sub-humid (26%) and humid (24%) agro-ecological zones [Figure 1(a)]. The majority of the studies have been conducted in East and West Africa followed by Southern Africa whereas very few studies were carried out in Central Africa. Over half of the studies were conducted in four countries: Kenya, Nigeria, Cameroon and Malawi [Figure 1(b)]. A large proportion of the studies were from Kenya; evidence within the study that are specific to this country are identified so.

Figure 1. Distribution of studies in four agro-ecological zones: (a) the spatial representation of the studies on ecosystem services (ESs) across Sub-Saharan Africa, (b) the distribution of the studies on ESs in different countries. Studies in highlands are captured within the respective humid and sub-humid agro-ecological zones.

There was variation in the context, design and spatial scale at which research was conducted (Figure 2) with more than half of the studies conducted on-farm (57%) and the rest carried out on-station (43%). The majority of the reported studies were manipulative experiments (84%), followed by observational (15%) and modelling studies (1%). ESs in the studies reviewed were mainly documented at field scale (68%) and less at farm (23%) and landscape scale (9%).

Figure 2. Representation of the studies reporting ESs of trees according to the context, design and spatial scale of the research, and the relevant agro-ecological zones for: (a) provisioning, (b) regulating and (c) supporting ESs.

ESs affected by trees

Fifteen ESs potentially affected by trees were reported in the reviewed studies. Provisioning ESs were the leading category reported (39%), followed by supporting (35%) and regulating ES (26%); studies on cultural ES provided by trees were scarce (Figure 3). Over half of the studies reporting provisioning, regulating and supporting ESs were conducted in the semi-arid zone (Figure 2). Fifty two percent of the studies focused on single ES; the rest (48%) reported data on multiple ESs – averaging three ES per study. The majority were examined as part of a broader investigation, indicating the possibility of assessing ESs using methods designed to address broader research questions.

Figure 3. The occurrence of ESs highlighted in different studies across Sub-Saharan Africa.

Provisioning ESs were dominated by food provision (Figure 3). Trees influence the production of food from crops and livestock; such effects are reported in the majority of the papers in this category. Some trees also produce fruits and nuts that can be used for human consumption, e.g. Adansonia digitata (baobab), Parkia biglobosa (African locust bean), Sclerocarya birrea (marula), Vitellaria paradoxa (shea tree) and Ziziphus jujuba (jujube red date). Five percent of the studies reviewed considered trees as providers of fodder (Figure 3). A total of 49 fodder tree and shrub species were evaluated mainly for their nutritive value and effect on animal growth and milk production. Provisioning of green manure was reported in 4% of the studies; where tree foliage was carried from outside the cultivated fields and applied on crops to supply nutrients. Wood provision was reported in 3% of the studies and linked to fuel wood and timber use. Studies of trees that provided wood were mainly carried out in alleys (Aihou et al. 1998), improved fallows (Girma et al. 2006; Jama et al. 2008) and rotational woodlots (Kimaro et al. 2007, Nyadzi et al. 2003, 2008). Non-timber tree products, such as gum (gum arabic) and genetic resources were reported in few studies, 2% in total.

Regulating ES documented were directly linked to agricultural production and productivity. Water regulation was reported in 10% of the studies, highlighting the regulatory role of trees on water resources. The main indicators of water regulation documented were the effects of trees on soil moisture, infiltration rate and evapotranspiration (Table A1). Trees influenced pest control in 6% of the studies, with weed control being most commonly studied and evenly distributed across agro-ecological zones, with the exception of arid zones. The majority of studies in this category focused on Striga weeds (a parasitic plant genus mostly attacking cereals such as maize, millet and sorghum), e.g. Gacheru and Rao (2005). Carbon storage was documented in 4% of the studies. Most of the studies highlighted carbon stored in the aboveground biomass; only few studies documented the sequestration potential of the trees belowground (e.g. in the roots and soil) or the carbon fluxes. Trees were reported to modify microclimate in 3% of the studies, through shading crops at field and farm scale, and as windbreaks or shelterbelts at landscape scale (Michels et al. 1998). Only nine studies (1%) reported the role of trees in regulation of soil erosion by slowing water flows and wind speed, holding the soil together with their roots and increasing infiltration.

Nutrient cycling (22%) and soil fertility improvement (14%) were the main supporting ESs documented, often evaluated alongside provisioning ESs. The trees mediated nutrient cycling by taking up inorganic nutrients (N, P, K, Mg, Ca, S and micronutrients) and cycling these to crops. Nutrient cycling was also mediated by application of un-decomposed fresh or wilted green material (green manure) from trees. Studies also explored the role of trees in rebuilding soil fertility by restoring soil organic matter (SOM), enhancing mineral nutrient concentrations to support crop productivity and restoring degraded lands (Barrios et al. 1996a, 1998).

ESs that increase and decrease with trees

Overall, beneficial effects of trees were evidenced in 58% of the studies reviewed while 15% reported a decline in ES and 28% found no effect of trees (Figure 4). The results show that trees, while supplying one or more ES, in some cases negatively affected others, thus creating trade-offs (Table 2). In the following, we discuss the impacts of trees on six examples of ESs that were reported to increase or decrease because of trees and were represented in many studies: crop yield, water regulation, soil fertility, microclimate, nutrient cycling and regulation of pests. Whereas a higher proportion of positive effects of trees on crop yield, water regulation, nutrient cycling and soil fertility were reported in semi-arid zones, studies reporting positive influence of trees on pest regulation were mostly from humid zones.

Table 2. The effect of trees on provision of selected ESs in Sub-Saharan African agricultural landscapes.

services	Effect	No. of studies	Agro-ecological zone				Agroforestry practice
			Arid	Semi-arid	Sub-humid	Humid
Food	Increase	141	1	80	31	29	a, d, g, h, k, l, n, o, p
	Neutral	29		16	6	7	a, c, d, g, k, n
	Decrease	37	3	18	7	9	a, b, c, d, g, h, i, o
Microclimate	Increase	11		8	2	1	d, g, h, o
	Neutral
	Decrease	7		5	2		a, d
Water regulation	Increase	35		27	6	2	a, c, d, g
	Neutral	10		6	2	2	d
	Decrease	24	3	17	3	1	a, b, d, g, l
Pest control	Increase	31		9	6	16	a, d, g, h, n
	Neutral	14		6	4	4	a, d, g
	Decrease	8		2	1	5	a, d, g
Nutrient cycling	Increase	76		41	17	19	a, d, g, h, l, n, o
	Neutral	42	1	17	9	15	a, d, g, h, l, n, o
	Decrease	10		5	1	4	d, g, h
Soil fertility	Increase	83	3	41	24	15	a, d, g, h, l, k, m, n, o
	Neutral	59	2	28	18	11	a, d, g, h, l, k, m, n, o
	Decrease	14		5	8	1	a, d, g, h, m, n

The effects are denoted as increase where there was significant positive effect, neutral where there was no significant effect and decrease for significant negative effect. The effects are summarized across different agro-ecological (climatic) zones as counts of studies. Agroforestry technologies associated with the effect are logged by letters and catalogued in Table A3.

Figure 4. Proportions (%) of ESs that increase and decrease by trees in Sub-Saharan Africa.

Crop yield

Crop yields increased under tree-based systems compared to tree-less systems in 68% of the studies, decreased in 18% and were unaffected in 14% of the studies (Figure 4). Crop yields increased due to improved microclimate, nutrient cycling and soil fertility under a variety of agroforestry technologies including: fallowing, tree–crop intercrop, alley cropping and biomass transfer (Table 2; see Table A2 for a full list of agroforestry practices and other management options including trees that have been reported from the studies). High yields after improved fallows, for example, are mainly attributed to increased soil inorganic nitrogen (Barrios et al. 1997; Chirwa et al. 2004). Nitrogen-fixing shrubs and trees were the species that increased the yields of subsequent crops. Yield increase under inter-cropping systems largely depended on climatic conditions and management of trees (Boffa et al. 2000; Bayala et al. 2002, 2008). Yield increases after application of green manures and prunings of shrubs depended mainly on the mode and time of application (Egbe et al. 1998; George et al. 2001; Ganunga et al. 2005). A combination of green manure with inorganic fertilizer had the largest impact on maize yield (Akinnifesi et al. 2007, 2010).

Trees on farm can in some cases also lead to reduced yields by taking up land that would otherwise be used for crop production or by competing with crops for nutrients, water and light. The former was notable for green manure trees and fallowing, which is often associated with loss of a growing season and this will have especially large impact where there is only one short growing season (Bünemann et al. 2004). Yields can also be reduced if trees compete with the crops for nutrients, water and light; a phenomenon common when the density of trees is high (Everson et al. 2009), and in close proximity to crops (Jama et al. 1995, 2000; Boffa et al. 2000). Certain tree species are considered competitive and provide too much shading when grown together with crops, and this may reduce yields (Mathuva et al. 1998; Bayala et al. 2002; Gaafar et al. 2006). Such species, e.g. eucalyptus, have further been criticized because of their tendency to deplete water from surrounding crop lands (Kidanu et al. 2004; Gindaba et al. 2007). Even for trees that have been shown to be less competitive, crop yields could decrease during dry seasons or with older (15 years) trees (Fadl 2009; Fadl & Sheikh 2010) when competition for water resources increases (Chikowo et al. 2004b).

Microclimate

Microclimatic conditions (air and soil temperature, relative humidity, solar radiation, wind speed, etc.) were favourably altered in 61% and negatively altered in 39% of the studies; there was no report of this ES being unaffected (Figure 4). Favourable alteration of microclimate resulted in increased growth and yield of crops due to reduced temperature and evapotranspiration, especially in semi-arid zones (Bayala et al. 2002). The majority of studies reported effects of trees on cereals, especially maize (a staple crop in several countries in SSA; Sileshi et al. 2009). Beneficial effects of moderate shading on growth and yields were also reported on a number of other crops, such as cocoa (Isaac et al. 2007), chilli pepper, millet (Kho et al. 2001) and coffee.

In studies where microclimatic conditions were reporting negative effects in tree–crop systems, it was mainly due to reduced photosynthetically active radiation in semi-arid zones (McIntyre et al. 1997; Ong et al. 2000; Bayala et al. 2002; Everson et al. 2009), higher relative humidity and lower air temperature compared to monoculture in sub-humid zones (Koech & Whitbread 2006; Schroth, Balle, et al. 1995). This resulted in reduced crop yields (McIntyre et al. 1997; Everson et al. 2009) and increased pest infestation (Koech & Whitbread 2006; Schroth, Balle, et al. 1995).

Water regulation

Trees regulate the amount of water available in the soil by controlling soil–plant–atmosphere water relations. Fifty one percent of the studies on water regulation reported a positive effect of trees, 35% found a negative effect and 14% showed no effect (Figure 4). Soil water content was higher on farms with rather than without trees. This was attributed to increased infiltration rate (Chirwa et al. 2003, 2004; Makumba et al. 2006; Sanou et al. 2010; Nyamadzawo et al. 2012), reduced soil evaporation (Jackson and Wallace 1999) and reduced transpiration (Mungai et al. 2000) in semi-arid zones. Another positive effect of trees on water dynamics in semi-arid zones was hydraulic redistribution by some tree and shrub species (Burgess et al. 1998; Bayala et al. 2008; Kizito et al. 2012). However, in some cases trees in simultaneous agroforestry practices showed competition with crops for water resources, resulting in lower amounts of soil water close to trees than in continuous cropping or natural grass fallows, or at greater distances from trees (Hartemink et al. 1996; Odhiambo et al. 2001; Livesley et al. 2004). These studies attributed reduced soil water content to uptake of water by trees, increased transpiration from trees and possibly reduced input of rainfall through canopy interception. Continuous cropping can also lower infiltration rates due to soil compaction as a result of decreased SOM and degradation of soil structure. Depletion patterns in soil moisture content varied with depth in the soil and distance from the trees (Gathumbi et al. 2002; Livesley et al. 2004), and was influenced by factors, such as plant and soil characteristics and climatic conditions (Kidanu et al. 2004; Gaafar et al. 2006). Depletion of soil moisture increased with tree density but was reduced by pruning (Bayala et al. 2002; Gaafar et al. 2006). The impact of trees on soil water content was significant during the dry season when refill from rainfall did not compensate depletion (Nyadzi et al. 2003).

Regulation of pests

Trees were found to most often reduce weeds, insect pests and disease causing plant pathogens. Sixty-eight percent of the studies on pest control reported a positive influence of trees, 15% found a negative effect and 26% found no effect (Figure 4). This result is in agreement with a global meta-analysis of the effect of agroforestry practices on pest control (Pumariño et al. 2015). The biomass of Striga and other arable weeds was reduced in tree-based systems compared with monoculture (Barrios et al. 1998; Sileshi & Mafongoya 2003; Gacheru & Rao 2005; Gworgwor 2007). The primary mechanisms by which trees reduced Striga infestation include a ‘trap crop effect’ that triggers suicidal germination of Striga, improvement of soil inorganic nitrogen and suppression of aboveground weed growth (Gacheru & Rao 2005). Trees also reduced the abundance, spread and damage caused by insect pests by modifying the microclimate, acting as barriers to movement of insects and harbouring natural enemies. For example, shade canopy reduced mirid (bug) density and the overall damage caused by insect pests in cocoa (Babin et al. 2010; Bisseleua et al. 2013). The role of trees in crop disease control has been less studied, but a number of studies in Kenya reported that disease control depended on microclimate and whether the tree species is a host plant or not (Desaeger & Rao 1999, 2000; Kandji et al. 2001, 2002; Schroth, Balle, et al. 1995).

Trees increased the infestations of insect pests and plant pathogens in some studies. For example improved fallows with tree species that are hosts of snout beetles increased snout beetle infestation of maize as compared to natural grass fallows or maize grown continuously in monoculture (Sileshi et al. 2000; Sileshi & Mafongoya 2003). In addition, higher levels of shade increased infestations by the white stem borer on coffee in eastern Uganda, probably due to a modified microclimate (Jonsson et al. 2014).

Nutrient cycling

Trees contribute to the flow of macro- and micro-nutrients in agricultural landscapes. Sixty percent of studies on nutrient cycling found a positive effect of trees, 8% found a negative effect while 33% found no effect (Figure 4). Certain tree species increased nutrient cycling by retrieving sub-soil inorganic N and P from deep soils and cycling these to crops through decomposing biomass (Aweto & Iyanda 2003; Gindaba et al. 2005). Biological nitrogen fixation represented the most effective type of nutrient cycling commonly used to accumulate nitrogen in the soil. Crops grown in the presence of N-fixing tree legumes had significantly greater foliar N concentration than those grown in the absence of trees (Barrios et al. 1997; Chirwa et al. 2004; Akinnifesi et al. 2006; Chikowo et al. 2007). A total of 17 tree species evaluated in the studies are known to fix N from the atmosphere, and their litter and prunings contributed to N cycling either directly or via manure if used as fodder for livestock. The negative effects of trees on nutrient cycling in crops included sub-optimal K uptake (Isaac et al. 2007), low concentrations of leaf N (Hartemink et al. 1996, 2000; Sileshi et al. 2008) and low concentrations of foliar Zn and Mn (Lehmann et al. 1999) aggravated by competition from the trees.

Soil fertility

Fifty-three percent of the studies on soil fertility reported an increase in a range of soil fertility indicators, 9% reported a reduction while 38% showed no effects of trees on some aspects of soil fertility (Figure 4). Trees improved soil fertility in cultivated land by increasing nutrient inputs from organic matter via litterfall and root decay (Barrios et al. 1997; Aihou et al. 1998; Ganunga et al. 2005). Other trees reduced loss of nutrients by reducing soil erosion in Kenya (Angima et al. 2000; Kinama et al. 2007) and improving soil physical properties in Zimbabwe (Mlambo et al. 2005). Specifically, legumes in alleys and fallows frequently increased SOM, SOC, inorganic-N, -K, plant-available P, exchangeable bases (Ca, K, Mg) and maintained higher soil pH than natural grass fallows or continuous maize cultivation (Barrios et al. 1996a, 1996b; Ikerra et al. 1999; Kang et al. 1999). Exceptions were reported in prolonged fallow periods with Gliricidia where the benefit was diminished (Kaya & Nair 2001). In parklands, the concentrations of SOC, SOM and nutrients, such as N, P and K were greater beneath tree canopies than on open ground between trees (Boffa et al. 2000; Kho et al. 2001). In these studies, the nutrient concentrations decreased with distance from tree stems and with soil depth, a pattern caused by nutrients accumulation resulting from greater organic inputs under tree canopies.

Trees sometimes reduced certain soil nutrients in relation to continuous cropping. Low concentrations of exchangeable bases (Ca, K and Mg) were reported particularly in alley cropping with leguminous trees (Kang et al. 1999; Schroth, Oliver, et al. 1995). Other parameters recorded in low quantities in agroforestry treatments as compared to cereal mono-cropping were total soil C and N, and plant available P (Olsen P) (Mathuva et al. 1998; Kang et al. 1999; Makumba et al. 2009). These cases of low concentration of nutrients in the soil were presumably because of nutrient uptake by trees (Isaac et al. 2007).

Biodiversity and ESs

The review shows that trees influence species diversity at field, farm and landscape scale, and that the effects of trees differ between taxa, arrangement of the trees in the landscape and land-use intensity. Biodiversity increased in 56% of the studies and decreased in 16%, while 28% found no effect (Table 3). The studies found that organisms were more abundant and species’ richness was greater in improved fallows and multi-strata agroforests than in monocultures (Table A2). In particular, the abundance, richness and overall diversity of soil organisms were positively influenced by presence of trees in the semi-arid zone (Adejuyigbe et al. 1999; Sileshi & Mafongoya 2003; Sileshi et al. 2008).

Table 3. Reported effects of trees on biodiversity in Sub-Saharan African agricultural landscapes.

Effect	No. of studies	Indicator of biodiversity				Scale			Agro-ecological zone			Agroforestry practice^a	Reference^b
		Plant diversity	Animal diversity	Insect diversity	Soil microorganism	Farm	Field	Land scape	Semi-arid	Sub-humid	Humid
Increase	14	3		8	9	2	11	1	8	1	5	a, d, f, g	1, 4, 5, 6, 7, 9
No effect	7	4	1		2	1	2	4		2	5	a, d, g, h	3, 8
Decrease	4			3	1	2	2		1	3		a, g, h	2, 9

The indicators from which increase and decrease of biodiversity was deduced were aggregated into plant diversity, insects and diversity of soil organisms. The agroforestry practice (a–h, see table footnote) in which the effect was recorded are listed in respective columns and described in detail in Table A3. References to the studies are numbered and listed in the footnote of table.

^a a – alley cropping, d –dispersed intercropping, f – home gardens, g – improved tree fallow, h –multistrata agroforestry.

^b 1. (Adejuyigbe et al. 1999); 2. (Bisseleua et al. 2013); 3. (Kang et al. 1999); 4. (Sileshi & Mafongoya 2006a); 5. (Sileshi et al. 2008a); 6. (Sileshi & Mafongoya 2006b); 7. (Oke & Odebiyi 2007); 8. (Kandji et al. 2002); 9. (Kandji et al. 2001).

Biodiversity was documented alongside an array of ESs provided by trees, such as food production, genetic resources, water regulation, microclimate, carbon storage, pest control, nutrient cycling and soil fertility. The general trend found was that increased belowground biodiversity positively co-varied with increased crop yield and improved soil fertility (Sileshi & Mafongoya 2006a; Sileshi et al. 2008). Increased aboveground biodiversity co-varied with increased fruit yield in multi-strata (cocoa) agroforests (Oke & Odebiyi 2007), increased infiltration rate (Sileshi & Mafongoya 2006b) and also increased pest control (Bisseleua et al. 2013). This indicates that biodiversity underpins supply of ES in agricultural landscapes through maintenance of ecological interactions (Barrios et al. 2012).

Discussion

Effects of trees on ESs in farms and agricultural landscapes

The review summarized 15 ESs provided by trees from studies conducted in farms and agricultural landscapes across SSA. Beneficial effects of trees were found in 58% of the studies reviewed while 15% reported a decline and 28% found no effect on ES. Thus, overall this review suggests that conserving and adding trees on farms and in agricultural landscapes of SSA is likely to increase the provision of multiple ESs. This is in line with other reviews that have shown that pest control is mostly improved under agroforestry practices (Pumariño et al. 2015), that legume tree species improve soil fertility (Jama et al. 2000; Akinnifesi et al. 2010) and increase maize yield (Jama et al. 2000; Sileshi et al. 2009), and that planting trees in agricultural fields increases infiltration capacity of the soils (Ilstedt et al. 2007). It is also in line with conclusions that trees can be a source of disservices depending on how they are managed; appropriate management practices are essential to enjoying the services and reducing the disservices from trees (Tscharntke et al. 2011).

This review identified a wide range of benefits of trees in agriculture. Most striking is the finding that crop yields increased under tree-based systems compared to tree-less systems in the majority of the reviewed studies (68%) and were unaffected in another 14%. Studies reporting the effect of trees on increased crop yields often also reported the role of trees on nutrient cycling and soil fertility (Makumba et al. 2006; Kimaro et al. 2008). Nutrient cycling and maintenance of soil fertility is essential in SSA where a large portion of the agricultural land is considered to be severely degraded (Boko et al. 2007). The studies also provided evidence of other benefits of trees, such as control of erosion, water regulation, modification of microclimate conditions, regulation of pests, as well as provision wood and non-wood tree products. The majority of the trees and shrubs providing these benefits were hedge-grown fodder species east Africa (Angima et al. 2002; Kinama et al. 2007; Mutegi et al. 2008), those dispersed in crop fields to provide shade (Bayala et al. 2002; Bisseleua et al. 2013), or in shelterbelts to modify microclimate and to check wind erosion (Michels et al. 1998; Onyewotu et al. 2004) in west Africa, and those planted in woodlots for wood (Kimaro et al. 2007; Nyadzi et al. 2003) in southern Africa.

The positive effects of trees on ES in agricultural landscapes were consistently largest in semi-arid areas. This includes enhanced nutrient cycling, improved soil fertility and improved microclimatic conditions, control of soil erosion and pest regulation under alley cropping (Mungai et al. 2000; Nyamadzawo et al. 2012), intercrops (Akinnifesi et al. 2007; Makumba et al. 2009) and improved fallows (Chikowo et al. 2004a; Sileshi et al. 2006). Biodiversity (abundance, richness and overall diversity of soil organisms) was positively influenced by presence of trees in the semi-arid zone (Adejuyigbe et al. 1999; Sileshi & Mafongoya 2006a; Sileshi et al. 2008). However, the inclusion of trees on farms in semi-arid zones requires adaptive management, since these areas not only have high-response potential but also risks of trade-offs.

Trade-offs of trees on farms and agricultural landscapes and how they can be reduced

A number of studies show that while trees affected some ES positively, they negatively affected others. Whether the trees can be integrated in the farming systems to promote only benefits or if they also introduce trade-offs depend on tree management as affected by the multi-functionality of tree species, their resource use efficiency and ability to favourably modify the microclimate for crops. Trade-offs associated with microclimate, water resources and nutrient depletion were found to be especially prevalent in alley cropping and for trees dispersed in cropped plots. These trees were introduced to reduce soil erosion, improve soil fertility and provide shade. However, abounding below- and above-ground competition for water, nutrients and light outweighed their benefit to crops in some environments.

The most common trade-offs of trees on farms were linked to microclimate and water regulation. Water depletion is a major problem in the semi-arid areas and particularly in dry seasons (Ong et al. 2000), while shading is a notable factor in reducing yield in parklands (Bayala et al. 2002). In dry regions or seasons, trees compete with crops for water resources through uptake and transpiration. This process balances with the positive effect of regulation of water by better infiltration due to trees and reduced soil evaporation (Ilstedt et al. 2007). The balance between these effects will be most pertinent in semi-arid environments where water-supply to crops is often the most limiting factor to crop growth. For example, while increase soil organic carbon resulted from the presence of trees in semi-arid Sudan, this was accompanied by a decline in soil moisture content and crop yield (Fadl 2009; Fadl & Sheikh 2010). Trees planted for green manure, wood or gum also lead to a decline in soil moisture content and crop yield in arid (Gaafar et al. 2006), semiarid (Mathuva et al. 1998) and sub-humid (Kidanu et al. 2004) areas. Conversely, improved soil fertility led to increased yields in the humid region despite a decline in soil moisture (Ikerra et al. 1999, 2006; Gathumbi et al. 2002). The magnitude and direction of the trade-offs between ES were influenced by the arrangement of trees in the system, their subsequent management, the tree species and the underlying climatic and edaphic factors.

Competition between trees and crops can be minimized by different management practices and selection of trees that are less competitive. Management options include increasing the spacing between trees and pruning the tree canopy to improve light penetration and to reduce the demand for water by trees or pruning of the root system to reduce water uptake. Several studies demonstrate the possibility to control water uptake and transpiration through pruning and regulating phenology (Livesley et al. 2004; Muthuri et al. 2004; Namirembe et al. 2009). A further benefit of tree pruning is that, if pruned material is left on the field, it adds organic matter and increase nutrient return to the soil (Mathuva et al. 1998; Livesley et al. 2002). Another measure to minimize competition between trees and crops is matching the agroforestry practice to the site. For example, crops on soils with low nitrogen content benefited from nitrogen-fixing trees, while adding legumes to soils where nitrogen was not limiting led to increases in nitrogen-based emissions while not increasing yields (Chikowo et al. 2004a, 2007). Exploitation of the complementarity between trees and crops has also been proposed on deep soils, particularly for deep-rooted tree species, which are able to utilize water and nutrients from below the crop-rooting zone (Gaafar et al. 2006). Deep-rooted species have reduced root volume near the surface, a feature that is good for minimizing competition with adjacent crops. Sequential arrangements, on the other hand, would be beneficial in shallow soils where trees are likely to compete with crops. Risks of increased pest problems can be minimized by ensuring that the trees are not important hosts or provide recourses for crop pests.

Geographical distribution of studies and knowledge gaps

The majority of ESs were reported in semi-arid zones. The relative occurrence of studies across the agro-ecological zones compares well with their proportional land area, except for the arid zones and inland humid areas where studies were rare. The lack of studies in the arid zone can be attributed to minimal agricultural production in most arid areas while humid inland areas of Central Africa are dominated by forests. The studies seem to aggregate around areas, which have experienced pressures on natural resources resulting in, for example, land degradation and where crop and livestock production are integrated. Pressures on natural resources can, for example, be attributed to increasing population density in these areas.

The role of trees in provision of wood, fruits and nuts and non-timber tree products, such as gum, genetic resources and medicine was underreported. The low number of studies on these ESs on farms and in agricultural landscapes indicates that the trends found in this review may not fully capture farmers’ practices. For example, wood is only cited in studies on trees planted in woodlot or alley hedgerows. In contrast, there are no studies on provision of wood from dispersed trees on cropland or pastures, boundary plantings, multi-story tree gardens or other agroforestry practices common for provision of timber and non-timber products in SSA. On-farm timber and non-timber tree products have increased recently following the decline in forest cover and increasing demand for these products (De Foresta et al. 2013). The small number of studies on non-timber tree products, thus, does not match to the widely acknowledged potentials, e.g. of genetic diversity to improve crop productivity, production of fruits and nuts and buffer against pests and diseases. However, it conforms to earlier reports that such ESs are not well documented, especially in Africa (Crossman et al. 2013; De Beenhouwer et al. 2013).

The review identified obvious lack of studies on cultural ESs provided by trees on farms in agricultural landscapes in SSA. Cultural services provided by trees seem to be understudied on farmland despite the fact that many cultural ESs from trees are documented to be produced in the wild (Egoh et al. 2012). People in agricultural landscapes obtain spiritual inspiration, cultural identity, recreation, etc., through interaction with trees. Certain trees, for example, are used as venues for learning and support experiences that uphold certain moral, religious and aesthetic values in many communities. Large trees are often used to provide shade and shelter for small businesses or for meetings and training events. These ESs are underreported, with virtually no quantitative data describing how trees in agricultural landscapes realize cultural needs. Recent reviews have attributed lack of information on cultural ESs to problems of definition, differences in aesthetic preferences and the slow integration of cultural services of trees within the ES framework (MA 2005; Daniel et al. 2012; Satz et al. 2013).

ESs are supplied at a range of spatial scales from field-level to landscape-level. However, ESs in the studies reviewed were mainly documented at field scale (68%) and less at farm (23%) and landscape scale (9%). This highlights a potential disparity in terms of the scale at which research was conducted in relation to the scale where a particular ES is actually produced and utilized. Mismatches between scales were noted in the provisioning (food, fodder and wood), which are generally utilized at farm-level but were mostly studied at field scale. This is in line with the literature indicating that studies on ESs are often conducted at a smaller scale than where they are produced and utilized (Balvanera et al. 2006). Another aspect of mismatch in scale is where the ES is produced at one scale, e.g. erosion control provided as a result of farmers’ work at field and farm scale, and benefits manifested at landscape level as less sedimentation in the river is experienced by communities or hydropower companies downstream in the watershed. It is in these situations that payment for environmental services can become a useful mechanism to link those producing ESs with the people utilizing them.

Conclusion

Trees on farms and in agricultural landscapes in SSA provide ESs that support food security, agricultural productivity and sustainable natural resource management, particularly in semi-arid environments. Research on supporting and regulating ES is often aimed at improving soil fertility and adjusting growing conditions in order to increase provisioning ES for improved crop production. While food production is important for food security and income, mainly focusing on provisioning ES risks failing to appreciate that ESs are interdependent. Longer-term trends and benefits can be missed when focusing on short-term gains.

Trade-offs were mainly observed between trees and provision of certain food crops, in particular during dry periods and in semi-arid zones, brought about by depletion of water resources, exhaustion of nutrients and unfavourable alteration of microclimatic conditions. It is not known whether the trade-offs observed are caused by interactions among ESs or simultaneous responses to the same driver(s) of ecosystem change. However, the trade-offs between ESs provided by trees and other farm enterprises (e.g. crops) can be minimized by management options that reduce competition and increase complementarity among crops and trees.

Very few studies on cultural ESs and a general underreporting of certain provisioning ES, such as wood, genetic resources and medicinal products in agricultural contexts indicate knowledge gaps. Data on provisioning ES were strongly biased towards staple food production, while information on provisioning of fruits and nuts is largely lacking. The lack of data on certain ESs provided by trees in agro-ecosystems makes it difficult to establish all the synergies and trade-offs associated with trees in the landscape. In addition, most studies on ESs are carried out at field scale although the potential benefits of the ESs are at farm or landscape scales. The poor representation of certain ESs in scientific studies and that effects of trees on ESs often are studied at a lower scale than where they are benefitted could furthermore impede policy development.

Acknowledgements

This study was supported by funds allocated to the Swedish University of Agricultural Sciences by the Swedish Ministry of Foreign Affairs as part of its special effort on global food security. Funding to EB to contribute to this chapter was provided by the CGIAR research programs on Forests, Trees and Agroforestry, and Humid Tropics.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by the Swedish University of Agricultural Sciences by the Swedish Ministry of Foreign Affairs as part of its special effort on global food security. Edmundo Barrios was funded by Consultative Group for International Agricultural Research (CGIAR) research programs on Forests, Trees and Agroforestry, and Humid Tropics.

Appendix

Table A1. Literature references reviewed, and on which the database is built. A total of 350 scientific papers that report ecosystem services provided by trees on farms and in agricultural landscapes of Sub-Saharan Africa. The papers refer to 205 sites dispersed in 23 countries across SSA. The database will be uploaded in Dataverse (https://dataverse.harvard.edu/dataverse/icraf).

Adejuyigbe et al. 1999. Agrofor. Syst. 47: 263–272; Adekunle (2009). Australasian Plant Pathology 38: 211–215; Aihou et al. 1999. Agrofor. Syst. 42: 213–227; Akinnifesi et al. (1997). Nutr. Cycl. Agroecosyst. 47: 71–80; Akinnifesi et al. 2006. Exp. Agric. 42: 441–457; Akinnifesi et al. 2007. Plant Soil 294: 203–217; Akobundu et al. (2002). Agrofor. Syst. 54: 161–170; Allou et al. (2006). J. Chem. Ecol. 32: 1743–1754; Amara et al. (1996). Agrofor. Syst. 34: 119–128; Angima et al. 2000. Journal of Soil and Water Conservation 55: 478–482; Angima et al. 2002. Agric. Ecosyst. Environ. 91: 15–23; Anim-Kwapong and Osei-Bonsu (2009). Agrofor. Syst. 76: 533–542; Arigbede et al. (2012). J. Anim. Physiol. Anim. Nutr. 96: 250–259; Asase et al. (2009). Afr. J. Ecol. 48: 338–346; Asfaw and Lemenih (2010). Forests: trees and Livelihoods 19: 359–377; Asfaw and Ågren (2007). Agrofor. Syst. 71: 35–48; Assefa (1998). Agrofor. Syst. 42: 13–23; Assogbadjo et al. (2012). For. Policy Econ. 14: 41–49; Augusseau et al. (2006). Biodivers. Conserv. 15: 613–630; Avohou et al. (2011). Biomass Bioenergy 35: 562–571; Aweto and Iyanda 2003 Land Degrad. Dev. 14: 51–56; Babin et al. 2010. Agric. For. Entomol. 12: 69–79; Banful et al. (2000). Agrofor. Syst. 49: 189–199; Barrios (1996). Soil Biol. Biochem. 28: 1459–1465; Barrios et al. (1996). Soil Biol. Biochem. 28: 185–193; Barrios et al. 1997. Soil Sci. Soc. Am. J. 61: 826–831; Barrios et al. 1998. Soil Sci. Soc. Am. J. 62: 1604–1609; Bayala et al. (2006). Nutr. Cycl. Agroecosyst. 76: 193–201; Bayala et al. 2008. Acta Oecologica 34: 370–378; Bayala et al. (2003). Arid Land Res. Manag. 17: 283–296; Bayala et al. (2005). Soil Biol. Biochem. 37: 533–539; Bayala et al. (2009). Agrofor. Syst. 79: 117–127; Bayala et al. 2008. Agrofor. Syst. 72: 187–194; Bayala et al. 2002. Agrofor. Syst. 54: 203–214; Bazié et al. (2012). Agrofor. Syst. 84: 377–388; Beedy et al. (2010) Agric. Ecosyst. Environ. 138: 139–146; Bernatchez et al. (2008). J. Arid Environ. 72: 1247–1256; Bisseleua et al. 2013. PLOS ONE 8: e56115; Bisseleua et al. (2009). Conservation Biology 23: 1176–1184; Bobo et al. (2006). J. Insect Conserv. 10: 29–42; Bobo et al. (2006). Biodivers. Conserv. 15: 4097–4117; Boffa et al. (2008). Afr. J. Ecol. 46: 24–32; Boffa et al. 2000. Agrofor. Syst. 49: 153–175; Bünemann et al. 2004. Plant Soil 264: 195–208; Burgess et al. 1998 Oecologia 115: 306–311; Chamshama et al. (1998). Agrofor. Syst. 40: 215–225; Chikowo et al. 2006. Nutr. Cycl. Agroecosyst. 76: 219–231; Chikowo et al. (2004). Agric. Ecosyst. Environ. 102: 119–131; Chikowo et al. (2004). Plant Soil 259: 315–330; Chikowo et al. (2004). Plant Soil 262: 303–315; Chirwa et al. 2003. Agrofor. Syst. 59: 265–277; Chirwa et al. (2006). Agrofor. Syst. 67: 93–106; Chirwa et al. (2007). Agrofor. Syst. 69: 29–43; Chirwa et al. 2003. Agrofor. Syst. 59: 243–251; Chirwa et al. 2004. Biol Fertil Soils 40: 20–27; Correia et al. (2010). Biodivers. Conserv. 19: 1725–1747; Deans et al. (1999). For. Ecol. Manag. 124: 153–167; Desaeger and Rao 1999. Agrofor. Syst. 47: 273–288; Desaeger and Rao 2000. Field Crops Research 65: 41–56; Diack et al. 2000. Arid. Soil Res. Rehab. 14: 205–218; Diagne et al. 2001. For. Ecol. Manag. 146: 35–43; Diels et al. 2004. Soil Biol. Biochem. 36: 1739–1750; Dossa et al. 2007. Agrofor. Syst. 72: 103–115; Duguma and Mollet 1997. Agrofor. Syst. 37: 45–57; Duguma and Mollet 1998. Agrofor. Syst. 40: 283–296; Egbe et al. 1998. Agrofor. Syst. 40: 309–321; Emanuel et al. 2005. For. Ecol. Manag. 214: 91–103; Everson et al. 2009. Agrofor. Syst. 75: 211–221; Fadl 2009. Forests: trees and Livelihoods 19: 27–46; Fadl 2013. J. For. Res. 24: 173–177; Fadl and Gebauer 2004. J. Agr. Rural Dev. Trop. 105: 149–154; Fadl and Sheikh 2010. Agrofor. Syst. 78: 243–252; Fifanou et al. 2011. Agrofor. Syst. 82: 1–13; Gaafar et al. 2006. Agrofor. Syst. 66: 1–11; Gachengo et al. 1999. Agrofor. Syst. 44: 21–36; Gacheru and Rao 2001. Int. J. Pest Manag. 47: 233–239; Gacheru and Rao 2005. Int. J. Pest Manag. 51: 91–100; Galiana et al. 2002. Soil Biol. Biochem. 34: 251–262; Ganunga et al. 2005. S Afr J Plant & Soil 22: 240–245; Gathumbi et al. 2003. Soil Sci. Soc. Am. J. 67: 573–582; Gathumbi et al. 2002. Soil Biol. Biochem. 34: 1059–1071; Gathumbi et al. 2004. For. Ecol. Manag. 187: 267–280; Gathumbi et al. 2002. Agron. J. 94: 518–526; George et al. 2002. Plant Soil 246: 53–63; George et al. 2001. Agrofor. Syst. 52: 199–205; Giday et al. 2013. J. Arid Environ. 94: 80–96; Gindaba et al. 2004. Biol. Fertil. Soils 40: 393–397; Gindaba et al. 2005. Biol. Fertil. Soils 42: 66–71; Gindaba et al. 2007. S Afr J Plant & Soil 24: 47–50; Girma et al. 2006. Agrofor. Syst. 68: 93–102; Girma et al. 2000. Agrofor. Syst. 50: 279–292; Githae et al. 2011. Afr. J. Plant Sci. 5: 475–482; Gnankambary 2008. Nutr. Cycl. Agroecosyst. 82: 1–13; Gockowski et al. 2010. J. Sustainable For. 29: 638–670; Gueye et al. 1997. Arid. Soil Res. Rehab. 11: 63–70; Gworgwor 2007. Agron. Sustainable Dev. 27: 89–94; Hagos et al. 2005. J. Arid Environ. 61: 47–59; Hailemariam et al. 2013. Agrofor. Syst. 87: 1261–1272; Hailu et al. 2000. Agrofor. Syst. 48: 9–24; Harawa et al. 2006. Nutr. Cycl. Agroecosyst. 75: 271–284; Hartemink et al. 1996. Soil Sci. Soc. Am. J. 60: 568–574; Hartemink et al. 2000. Geoderma 98: 11–33; Haule et al. 2003. Malawi Journal of Agricultural Sciences 2: 21–33; Hauser et al. 2000. Agrofor. Syst. 49: 177–188; Hauser et al. 2000. Agrofor. Syst. 49: 111–122; Hauser et al. 2005. Agrofor. Syst. 64: 73–82; Heineman et al. 1997. For. Ecol. Manag. 91: 103–135; Henry et al. 2009. Agric. Ecosyst. Environ. 129: 238–252; Hove et al. 2003. Agrofor. Syst. 59: 231–241; Hove et al. 2001. Anim. Feed Sci. Technol. 91: 95–106; Howard et al. 1997. Agrofor. Syst. 35: 15–29; Ikerra et al. 1999. Plant Soil 211: 155–164; Ikerra et al. 2001. Agrofor. Syst. 52: 13–21; Ikerra et al. 2006. Nutr. Cycl. Agroecosyst. 76: 249–260; Isaac et al. 2005. Agrofor. Syst. 65: 23–31; Isaac et al. 2011. For. Ecol. Manag. 261: 582–588; Isaac et al. 2007. Nutr. Cycl. Agroecosyst. 78: 155–165; Isaac et al. 2007. Plant Soil 298: 243–254; Iyamuremye et al. 2000. Arid. Soil Res. Rehab. 14: 359–371; Jackson 2000. Agric. For. Meteorol. 100: 323–336; Jackson et al. 1996. Biol Fertil Soils 23: 50–56; Jackson and Wallace 1999. Agric. For. Meteorol. 94: 203–215; Jackson et al. 2000. For. Ecol. Manag. 126: 133–148; Jaiyeoba 1996. J. Arid Environ. 33: 473–482; Jama et al. 1995. Agrofor. Syst. 31: 257–274; Jama and Nair 1996. Plant Soil 79: 275–285; Jama et al. 2008. Agrofor. Syst. 73: 155–166; Jones et al. 1998. Plant Soil 201: 197–207; Jonsson et al. 1999. Exp. Agric. 35: 39–53; Kabi and Bareeba 2008. Anim. Feed Sci. Technol. 140: 178–190; Kakengi et al. 2001. Agrofor. Syst. 52: 73–82; Kakengi et al. 2005. Asian-Aust. J. Anim. Sci. 18: 42–47; Kandji et al. 2001. Applied Soil Ecology 18: 143–157; Kandji et al. 2002. Eur. J. Soil Biol. 38: 197–203; Kandji et al. 2003. Agrofor. Syst. 57: 51–57; Kang et al. 1999. Nutr. Cycl. Agroecosyst. 54: 145–155; Kang et al. 1997. Soil Use Manage. 13: 130–136; Kanmegne et al. 2003. Agrofor. Syst. 58: 119–124; Kanmegne et al. 1999. Agrofor. Syst. 46: 239–249; Kaonga and Bayliss-Smith 2009. Agrofor. Syst. 76: 37–51; Kaonga and Coleman 2008. For. Ecol. Manag. 256 1160–1166; Karachi and Matata 2000. J. Trop. For. Sci. 12: 138–246; Karachi 1997. Agrofor. Syst. 37: 253–263; Kasongo et al. 2009. Soil Use Manage. 25: 21–27; Kaya and Nair 2001. Agrofor. Syst. 52: 1–11; Kaya and Nair 2004. Agrofor. Syst. 60: 267–276; Kelly et al. 2004. Molecular Ecology 13: 1231–1240; Kho et al. 2001. Agrofor. Syst. 52: 219–238; Kidanu et al. 2004 Agricultural Systems 80: 151–170; Kim 2012. Agrofor. Syst. 86: 175–184; Kimaro et al. 2008 Agric. Ecosyst. Environ. 125: 73–83; Kimaro et al. 2007. Agrofor. Syst. 71: 175–184; Kinama et al. 2005. Agric. For. Meteorol. 130: 149–162; Kinama et al. 2007. Arid Land Res. Manag. 21: 1–19; Kindt et al. 2006. Biodivers. Conserv. 15: 3235–3249; Kindt et al. (2006b). Biodivers. Conserv. 15: 1253–1270; Kiwia et al. 2009. Agrofor. Syst. 76: 455–465; Kizito et al. 2012. J. Arid Environ. 83: 69–77; Koech and Whitbread 2006. New Forests 31: 385–402; Kuyah et al. 2012. Agric. Ecosyst. Environ. 158: 225–234; Kuyah et al. 2013. Biomass Bioenergy 55: 276–284; Kuyah et al. 2012. Agric. Ecosyst. Environ. 158: 216–224; Laird et al. 2007. Biodivers. Conserv. 16: 2401–2427; Lamien et al. 2007. Agrofor. Syst. 69: 1–11; Larbi et al. 2005. Agrofor. Syst. 65: 197–205; Larbi et al. 2000. Agrofor. Syst. 48: 157–168; Larbi et al. 1998. Agrofor. Syst. 39: 275–290; Larbi et al. 1998. Anim. Feed Sci. Technol. 72: 81–96; Lehmann et al. 1998. J. Arid Environ. 38: 479–492; Lehmann et al. 1999. Biol Fertil Soils 30: 90–95; Lehmann et al. 2002. Nutr. Cycl. Agroecosyst. 62: 1–9; Lehmann et al. 1998. For. Ecol. Manag. 111: 157–169; Lehmann et al. 1998. Soil Biol. Biochem. 30: 1545–1552; Lehmann et al. 1995. Agrofor. Syst. 29: 21–36; Lehmann et al. 1999. Plant Soil 210: 249–262; Lesueur and Sarr 2008. Agrofor. Syst. 73: 37–45; Lisau et al. 2005. Agrofor. Syst. 64: 61–0071; Livesley et al. 2002. Plant Soil 247: 177–187; Livesley et al. 2004. Plant Soil 264: 129–139; Lose et al. 2003. Agric. Ecosyst. Environ. 100: 137–151; Lott et al. 2009. Agric. For. Meteorol. 149: 1140–1151; Lupwayi and Haque 1999. Biol Fertil Soils 28: 182–195; Lupwayi and Haque 1999. Biol Fertil Soils 28: 204–211; Lupwayi et al. 1999. Biol Fertil Soils 28: 196–203; Mafongoya and Dzowela 1999. Agrofor. Syst. 47:139–151; Mafongoya and Nair 1997. Agrofor. Syst. 35: 31–46; Mafongoya et al. 1997. Agrofor. Syst. 35: 47–35; Mafongoya et al. 2000. Biol Fertil Soils 30: 298–305; Mafongoya et al. 2003. Agrofor. Syst. 59: 279–288; Makumba et al. 2009. Afr. J. Agric. Res. 4: 278–288; Makumba et al. 2007. Agric. Ecosyst. Environ. 118: 237–243; Makumba and Akinnifesi 2008. Afr. J. Biotechnol 7: 3266–3274; Makumba et al. 2005. Exp. Agric. 42: 51–63; Makumba et al. 2006. Agric. Ecosyst. Environ. 116: 85–92; Mapiye et al. 2009. Livestock Science 126: 206–214; Maroko et al. 1998. Biol Fertil Soils 26: 229–234; Maroko et al. 1999. Soil Sci. Soc. Am. J. 63: 320–326; Masama et al. 1997. Anim. Feed Sci. Technol. 69: 233–240; Masse et al. 2004. Soil Use Manage. 20: 92–95; Matata et al. 2011. J. Soil Sci. Environ. Manage. 2: 311–317; Mathuva et al. 1998. Field Crops Research 55: 57–72; Mclntyre et al. 1997. Field Crops Research 52: 151–160; Mekonnen et al. 1999. Agrofor. Syst. 42: 265–282; Mekonnen et al. 1997. Plant Soil 188: 319–327; Mekonnen et al. 2006. New Forests 32: 285–298; Mekonnen et al. 2007. Forests: trees and Livelihoods 17: 309–323; Michels et al. 1998. Exp. Agric. 34: 449–464; Mohammed et al. 1999. Theor. Appl. Climatol. 62: 221–231; Mouen Bedimo et al. 2007. Eur J Plant Pathol 119: 391–400;

Mouen Bedimo et al. 2008. Phytopathology 98: 1320–1325; Mpairwe et al. 1998. Agrofor. Syst. 41: 139–150; Muchane et al. 2010. Agrofor. Syst. 78: 139–150; Mugendi et al. 2000. Biol Fertil Soils 31: 97–101; Mugendi et al. 1999. Agrofor. Syst. 46: 51–64; Mugendi et al. 1999. Agrofor. Syst. 46: 39–50; Mugwe et al. 2009. Agrofor. Syst. 76: 467–485; Mugwe et al. 2009. Soil Use Manage. 25: 434–440; Mukuralinda et al. 2009. Nutr. Cycl. Agroecosyst. 84: 155–166; Mulatya et al. 2002. Agrofor. Syst. 56: 65–72; Mungai et al. 2001. J. Environ. Sci. 13: 164–169; Mungai et al. 2000. Theor. Appl. Climatol.: 63–78; Mungai et al. 2001. J. Environ. Sci. 13: 291–298; Murovhi et al. 2012. South Africa Agrofor. Syst. 86: 61–71; Murovhi and Materechera 2006. The Southern African Forestry Journal 208: 23–30; Mutegi et al. 2008. Agrofor. Syst. 74: 37–49; Muthuri et al. 2004. Land Use and Water Resources Research 4: 1–11; Muthuri et al. 2005. For. Ecol. Manag. 212: 23–39; Muthuri et al. 2009. Agric. Ecosyst. Environ. 129: 497–507; Mweta et al. 2007. Scientific Research and Essay 2: 446–453; Namirembe et al. 2009. Agrofor. Syst. 75: 197–210; Ndiaye et al. 2000. Arid. Soil Res. Rehab. 14: 317–327; Ndufa et al. 2009. Agron. J. 101: 1352; Ndufa et al. 1999. For. Ecol. Manag. 122: 231–241; Nherera et al. 1998. Anim. Feed Sci. Technol. 74: 15–28; Niang et al. 2002. Agrofor. Syst. 56: 145–154; Niang et al. 1998. Agrofor. Syst. 39: 263–274; Niang et al. 1996. Agrofor. Syst. 34: 171–178; Nolte et al. 2003. For. Ecol. Manag. 179: 535–545; Norgrove and Hauser 2013. Agrofor. Syst. 87: 349–354; Nyadzi et al. 2006. Agric. Ecosyst. Environ. 116: 93–103; Nyadzi et al. 2003. Agrofor. Syst. 59: 215–229; Nyadzi et al. 2003. Agrofor. Syst. 59: 253–263; Nyamadzawo et al. 2007. Soil and Tillage Research 96: 182–194; Nyamadzawo et al. 2008. Nutr. Cycl. Agroecosyst. 81: 85–93; Nyamadzawo et al. 2003. Agrofor. Syst. 59: 207–214; Nyamadzawo et al. 2008. Agrofor. Syst. 72: 161–168; Nyamadzawo et al. 2012. Water SA 38; Nyberg et al. 2000. Plant Soil 218: 83–89; Nyberg and Högberg 1995. Agrofor. Syst. 32: 45–52; Nziguheba et al. 2005. Biol Fertil Soils 41: 240–248; Nziguheba et al. 2002. Agrofor. Syst. 55: 165–174; Odhiambo et al. 2001. Plant Soil 235: 221–233; Ogol et al. 1998. Environ. Entomol. 27: 1444–1451; Ogol et al. 1999. Agric. Ecosyst. Environ. 76: 1–15; Oke and Odebiyi 2007. Agric. Ecosyst. Environ. 122: 305–311; Okogun et al. 2000. Agrofor. Syst. 50: 123–136; O’neill et al. 2002. Journal of Sustainable Agriculture 29: 57–67; Ong et al. 2000. Agric. Ecosyst. Environ. 80: 121–141; Onyewotu et al. 2004. Theor. Appl. Climatol. 79: 255–262; Osman et al. 2011. Afr. J. Agric. Res. 6: 882–891; Oteng’i et al. 2007. Afr. J. Agric. Res. 2: 89–104; Oteng’i et al. 2000. Agrofor. Syst. 50: 137–156; Ouédraogo et al. 2006. Agrofor. Syst. 68: 69–80; Ouédraogo-Koné et al. 2008. Agrofor. Syst. 74: 213–221; Ouinsavi and Sokpon 2008. Agrofor. Syst. 74: 17–26; Pamo et al. 2007. Anim. Feed Sci. Technol. 135: 273–282; Pamo et al. 2006. Small Rumin. Res. 65: 31–37; Partey and Thevathasan 2013. Commun. Soil Sci. Plant Anal.; Paterson et al. 1999. Tropical Animal Health and Production 31: 115–126; Phiri et al. 1999. Agrofor. Syst. 49: 153–162: Phiri et al. 2003. Agrofor. Syst. 59: 197–205; Pouliot et al. 2011. Agrofor. Syst. 85: 477–488; Raddad et al. 2005. Plant Soil 275: 261–269; Raddad and Luukkanen 2007. Agric. Water Manag. 87: 61–72; Raddad et al. 2006. Agrofor. Syst. 68: 193–207; Radersma and Ong 2004. For. Ecol. Manag. 188: 77–89; Rao and Mathuva 2000. Agric. Ecosyst. Environ. 78: 123–137; Rao et al. 2002. Exp. Agric. 38: 223–236; Reda et al. 2005. Journal of Agronomy and Crop Science 191: 20–26; Reyes et al. 2009. Agrofor. Syst. 76: 513–523; Rhoades 1995. Agrofor. Syst. 29: 133–145; Roothaert and Franzel 2001. Agrofor. Syst. 52: 239–252; Roothaert et al. 2003. Exp. Agric. 39: 423–440; Roupsard et al. 1999. Functional Ecology 13: 460–472; Rubanza et al. 2007. Small Rumin. Res. 70: 165–173; Sakala et al. 2003. BIol. Agric. Hortic. 21: 121–130; Salako et al. 1999. Geoderma 91: 103–123; Samba et al. 2001. For. Ecol. Manag. 154: 277–288; Sanginga et al. 1996. Biol Fertil Soils 23: 26–32; Sanginga et al. 1995. Biol Fertil Soils 20: 205–211; Sanon et al. 2008. Livestock Science 117: 192–202; Sanou et al. 2012. Exp. Agric. 48: 283–300; Sanou et al. 2012. Agrofor. Syst. 85: 431–441; Sanou et al. 2010. Soil Use Manage. 26: 75–81; Schroth et al. 1995. Agrofor. Syst. 29: 147–163; Schroth et al. 1995. Agrofor. Syst. 32: 261–276; Sebata et al. 2011. Anim. Feed Sci. Technol. 170: 122–125; Shackleton et al. 2003. Forests: trees and Livelihoods 13: 217–232; Shepherd et al. 2000. Biol Fertil Soils 31: 348–355; Shepherd et al. 1996. Fertilizer Research 44: 87–99; Sileshi et al. 2006. Zambian Journal of Agricultural Science 8: 6–12; Sileshi et al. 2008. Soil Biol. Biochem. 40: 3065–3075; Sileshi and Mafongoya 2003. Applied Soil Ecology 23: 211–222; Sileshi and Mafongoya 2006. Agric. Ecosyst. Environ. 115: 69–78; Sileshi and Mafongoya 2006. Applied Soil Ecology 33: 49–60; Sileshi and Mafongoya 2007. Biol. Fertil. Soils 43: 333–340; Sileshi et al. 2005. Agric. For. Entomol. 7: 61–69; Sileshi et al. 2000. Agrofor. Syst. 49: 41–52; Sileshi et al. 2011. Agric. Water Manag. 98: 1364–1372; Sileshi et al. 2011. Agrofor. Syst. 83: 361–372; Siriri et al. 2010. Agrofor. Syst. 78: 65–77; Siriri and Raussen 2003. Agric. Ecosyst. Environ. 95: 359–360; Siriri et al. 2013. Agrofor. Syst. 87: 45–58; Sirrine et al. 2010. Int. J. Agr. Sustain. 8: 290–304; Sjögren et al. 2010. J. For. Res. 21: 379–386; Smith et al. 1999. Functional Ecology 13: 256–264; Smith et al. 1999. Plant Soil 211: 191–205; Sonwa et al. 2007. Biodivers. Conserv. 16: 2385–2400; Ståhl et al. 2005. Agrofor. Syst. 65: 67–79; Ståhl et al. 2002. Plant Soil 243: 103–117; Tahir et al. 2009. J. Arid Environ. 73: 499–505; Takimoto et al. 2008. Agric. Ecosyst. Environ. 125: 159–166; Takimoto et al. 2009. Agrofor. Syst. 76: 11–25; Teklay and Malmer 2004. Soil Biol. Biochem. 36: 777–786; Teklay et al. 2006. Nutr. Cycl. Agroecosyst. 75: 163–173; Thor Smestad et al. 2002. Agrofor. Syst. 55: 181–194; Tian et al. 2003. J. Plant Nutr. 26: 475–486; Tian et al. 2005. Nutr. Cycl. Agroecosyst. 71: 139–150; Tian and Kolawole 2004. J. Plant Nutr. 27: 571–583; Tilander and Bonzi 1997. Plant Soil 197: 219–232; Tilander et al. 1995. Agrofor. Syst. 30: 363–378; Tomlinson et al. 1998. For. Ecol. Manag. 107: 173–182; Tossah et al. 1999. Agrofor. Syst. 42: 229–244; Traore et al. 2004. Arid Land Res. Manag. 18: 359–368; Tschakert et al. 2004. J. Arid Environ. 59: 511–533; Vanlauwe et al. 2001. Agrofor. Syst. 53: 21–30; Vanlauwe et al. 1999. Agrofor. Syst. 42: 245–264; Vanlauwe et al. 1998. Eur. J. Soil Sci. 49: 417–425; Vanlauwe et al. 2005. Plant Soil 273: 337–354; Vanlauwe et al. 1998. Soil Sci. Soc. Am. J. 62: 454–460; Vanlauwe et al. 1998. Soil Sci. Soc. Am. J. 62: 461–466; Vanlauwe et al. 2001. Plant Soil 231: 187–199; Vanlauwe et al. 2001. Plant Soil 231: 201–210; Vanlauwe et al. 1996. Soil Biol. Biochem. 28: 739–749; Venter et al. 2011. For. Ecol. Manag. 261: 630–639; Wajja-Musukwe et al. 2008. Tree Physiology 28: 233–242; Walker et al. 2007. Agrofor. Syst. 70: 197–209; Wallace et al. 1999. Agric. For. Meteorol. 94: 189–202; Waltert et al. 2005. Ecological Applications 15: 1351–1366; Wilson et al. 1998. Exp. Agric. 34: 85–99

Table A2. Indicators of ecosystem services (ESs) provided by trees in sub-Saharan African agricultural landscapes. The indicators refer to the unit of study for assessing the contributions of trees to ESs. Numbers indicate the number of studies reporting the indicator for ES affected by trees. The presence of trees resulted in either increase, decrease or no effects on the indicators of ESs; indicated in Table 2, and Figure 3 and 4.

ESs	ES indicator	No. of studies
Food	Crop yield	162
	Yield of fruits and nuts	6
	Milk yield	3
	Number of fruit trees	4
Fodder	Digestibility (dry matter, crude protein, organic matter)	13
	Intake (feed, dry matter, crude protein)	7
	Animal growth rate	8
	Nutrient content (Ca, P, Mg, K, N)	21
	Biomass, dry matter	11
Wood	Biomass	22
	Wood products	1
	Number of trees	5
Organic fertilizer (green manure, compost, mulch)	Biomass	23
Gum	Gum yield	6
Genetic resources	Genetic diversity	4
	Seeds	1
Medicine	Number of farmers	3
Water cycling	Soil moisture content	55
	Infiltration rate	13
	Evapotranspiration	13
Pest regulation	Insect counts	19
	Damaged plants	9
	Disease incidences	4
	Weed biomass	6
	Weed abundance	13
Carbon storage	Carbon stocks	28
Microclimate	Transmitted light/light intensity	18
	Relative humidity	1
	Air temperature	3
	Soil temperature	6
Erosion control	Nutrient loss	4
	Runoff	7
	Soil loss	8
Wind break	Wind speed	4
Nutrient cycling	Nitrogen derived from the atmosphere/litter	20
	δ^15N values	11
	Foliar nutrient content (N, C, Ca, Cu, K, Mg, Mn, N, P, S, Zn, Fe)	125
	Nutrient balance (Ca, K, Mg, N, P)	3
	Nutrient mineralization from organic matter (N and partly P) and soil mineral weathering (K)	2
	Nutrient uptake (Ca, K, Mg, N, P)	36
Soil fertility	Aggregate stability	2
	Bulk density	7
	Cation exchange capacity (CEC, CECe)	19
	Exchangeable ions (Al, K, P, Ca, Mg, Mn, Na)	39
	Extractable elements (Al, Ca, K, Mg, Mn, Zn, Fe, P)	20
	Microbial biomass (C, N, P)	2
	Nutrient concentration (K, Mg, Ca, Fe, Mn, N, Na, P, S, Zn)	86
	Particulate organic matter (C, N)	8
	pH	64
	Soil (organic) carbon	62
	Soil organic matter (SOM)	21
	Soil texture	10

Table A3. Agroforestry practices (and other management options including trees) that have been reported from studies in agricultural landscapes in sub-Saharan Africa. A short description of the practices is given. The effects of trees on ES provision and biodiversity are summarized in Tables 2 and 3.

Code	Agroforestry practice^a	Description/objective
a	Alley cropping	Short- and long-term rows of trees within crops field, including hedgerows
b	Boundary planting	Short- and long-term trees along the boundaries of crop and pasture fields, including enclosures and hedges
c	Contour planting	Trees planted across slopes following contour lines to stabilize soil (erosion control), including trees on terraces
d	Dispersed intercropping	Intercropping of short- and long-term trees mixed and crops or pasture, including parklands
e	Fodder bank	Small, dense stands of leguminous trees and shrubs planted to serve as reservoir from where farmers collect high quality fodder for dry season feeding
f	Home gardens	Multipurpose (fruits, nuts, fodder, shade, etc.) trees and shrubs managed together with crops, vegetables and livestock within the compound of an individual household
g	Improved tree fallow	Planted tree species (successive or simultaneously with crops) used as an alternative to natural grass or bush fallow
h	Multi-strata agroforestry	Perennial tree crop systems such as coffee, cacao or tea with higher shade trees, including multi-storey tree gardens
k	Windbreaks (Shelterbelts)	Rows of trees or shrubs planted around edges of fields and homestead to lessen or break the force of wind
l	Woodlot	A stand in a section of the farm fully devoted to trees
Other management aspects
m	Beneath canopy	Experimental design that compares the variation of defined parameters under the canopy versus in the open
n	Biomass transfer	Plant biomass collected outside the crop-growing area is moved to crop fields to provide an input of nutrients
o	Pruning	Trees are pruned and the prunings used as mulch and green manure in crop fields
p	Fodder supplements	Fresh or dried tree forage given to livestock to supplement native forage, crop residues or grazing

^a Nair (1993); Sinclair (2004); Akinnifesi et al. (2010).