Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Home garden system dynamics in Southern Ethiopia

Home garden system dynamics in Southern Ethiopia Agroforest Syst (2018) 92:1579–1595 https://doi.org/10.1007/s10457-017-0106-5(0123456789().,-volV)(0123456789().,-volV) . . . Beyene Teklu Mellisse Gerrie W. J. van de Ven Ken E. Giller Katrien Descheemaeker Received: 28 June 2016 / Accepted: 18 July 2017 / Published online: 31 July 2017 The Author(s) 2017. This article is an open access publication Abstract Home gardens in southern Ethiopia are cereal-vegetable systems. In densely populated, mar- regarded as efficient farming systems, allowing inter- ket proximate areas a major trend was expansion of actions and synergies between crop, tree and livestock khat, from 6 to 35% of the area share per farm, while components. However, these age-old traditional home the combined area share of enset and coffee decreased gardens are evolving rapidly in response to changes in from 45 to 25%. Concurrently, the cattle herd size fell both the socio-economic and biophysical environ- from 5.8 TLU to 3.9 TLU per household. In medium ment. Altered cropping patterns, farm size and com- populated, less accessible areas the trend was consol- ponent interactions may affect the systems’ idation of combined production of food and cash sustainability. Home gardens exhibit a huge diversity crops. Enset and coffee together maintained a share of in farms and farming systems, which needs to be over 45%. Easy transport and marketing of the understood in order to design interventions for perishable cash-generating khat compared with tradi- improvement. Dynamics of home gardens were stud- tional crops favoured its cultivation among small- ied over two-decades (1991–2013) based on a survey holders located close to markets. The insights in home of 240 farm households and focus group discussions. garden change in response to increasing population Farms were grouped into five types: Khat-based, pressure, decreasing farm size and market develop- Enset-cereal-vegetable, Enset-based, Enset-coffee and ment may help to design interventions to increase Enset-livestock. Farm trajectories revealed a shift system sustainability. from food-oriented Enset-based and Enset-livestock systems to (1) cash crop oriented khat-based systems, Keywords Home garden types  Cropping patterns and (2) combined food and cash crop oriented Enset- Cash crops  Commercialization  Diversification Catha edulis B. T. Mellisse (&)  G. W. J. van de Ven K. E. Giller  K. Descheemaeker Plant Production Systems, Wageningen University and Introduction Research (WUR), PO Box 430, Wageningen, The Netherlands Home gardens are characterized by multipurpose trees e-mail: beyene.mellisse@wur.nl; beyteklu@gmail.com and shrubs in intimate association with annual and perennial crops and livestock around the homestead B. T. Mellisse (Kumar and Nair 2006). Home gardens ensure a Wondo Genet College of Forestry and Natural Resources, sustained availability of multiple products and Hawassa University, P.O.Box 128, Shashemene, Ethiopia 123 1580 Agroforest Syst (2018) 92:1579–1595 generate income (Kumar and Nair 2004). Despite their the traditional land use systems (Abebe et al. 2010; small size (Kumar and Nair 2006), home gardens fulfil Dessie and Kinlund 2008). Yet, little is known about most of the basic food and nutritional needs of the the rate at which those systems have shifted to mono- households, while the multi-storied configuration and cropping systems, how that differs across the region, high species diversity maintain their structure and and which factors could explain it. Other studies function in the face of external stress (Kumar and Nair (Herrero et al. 2009; Ebanyat et al. 2010) suggest 2004). trends in institutional support, resource endowment, In the Southern Nations Nationalities and Peoples’ prices, social conditions and technology as drivers of Regional State (SNNPRS) of Ethiopia, home gardens farming system change, but these have not yet been are the prevalent land use system covering about investigated for the particular case of home gardens in 576,000 ha, which is 31% of the region’s cultivable southern Ethiopia. land (Abebe 2005). The traditional ‘enset-coffee’ Similar to many other areas in sub-Saharan Africa, home gardens are characterised by the combination of SNNPRS is characterized by a huge diversity in farms two perennial crops: enset (Enset ventricosum and farming systems, even though they are grouped (Welw.) Cheesman) and coffee (Coffea arabica L.). under the common term home gardens. Recognizing Enset is a herbaceous, multipurpose crop and a staple this variability within and among farming systems and food for more than 15 million people (Abebe et al. localities is the first step in the design of new technologies to improve agricultural production (Gil- 2010). Enset leaves are a key livestock feed and used as mulch to reduce soil erosion and runoff (Amede and ler et al. 2011; Descheemaeker et al. 2016). Diro 2005). Hence, enset fulfils both productive and With the wider aim to contribute to the develop- protective functions (Kippie 2002). Coffee is mainly a ment of tailored interventions to improve agricultural cash crop, but is also consumed (Kippie 2002; Abebe production and productivity, this paper aims to: (1) 2005). understand the diversity in home garden systems of Increasing population pressure and subdivision of southern Ethiopia; (2) analyse trends in farm size, farms have led to fragmentation of land, a decline in cropping pattern and livestock population over the the area under coffee and enset (Tsegaye and Struik past two decades; (3) explain the spatial patterns of 2001; Abebe 2005) and gradual replacement of the change in home gardens in relation to market access, age-old diverse system. A major change is the agro-ecological conditions and population density. expansion of a mono-cropping system of khat (Catha We combined quantitative and qualitative methods, edulis Forsk) at the expense of enset-coffee home including secondary data collection on demography gardens (Tsegaye and Struik 2001; Abebe et al. 2010). and farm size, a farm typology based on the area share Khat is grown for its economically important leaves of crops and grazing land, and constructed a 20-year and tender twigs, which are chewed for their stimu- timeline of farming system change based on detailed lating effect. Due to the decline in enset cultivation, household interviews concerning current and past many households have become more dependent on the farm assets and practices. food market (Amede and Diro 2005). Market depen- dency for family food is further exacerbated by decreasing farm area and productivity of food crops Materials and methods (Amede et al. 2001). In countries like Ethiopia, where smallholder farmers have no access to insurance, the The study area market dependency increases the vulnerability to economic or environmental shocks. The replacement Sidama and Gedeo are representative zones of the of enset has also induced a shortage of livestock feed SNNPRS region with a prevalence of enset-coffee with direct repercussions on herd size, herd compo- home gardens. These home gardens are characterised sition, livestock production and hence, nutritional by the production of enset, coffee and multi-purpose quality of human diets (Tsegaye and Struik 2001). trees, accompanied by root and tuber crops, vegeta- Most studies on the home gardens of southern Ethiopia bles, annual cereal crops and livestock keeping claimed that expansion of khat has resulted in (Kippie 2002; Abebe et al. 2010). Sidama is located 0 0 0 0 homogenization of the structure and composition of within 545 –645 N and 385 –3941 E, covering a 123 Agroforest Syst (2018) 92:1579–1595 1581 total area of 7672 km (Abebe 2005) with 3.50 million 2010) and (3) three classes of population density, i.e. inhabitants (CSA 2007). Gedeo is located within high ([800 person/km ), medium (500–800 person/ 2 2 0 0 0 0 550 –612 N, 3803 –3818 E, covering a total area km ) and low (\500 person/km ). of 1347 km (Kippie 2002) with 0.84 million inhab- On average 10 households per PA, 240 households itants (CSA 2007). Most kebeles (smallest Ethiopian in total, were selected for the household survey. The administrative unit, also called peasant association or districts’ Bureau of Agriculture categorized farms as PA) in both zones are classified as ‘rural’ (CSA 2007). small, medium and large based on the area of their land The rainfall distribution is bimodal with a long (June holding and the number of livestock. Farms with to September) and short (March to May) rainy season. [1.25 ha of land and [5 TLU were categorized as Both Sidama and Gedeo straddle two agro-ecological large, 0.5–1.25 ha and 1–5 TLU as medium, and zones, the moist mid-altitude (In Amharic: woina- \0.5 ha and 0–1 TLU as small. Stratified random dega) and the moist highland (In Amharic: dega). The sampling ensured that the number of selected house- moist mid-altitude zone ranges in elevation from 1500 holds in each group was in line with their share in each to 2300 m a.s.l. and receives 1200–1600 mm rainfall of the PAs. annually; the average annual temperature ranges from 16 to 22 C. The moist highland zone comprises an Data collection elevation range of 2300–3200 m a.s.l.; mean annual rainfall amounts to 1600–2000 mm and the average We combined qualitative and quantitative data col- annual temperature ranges from 15 to 19 C. The soils lection approaches to triangulate the relevance and in the mid-altitude and highland agro-ecologies are accuracy of the information gathered (Ebanyat et al. mainly characterized by clay-loam to silt-loam and 2010). Eight focus group discussions (FGDs) were sandy-loam to sandy-clay textural classes organized, four in midland PAs (Ywo, Kochow, respectively. Soyama and Dero) and four in highland PAs (Gikeatoye, Guguma, Duba and Elalcha). In each Selection of study sites and farm households FGD eight to ten farmers participated representing different age groups, both sexes and each farm size Within Sidama and Gedeo zones, four districts (In category. The discussion focused on positive and Amharic: woreda) were selected (Wondo Genet, negative aspects of khat expansion, constraints to Melga, Dale and Bule) and within each district, 6 maintain enset and coffee, population growth and its PAs, giving 24 PAs in total (Fig. 1). The multistage impact on resource availability (land), change in sampling approach for districts and PAs was designed income sources, emergence of new activities, eco- to encompass differences in distance to markets, agro- nomic infrastructure (roads and transport service), ecological conditions and population density market development and price trends. (Table 1). These factors were hypothesised to affect A detailed household survey was administered farmers’ crop allocation decisions through their influ- between August and November 2013 with 240 farm ence on the ease of transportation of agricultural households. We used a semi-structured questionnaire products, crop suitability and farm size, respectively. to collect data on changes, trends and events in the past Distance from a major market and altitude were two decades with the aim to construct a timeline. To measured using a global positioning system (GPS) capture the changes in cropping and livestock systems, following roads and foot paths. The population density household heads were asked to recall the year of khat of each PA was collected from the Ethiopian Central introduction, land allocation to various annual and Statistics Agency (CSA 1991–2013). The PAs were perennial crops (food crops, cash crops), total land categorized into (1) three classes of distance to market, holding and herd size from 1991 to 2013 in a stepwise i.e. near (\36 km), medium (36–70 km) and far fashion. The year of the regime change in Ethiopia, ([70 km), roughly corresponding to travelling times 1991, and the following national election years (1994, of 1.5, 3 and more than 3 h by car, respectively, (2) 2000, 2005 and 2010) were used as benchmarks to two elevation classes, based on agro-ecological zon- help respondents’ recall. Household heads were fur- ing, i.e. high ([2300 m a.s.l.) and medium ther asked about demographic characteristics (family (1500–2300 m a.s.l.) (Kippie 2002; Abebe et al. size, level of education), production objectives, 123 1582 Agroforest Syst (2018) 92:1579–1595 Table 1 Peasant associations (PA), numbers of surveyed households in each of the three farm size categories, distance to market, elevation and population density. Source: Ethiopian Central Statistical Authority (CSA 1991–2013) No. Site (PA) District Sample Sampled farms by farm size Distance to Elevation Population (Woreda) size major market (m a.s.l.) density Large Medium Small 2 place (km) (person/km ) Mid-land agro-ecology (1500–2300 m a.s.l.) 1 Ywo Wondo Genet 11 1 4 6 N 12.4 1679 H 958 2 Chuko Wondo Genet 12 1 3 8 N 26.4 1756 H 1551 3 Weterakechema Wondo Genet 7 1 2 4 M 36.4 1755 M 714 4 Kochow Melga 10 2 4 4 N 24.4 2145 H 834 5 Manicho Melga 11 1 8 2 N 25.4 2190 H 858 6 Berana Melga 9 1 6 2 N 30.9 2216 H 915 7 Soyama Dale 9 1 6 2 M 36.3 1787 H 852 8 Megera Dale 10 1 4 5 M 47.8 1746 M 566 9 Awada Dale 11 2 2 7 M 44.7 1753 L 433 10 Dero Bule 11 1 4 6 F 126.8 2245 M 792 11 Bassura Bule 11 3 3 5 F 114.8 2253 M 773 12 Osselemajo Bule 8 2 2 4 F 118.5 2250 M 737 Sub total 120 17 (14%) 48 (40%) 55 (46%) Highland agro-ecology (2300–3200 m a.s.l.) 13 Weteragedo Wondo Genet 8 3 3 2 M 38.1 2597 H 945 14 BaboChorora Wondo Genet 9 2 4 3 M 40.5 2624 M 646 15 Gikeatoye Wondo Genet 13 3 5 5 M 44.9 2599 H 1565 16 Fitoketemuna Melga 12 4 4 4 M 41.1 2561 H 950 17 Gerewe Melga 9 3 4 2 N 34.6 2634 M 508 18 Guguma Melga 9 3 2 4 M 47.1 2677 M 583 19 Hayello Dale 8 2 2 4 F 70.2 2328 H 1023 20 Duba Dale 10 1 2 7 M 47.4 2338 H 1153 21 Gedamo Dale 12 1 3 8 M 64.7 2339 L 102 22 Elalcha Bule 11 3 4 4 F 126.9 2902 L 469 23 Akollo 1st Bule 11 2 4 5 F 142.9 2787 L 272 24 Suko Bule 8 2 2 4 F 135.1 2709 L 43 Sub-total 120 29 (24%) 39 (33%) 52 (43%) Total 240 46 (19%) 87 (36%) 107 (45%) N near, M medium, F far, H high, M medium, L low sources of income, constraints to crop production and f  j ij j¼1 M ¼ livestock rearing and dependency on the market for i food (number of months per year). Farmers also listed where M = mean score of factor i; f = the percent- i ij and ranked political, cultural, environmental and age of respondents that gave a score of j for factor economic factors triggering the changes in land i and n = the total number of respondents that listed allocation and farming practices based on a score of factor i. one (low importance), two (average importance) or Enumerators with a certificate in agriculture and three (high importance). The mean score of each factor speaking the local language were recruited from the was calculated using the formula 123 Agroforest Syst (2018) 92:1579–1595 1583 localities and trained in data collection. Retail prices of southern Ethiopia. Five different home garden types coffee, khat, kocho (the processed edible part of enset), were identified and characterized by the area share of tef (Eragrostis tef (Zucc.) Trotter) and maize (Zea the dominant crops and grazing land (Table 2; Fig. 2). mays L.) were collected from CSA. The price in Our decision tree (Fig. 2) dictated that throughout the Ethiopian birr was converted to US$ at a rate of 2.5, study period, farms with an area share of khat over 6.3, 8.6, 17.5 and 19.1 Eth. birr to one US$, the average 30% were classified as Khat-based systems. With less exchange rate in 1991–1994, 1995–1999, 2000–2008 than 30% khat, home gardens were described as Enset- and 2009–2012 and 2013–2014 respectively (NBE coffee if at least 35% of their area was covered with 1991–2014). The revenue per hectare was calculated coffee. Subsequently, Enset-livestock systems were by multiplying the price of each crop with the average identified if grazing land covered more than 20%. Of yield (assumed to be constant over the study period). the remaining two types, Enset-based and Enset- cereal-vegetable systems were distinguished based on Farm typology an enset area share of at least and less than 35% enset, respectively. We constructed a farm typology based on crop In Khat-based and Enset-based home gardens khat allocation data from the household survey after and enset covered 51 and 58% of the farm land exploring the data using Ward’s minimum variance respectively (Table 2). The combined area of enset, method (Joe and Ward 1963), followed by a K-means cereals and vegetables covered more than 60% of the clustering (Blazy et al. 2009). The area shares of nine area in the Enset-cereal-vegetable system, while 89% crops (enset, coffee, khat, maize, other annual cereals, of the area in Enset-coffee home gardens was covered beans, root and tuber crops, vegetables and sugarcane) by enset and coffee. In Enset-livestock home gardens and grazing land were considered. Each of the the combined area of enset and grazing land covered identified farm types was named after the crops with more than 50% of the farm area. The highest the major area shares in 2013. In order to classify farms proportion of cash crops was observed in Khat-based based on the same criteria over time, thresholds of area home gardens (60%) while the highest proportion of shares were derived based on the variation within each food crops was observed in Enset-based home gardens farm type, following Falconnier et al. (2015). Consid- (74%). The largest proportion of grazing land was ering these thresholds, we developed a decision tree to documented in Enset-livestock systems (34%). The classify each farm into a distinct farm type for the six average family size in Khat-based home gardens was benchmarks in the timeline 1991–2013. significantly larger than that in Enset-based and Enset- livestock home gardens (Table 2). The average farm Data analysis size showed no significant difference between home garden types, and the Khat-based home gardens had A one way analysis of variance (ANOVA) was used to the smallest land to labour ratio. test the difference between the means of five home garden types in family size, farm size, land to person Distribution of home garden types across districts ratio, herd size, income and market dependency. and socio-economic and biophysical environments Differences were deemed significant at P \ 0.05. Analyses were performed using the Statistical Package The four districts of the study area showed a different prevalence of the five home garden types. Wondo for Social Sciences (SPSS) version 20. Qualitative data was used to support the discussion of results. Genet and Melga were characterized by the domi- nance of two home garden types, namely Khat-based and Enset-cereal-vegetable in Wondo Genet and Results Khat-based and Enset-livestock in Melga. In both districts Enset-coffee systems were absent. Dale and Characteristics of the home garden types Bule showed a more even representation of four home garden types. The Khat-based home garden The variability in land use indicated a large diversity type was not present in Bule and nearly absent in in farming systems in Sidama and Gedeo zones of Dale. 123 1584 Agroforest Syst (2018) 92:1579–1595 Fig. 1 Map of the study districts (woredas) within Sidama and Gedeo zones, SNNPRS, Ethiopia Thepresenceofthe homegardentypesdiffered Dynamics in cropping pattern and livestock holding in the four districts in relation to market access, population density and agro-ecology (Fig. 3). Close to markets, 56% of the The overall trend in cropping patterns in Wondo Genet farms were Khat-based (Fig. 3a). At medium and far distance their share was 17 and 0% respectively. and Melga indicated a shift from food to cash crop production over time, while in Dale farmers continued The shares of Enset-cereal-vegetable, Enset-based and Enset-coffee home garden types all decreased to produce a combination of food and cash crops and in with proximity to the market. The distribution of Bule they remained food crop oriented (Fig. 4). In home garden types in relation to population density 1991 the combined area share of enset and coffee showed a similar pattern (Fig. 3b), with Khat-based amounted to more than 45% in all districts of the study systems gaining importance with increasing popu- area. In 2013 it had fallen below 25% in Wondo Genet lationdensity andanopposite trend for Enset- and Melga, whereas the area share of khat increased from about 5% in 1991 to about 35% in 2013 (Fig. 4). cereal-vegetable, Enset-based and Enset-coffee home garden types. Population density had no clear Even though khat cultivation was observed in 1991 also in Dale, it expanded less there from 0.9% in 1991 effect on the presence of Enset-livestock home gardens. In the mid-altitude zone Khat-based and to 8% in 2013. In Bule, khat has not played any role worth mentioning in the past two decades (Fig. 4). Enset-coffee home gardens were more prevalent, while in the high-altitude zone Enset-based and Unlike the cropping pattern, the grazing land area was Enset-livestock systems were more common not noticeably influenced by the expansion of khat and (Fig. 3c). was maintained at 8–12% in Wondo Genet and 123 Agroforest Syst (2018) 92:1579–1595 1585 Table 2 Average (±standard error of the mean) crop alloca- P \ 0.05 are indicated with different letters within a row; tion and household characteristics for the five home garden dominant crops and grazing land that lend their name to the types; significant differences in household characteristics at type are indicated in bold Description Home garden types Khat-based Enset-cereal- Enset-based Enset-coffee Enset- (K-b) vegetable (E- (E-b) (E-c) livestock c-v) (E-l) Number of farms 58 51 43 30 58 Share in the farm area (%) Food crops Enset 26.9 ± 1.7 18.8 – 1.3 57.6 – 2.2 39.9 – 2.2 21.7 – 1.4 Annual cereals 5.9 ± 1.4 24.3 – 1.9 11.6 ± 2.2 1.8 ± 1.0 13.5 ± 1.5 Other food crops 0.4 ± 0.2 11.8 ± 1.3 4.8 ± 1.6 2.3 ± 1.1 7.1 ± 1.5 Total food crops 33.2 ± 2.1 54.9 ± 2.5 74.0 ± 2.6 44.0 ± 2.4 42.3 ± 2.2 Cash crops Coffee 3.0 ± 0.6 5.6 ± 1.3 10.8 ± 1.8 49.4 – 2.0 3.4 ± 1.0 Khat 51.2 – 2.0 11.0 ± 1.7 4.1 ± 1.4 2.0 ± 1.3 14.7 ± 1.6 Vegetables 3.8 ± 1.0 18.7 – 2.2 5.8 ± 1.8 1.4 ± 0.8 4.8 ± 1.0 Other cash crops 1.7 ± 0.8 3.4 ± 1.1 1.8 ± 0.9 2.1 ± 0.8 0.8 ± 0.2 Total cash crops 59.7 ± 2.2 37.8 ± 2.6 21.8 ± 2.6 54.9 ± 2.5 23.7 ± 1.8 Grazing land 7.1 ± 1.2 7.3 ± 1.1 4.2 ± 1.3 1.1 ± 0.5 34.0 – 1.7 Household characteristics Family size (#) 9.0 ± 0.5a 8.8 ± 0.4a 7.0 ± 0.5b 8.8 ± 0.8ab 7.3 ± 0.5b Farm size (ha) 0.8 ± 0.1a 0.8 ± 0.1a 1.0 ± 0.1a 0.80 ± 0.2a 1.0 ± 0.1a Land: labour ratio 0.13 ± 0.0b 0.18 ± 0.0ab 0.27 ± 0.1a 0.25 ± 0.0a 0.25 ± 0.0a Cattle (TLU) 3.8 ± 0.4a 3.8 ± 0.4a 2.8 ± 0.5b 1.6 ± 0.2b 4.5 ± 0.4a Small ruminant 0.4 ± 0.0b 0.7 ± 0.1a 1.0 ± 0.2a 1.5 ± 0.3a 1.0 ± 0.1a (TLU) Income (10 Eth. 28.0 ± 8.0a 18.0 ± 2.0ab 7.0 ± 1.0c 12.0 ± 3.0b 16.0 ± 3.0b Birr) Market dependency 5.3 ± 0.4a 2.9 ± 0.4b 3.9 ± 0.5a 3.1. ± 0.6ab 3.2 ± 0.4b for food (months) Number of household heads Illiterate 10 (17%) 7 (14%) 13 (30%) 2 (7%) 21 (36%) according to educational level Write and read 7 (12%) 3 (6%) 7 (16%) 5 (16%) 14 (24%) Elementary 33 (57%) 27 (53%) 17 (40%) 17 (57%) 20 (34%) Secondary 8 (14%) 14 (27%) 6 (14%) 6 (20%) 3 (5%) TLU tropical livestock unit Other food crops include: beans, root and tuber crops Other cash crops include: sugarcane 20–23% in Melga for the last two decades. Grazing Bule from 4.8 to 3.4 TLU over the same time. In land area remained below 10% in Dale, and in Bule it Wondo Genet the cattle holding steadily decreased varied between 8 and 17%, with a share of 13% in from 5.8 TLU in 1991 to 3.9 TLU in 2013. The number 2013. of small ruminants varied between districts, but In the period 1991–2013 the cattle and small changed little over time. Only in Wondo Genet it ruminant population dynamics varied among the four tended to decline from 0.9 TLU in 1991 to 0.5 TLU in districts. In three districts the cattle holding decreased 2013, but also here it varied between years. In Melga and only in Melga it increased from 5.1 TLU in the number of small ruminants per farm represented 1994–2000 to 5.9 TLU in 2005–2013. In Dale the 0.5 TLU, in Dale it varied between 0.1 and 0.3 and in cattle holding decreased from 2.7 to 2.1 TLU and in Bule between 1.0 and 1.3 TLU. 123 1586 Agroforest Syst (2018) 92:1579–1595 Fig. 2 Decision tree and threshold values for area share used for farm classification in five home garden types. Number of farms (n) refers to the year 1991 Fig. 3 Share of home garden types in 2013 classified according to a market access, b population density and c elevation 123 Agroforest Syst (2018) 92:1579–1595 1587 Dale Wondo Genet 1991 1994 2000 2005 2010 2013 1991 1994 2000 2005 2010 2013 Bule Melga 80 80 20 20 1991 1994 2000 2005 2010 2013 1991 1994 2000 2005 2010 2013 Fig. 4 Land use dynamics in the four districts of the study area in the period 1991–2013 Dynamics in the prevalence of home garden types half of the Enset-livestock home gardens. Also in in the period 1991–2013 Melga the major emerging type was the Khat-based home garden, increasing from 3 to 39% of the farms at In 1991, about 30% of the sampled farms were the expense of Enset-based and Enset-coffee home engaged in Enset-based and 30% in Enset-livestock gardens, which fully disappeared by 2013. farming, while only 3% was engaged in Khat-based The farm transitions were different in Dale and farming (Fig. 5). The proportion of Enset-livestock Bule districts. In Dale the emergence of the Khat- and Enset-based home gardens decreased to 24% and based home garden type was restricted to only 3% of less than 20% of the farms respectively in 2013 the farms, but an increase in the proportion of the (Fig. 5). The share of Enset-coffee home gardens Enset-coffee system was noticed (Fig. 6). In Bule the declined from 19% in 1991 to 13% in 2013. In Enset-based and Enset-cereal-vegetable home gardens contrast, the share of khat-based farms increased from gained some importance at the expense of the Enset- 3% in 1991 to 24% in 2013. The percentage of farms in livestock and Enset-coffee home gardens (Fig. 6). the Enset-cereal-vegetable system slightly increased from 17% in 1991 to 21% in 2013 (Fig. 5). Determinants of farming system dynamics The transitions in home garden systems varied across the four districts of the study area (Fig. 6). In Respondents identified population pressure, policy, Wondo Genet most Enset-based home gardens in 1991 infrastructure, profitability, decline in soil fertility, transitioned to Khat-based or Enset-cereal-veg- cultural change, lack of input, labour availability and etable home gardens in 2013, and the proportion of harvesting frequency of khat as major determinants of Enset-based farms dropped from 44% in 1991 to 7% in land use change in the area. Among these nine factors, 2013. The farms practicing Enset-coffee farming in population pressure was ranked as the major driver of 1991 have all transitioned to either Khat-based or land use change. In all four districts population Enset-cereal-vegetable home gardens in 2013, as have pressure increased (Fig. 7a) and contributed to the Area share (%) 1588 Agroforest Syst (2018) 92:1579–1595 Fig. 5 Proportion of each farm type in 1991 and 2013 and their transition during this period in Sidama and Gedeo zones (n = 240 farms). The stacked bar in the middle represents the transition of farms from a given type in 1991 to the types in 2013; the height of each stacked cell and bar is representative of the number of farms Fig. 6 Changes in the 1991 khat-based prevalence of home garden 10% 10% 20% 7% 8% types in the four districts of Enset-cereal-vegetable the study area over the Wondo Enset-based 55% period 1991–2013 18% Genet 28% Enset-coffee 44% Enset-livestock 3% 14% 39% 45% Melga 50% 30% 8% 3% 8% 20% 0% 20% 20% 20% Dale 27% 27% 33% 33% 20% 27% 33% 33% Bule 17% 17% 30% 23% 123 Agroforest Syst (2018) 92:1579–1595 1589 Fig. 7 Population density (a) and change in land holding per household (b)in the four districts between 1991 and 2013 1991 1994 2000 2005 2010 2013 1992/94 1995/00 2001/05 2006/10 2011/13 0.0 -1.0 -2.0 -3.0 -4.0 -5.0 -6.0 decline in farm size and land fragmentation. In Wondo The price of staple foods (maize, kocho and teff) Genet where population density doubled between rose gradually over the two decades (Fig. 8a). Among 1991 and 2013 the rate of decline in land holding per the food crops, the teff price showed the largest household increased from 0.5% per year in the period increase, from 0.4 US$ per kg in 1990/1 to 0.8 US$ in 1992/94 to 5.1% per year in the period 2011/13 2013/14. The price of maize was very close to that of (Fig. 7). In Melga, Dale and Bule districts, where the kocho, increasing from about 0.2 to 0.3 US$ per kg increase in population density was less pronounced, over the whole period. The price of khat fluctuated the rate of decline in land holding did not exceed 3%. around 3.5 US$ per kg of fresh weight in 1990/1 to A market liberalization policy implemented in 1991 2013/14, and for coffee, it fluctuated between 1.3 and and constitutionalized in 1995, which eliminated crop 4.3 US$ per kg in the same period (Fig. 8a). and livestock sale quota and price controls, was ranked Besides the price fluctuation, the greater income as a second determinant. As a third factor farmers per hectare from khat may have further discouraged mentioned the cultivation of khat, which became production of coffee. The average revenue per popular thanks to its frequent harvesting and high hectare from khat by far exceeded the average return per unit area. According to the respondents, the revenues from food crops (maize, teff and kocho) shift to khat cultivation was further fuelled by the (Fig. 8b). As a result of the continuous fragmenta- development of infrastructure (road, mobile phone) tion of land, the inability to produce sufficient and urban expansion. amount of family food also discouraged smallhold- Declining soil fertility coupled with a lack of farm ers to allocate land to food crops. Based on the inputs (seed, fertilizer and manure) also influenced the assumption that 50% of the energy demand is land use dynamics of the study area. Farmers further provided by enset and 50% by maize, the area referred to cultural change, and in particular to required to produce year-round food for an average changing diets. Labour availability was ranked last household of three adults and five children was in terms of its influence on land use change. 0.46 ha. A yield of 6.2 ton per ha for enset and Change in land Population density holding/household/year (%) (Person/km2) 1590 Agroforest Syst (2018) 92:1579–1595 Years Fig. 8 Annual retail prices of food and cash crops (a) and the average maize, teff and coffee yield per ha summarized from revenue per hectare (b) over the period 1991–2014. (Source: the CSA statistical bulletin (2000–2013); average yield for Retailer prices calculated and summarized from price lists for kocho and khat per ha from ATA (2012) the period 1991–2014 from CSA and Regional Custom Offices; 2.3 ton per ha for maize were considered (ATA where khat was introduced and expanded over the last 2012). If these households shift to khat and buy all two decades (1991-2013). This trend was observed in their food, they need only 0.05 ha, assuming a khat particular in Wondo Genet and Melga (Fig. 4) and revenue of 9368 US$ per ha (Fig. 8b) and a price simplified the cropping pattern with khat monoculture 0.3 US$ per kg for enset (kocho) and maize now occupying more than one-third of the farm area. (Fig. 8a). A similar expansion of khat was observed in the eastern highlands of Ethiopia, where it replaced sorghum (Feyisa and Aune 2003; Mekbib 2009) and Discussion in the Jima zone of southern Ethiopia where it replaced coffee (Dube et al. 2014). Expansion of khat cultivation Khat’s popularity with smallholders is related to its profitability, its regular harvesting (two to three times Cropping patterns in Sidama and Gedeo zones of per year) and its quick establishment (one to two southern Ethiopia changed most strongly in areas years). On the one hand, the process of -1 Revenue per hectare (thousand US$) Price (US$ kg ) 1990/1 1990/1 1992/3 1992/3 1994/5 1994/5 1996/7 1996/7 1998/9 1998/9 2000/01 2000/01 2002/03 2002/03 2004/05 2004/05 2006/07 2006/07 2008/09 2008/09 2010/11 2010/11 2012/13 2012/13 Agroforest Syst (2018) 92:1579–1595 1591 commercialization and specialization in cash crop Similar diversification with French beans or tomatoes monoculture may expose smallholder farmers to was reported among smallholder coffee growers in unpredictable market fluctuations, disease outbreaks central Kenya (Dorsey 2015). The degree of crop- and other adverse shocks, with a risk to become food livestock integration varied between farming systems, aid dependent (Amede and Diro 2005). Farmers are and the prevalence of competitive versus complemen- particularly vulnerable if institutions, policies, and tary roles of crops and livestock partly explains markets are non-supportive (Von Braun 1995), as in dynamics in livestock herd sizes. The advantage of the case of the study area, where e.g. crop insurance combining crop and livestock activities in the tradi- schemes are non-existent. Similarly, other studies tional systems was related to the availability of cheap reported the incapability of small commercialized fodder for livestock, such as enset leaves, and of farms to continue cash crop cultivation following a animal manure, contributing to enset and coffee decline in output prices, after which they reverted back growth. In the traditional home gardens, livestock to subsistence cropping (Wiggins et al. 2011). On the also played a critical role in providing a protein and other hand, the increased income from cash crops may nutrient rich diet to the household, given the low enable smallholders to buy food crops rather than nutritious value of enset (Brandt et al. 1997; Tsegaye having to produce everything themselves (Timmer and Struik 2001). The transition towards khat cultiva- 1997). tion reduced herd sizes due to decreased fodder Khat’s popularity with smallholders is further availability, especially from enset, and also lowered driven by disadvantages of the traditional crops, such the interest in livestock rearing as a source of cash and as the long time before enset plants can be harvested manure. For example, in Wondo Genet the khat (3–4 years after transplanting) and the unreliable price expansion went hand in hand with a decline in small and long establishment period of coffee. Thus, in the ruminant numbers. If the demand for cash is met by context of the study area, the incorporation of khat in khat, the need for keeping small ruminants recedes. the production system may be viewed as an innovation Moreover, khat is typically fertilized with mineral to take advantage of market opportunities towards fertilizer, reducing the need for animal manure. In achieving food security and improving household addition, the feeding behaviour of small ruminants, income (Rehima et al. 2013). browsing leaves and twigs, is incompatible with khat, as leaves and twigs are the economic product. Farming system transitions Similarly, less interest in keeping draught animals was reported where khat, which does not require The historical analysis of farming system transitions ploughing, replaced annual cropping systems in east- indicated a shift away from the traditional home ern Ethiopia (Kandari et al. 2014). Contrastingly, gardens based on enset, coffee and livestock (Fig. 5), small ruminant herd sizes did not decrease in Dale and illustrating the inability of these systems to accom- Bule, where the share of enset and coffee remained modate the increasing population. Whereas the tradi- fairly constant (Fig. 4). This could be related to the tional systems are known to support very dense availability of browse tree species for goats and -2 populations of up to 500 persons km (Kippie understory grazing for sheep in coffee fields. 2002; Abebe 2005), the population density of all The decrease in cattle holding per household in the study districts, except Bule, has surpassed this density khat-dominated district of Wondo Genet occurred together with the replacement of enset (Fig. 4). Enset since 2000 (Fig. 7). In areas far away from major roads and markets leaves have a high protein content (Brandt et al. 1997; farmers were less inclined to shift to khat, because of Tsegaye and Struik 2001; Solomon et al. 2008) and a less stringent land constraints on the one hand and feeding value that is comparable to a good quality difficulties related to the marketing of the perishable grass (Fekadu and Ledin 1997). Similarly, Kandari khat twigs on the other hand. Hence, farms transi- et al. (2014) reported a decline in per capita herd size tioned to Enset-based systems or modified the tradi- following complete replacement of major fodder crops tional systems by incorporating annual crops like by khat in Harar region of eastern Ethiopia. In Melga cereals and vegetables, leading to the expansion of we observed the opposite trend of a slight increase in Enset-cereal-vegetable systems in Bule (Fig. 6). cattle ownership, which could be related to the 123 1592 Agroforest Syst (2018) 92:1579–1595 absence of feed shortage in this district, allowing was present. The district is far from the market farmers to expand their herd size. This hypothesis is ([100 km), population density in 2013 was medium -2 supported by the relatively large grazing land area per (750 people km ) in the mid-altitude zone and low in -2 household and the fact that smallholders move their the highlands (250 people km ). Furthermore, the cattle to nearby lowland areas during the rainy season highland altitude of 2700–2900 m a.s.l. is above the and conserve their own pasture land for dry season suitable range for khat (Lemessa 2001). Melga closely grazing (focus group discussion). The smallest herd resembled the situation in Wondo Genet, with the sizes were observed in Dale, where the area of grazing difference that population growth and also land land was particularly small. Indeed, livestock feed fragmentation where somewhat lower. As a result, scarcity due to lack of grazing land is often reported as the area share of khat-based systems and the area the major constraint of livestock production also in under cash crops were smaller than in Wondo Genet in other districts of Sidama zone (Samuel 2014). 2013. Dale also took an intermediate position. It is at medium distance from the market, leading to a risk of The main drivers of changes in land use and farm quality loss for khat during transport (Distefano 1983). type transition It had a medium population density, similar to Melga, and the altitude did not exceed 2350 m a.s.l., posing no The main land use change, the replacement of Enset- agro-ecological constraint to khat. The improved road oriented systems by Khat-based systems, was observed networks around Hawassa also benefitted the districts in areas close to markets (Fig. 3). The improved road further from the market, such as Dale. networks and the proximity to the regional capital city, An increase in population also results in a growing Hawassa, opened up market opportunities for the labour force. Khat requires about three times more farmers, especially in Wondo Genet. The importance labour than cereal crops on a per hectare basis of a nearby market is related to khat’s perishable (Getahun and Krikorian 1973). However, in the nature, with a maximum shelf-life of four days after densely populated study area there are no labour picking (Distefano 1983). Also in the Hararghe shortages that could have constrained khat expansion. highlands of eastern Ethiopia, the proximity of a Some authors argue that khat may expand even market and easy transportation were determinants of further (Dessie 2013). On the other hand, the khat expansion (Woldu et al. 2015). In addition, the controversy and uncertainty about a possible prohi- high income potential motivated smallholders to bition of its cultivation seem to restrain smallholders allocate more land to khat cultivation. The revenue from allocating all of their land to khat (Dessie and per ha from khat exceeded that of maize and teff 16 Kinlund 2008). times and of coffee three times (Fig. 8b). Dessie (2013) reported similar values, fifteen times for cereals and Implication and options for future developments four times for coffee. In Wondo Genet, the expansion of khat was further stimulated by the growing popu- The traditional home garden systems in the SNNPRS lation density (Fig. 7). The resulting land fragmenta- can no longer absorb the increasing population. Part of tion, particularly after 2005 when population density them have transitioned to intensive khat-based sys- -2 exceeded 800 persons km (Figs. 4, 7), left khat tems, replacing the cultivation of enset and coffee by cultivation as one of the few alternatives to make a khat and hence producing less food. This development living from the tiny farm areas. The strong positive is associated with various risks such as an increased relationship between area share of khat and population dependency on one crop, khat, and hence vulnerability density was also confirmed in eastern Ethiopia (Tefera to harvest failure, price fluctuations, and pests and 2009). In Wondo Genet the total area allocated to cash diseases. This study has shown that the trends in land crops increased over time to almost 50% (Fig. 4). The use and farm types differed between districts depend- inclination of smallholders towards cultivating cash ing on market access, local population density and crops when smaller farm sizes impair food self- elevation. Taking into account this observed hetero- sufficiency was also described by Woldu et al. (2015). geneity is crucial in order to identify adapted inter- In contrast to Wondo Genet, in Bule none of the ventions to achieve social and economic development identified drivers for khat introduction and expansion of smallholder farmers. We have identified several 123 Agroforest Syst (2018) 92:1579–1595 1593 Conclusion options for locally adapted interventions. For the khat- based home gardens, diversifying agricultural produce We found a huge diversity of home garden systems in by intercropping khat with annual crops such as cereals and legumes could reduce the risk of depend- southern Ethiopia and their dynamics over the last two decades. Five home garden types were distinguished ing primarily on one crop. Khat is the main crop receiving artificial fertilizer, which in intercropping based on cropping patterns. Smallholders specialized in khat production typically allocate about half of their systems will also benefit the food crops, contributing to food security. Crop residues can be left in the field to farm land to khat, as such depending on the market for improve soil fertility. The integration with livestock is about half of the year for family food requirements, but difficult as not much feed becomes available from this earning a good income. Contrastingly, farmers special- intercropping system. ized in food production allocated about 75% of their farm to food crops, thus generating significantly less Enset-oriented home gardens offer scope in areas with lower population densities as enset takes several income. Over the past two decades cropping patterns had changed considerably, showing: (1) a shift from years to yield. A more intense coupling of enset with crossbred (i.e. Boran-Holstein–Friesian) milking cows food crop to cash crop production in densely populated areas close to markets; (2) a continuation of combined or small ruminants could be of mutual benefit, as the enset leaves can be used as animal feed and enset food and cash crops in medium populated, less acces- sible areas. In line with changes in land use and livestock needs manure for a good production. The time required for enset to reach maturity is up to eight populations, two main trends in home garden systems years without manure application, compared to three were observed: (1) a transition to cash crop oriented to four years with manure application (Shumbulo et al. home gardens, and (2) development of combined food 2012). Milk and meat could improve the nutritional and cash crop oriented home gardens. These dynamic changes were influenced by changes in population quality of family diets, whereas other livestock functions like banking functions also contribute to density and market access, changes in prices, a decline in soil fertility and a policy of market liberalization. Our rural livelihoods (Moll 2005;Thorntonetal. 2007). In areas where grazing land is scarce the integration with insights on how enset and coffee based home gardens of southern Ethiopia have responded to increasing popu- livestock is not feasible. Here, introducing locally adapted improved coffee and enset varieties could lation pressure and commercialization provide insights contribute to enhance the income and food security of that can inform the design of alternative options for smallholder farmers through improving the productiv- sustainable development. ity. The shortage of manure to maintain soil fertility Acknowledgements This research was funded by the can be addressed by the use of compost, prepared government of the Netherland through the project capacity from coffee husk waste (Abebe 2013) and the inedible building for scaling up of evidence-based best practices in portions of enset leaves and, pseudostem sheath and agricultural production in Ethiopia (CASCAPE). We thank the corm (Tamire and Argaw 2015). There is also scope to farmers who participated in this research. Our thanks are extended to experts from Bureau of Agriculture in Wondo integrate high value and productive fruit trees with the Genet, Melga, Dale and Bule districts for making available aim to maintain the diversified nature of the system secondary data and for their cooperation during the survey data and smallholders’ sources of income. A continuous collection process. We also wish to thank Mr. Azmach Tensaye soil cover with diversified living plants facilitates the for his help in translating questionnaire into local language and supervising enumerators during data collection process. capture and infiltration of rainwater and protects the soil, besides being a source of organic matter through Open Access This article is distributed under the terms of the litter accumulation (Mollison and Slay 1991). How- Creative Commons Attribution 4.0 International License (http:// creativecommons.org/licenses/by/4.0/), which permits unre- ever, in-depth research is required to quantify the stricted use, distribution, and reproduction in any medium, effects of intercropping of the khat-based systems on provided you give appropriate credit to the original productivity and of further integration of enset and author(s) and the source, provide a link to the Creative Com- crossbred animals. mons license, and indicate if changes were made. 123 1594 Agroforest Syst (2018) 92:1579–1595 References determinants of sustainability in smallholder farming sys- tems of Eastern Uganda. Popul Environ 31:474–506. doi:10.1007/s11111-010-0104-2 Abebe T (2005) Diversity in home garden agroforestry systems Falconnier GN, Descheemaeker K, Van Mourik T, Sanogo O, of Southern Ethiopia. Dissertation, Wageningen Univer- Giller K (2015) Understanding farm trajectories and sity, Wageningen development pathways: two decades of change in southern Abebe T (2013) Determinants of crop diversity and composition Mali. Agric Syst 139:210–222. doi:10.1016/j.agsy.2015. in Enet-coffee home gardens of southern Ethiopia. J Agr 07.005 Rural Dev Trop Subtrop 114(1):29–38 Fekadu D, Ledin I (1997) Weight and chemical composition of Abebe T, Wiersum K, Bongers F (2010) Spatial and temporal the plant parts of enset (Ensete ventricosum) and the intake variation in crop diversity in agroforestry home gardens of and degradability of enset by cattle. Livest Prod Sci southern Ethiopia. Agrofor Syst 78:309–322 49:249–257 Amede T, Diro M (2005) Optimizing soil fertility gradients in Feyisa T, Aune J (2003) Khat expansion in the Ethiopian the enset (Ensete ventricosum) systems of the Ethiopian highlands. Effects on the farming system in Habro District. highlands: trade-offs and local innovations. In: Bationo A, Mt Res Dev 23:185–189 Kimetu J, Kihara J (eds) Improving human welfare and Getahun A, Krikorian A (1973) Chat: coffee’s rival from Harar, environmental conservation by empowering farmers to Ethiopia, botany cultivation and use. Econ Bot 27:353–377 combat soil fertility degradation. Tropical Soils Biology Giller KE, Corbeels M, Nyamangara J, Triomphe B, Affholder and Fertility Institute of CIAT, Addis Ababa F, Scopel E, Tittonell P (2011) A research agenda to Amede T, Takele B, Endrias G (2001) Reversing the degrada- explore the role of conservation agriculture in African tion of arable land in Ethiopian highlands. Managing smallholder farming systems. Field Crops Res African Soils No 23 124(3):468–472. doi:10.1016/j.fcr.2011.04.010 ATA (2012) Report on results of crop yield sample survey. Herrero M, Thornton P, Notenbaert A, Msangi S, Wood S, Agricultural Transformation Agency (ATA), Addis Ababa Kruska R, Dixon J, Bossio DA, van de Steeg J, Freeman H, Blazy J, Ozier-Lafontaine H, Dore` T, Thomas A, Wery J (2009) Li X, Parthasarathy R (2009) Drivers of change in crop- A methodological framework that accounts for farm livestock systems and their potential impacts on agro- diversity in the prototyping of crop management systems. ecosystems services and human well-being to 2030. Application to banana-based systems in Guadeloupe. Agric CGIAR Systemwide Livestock Programme (SLP). ILRI, Syst 101:30–41. doi:10.1016/j.agsy.2009.02.004 Nairobi Brandt SA, Spring A, Hiebsch C, McCabe G, Taboge E, Diro M, Joe H, Ward J (1963) Hierarchical grouping to optimize an Wolde-Michael G, Ynitso G, Shigeta M, Tesfaye S (1997) objective function. J Am Stat Assoc. doi:10.1080/ The ‘tree against hunger’: Enset-based agricultural systems 01621459.1963.10500845 in Ethiopia. AAAS, Washington, DC Kandari L, Yadav H, Thakur A, Kandari T (2014) Chat (Catha CSA (1991–2013) The population estimates derived from pro- edulis): a socio-economic crop in Harar region, eastern jecting the 1994 and 2007 census, analytical report for Ethiopia. Springer Plus 3:579. doi:10.1186/2193-1801-3- Southern Nations, Nationalities and People’s Region. Central Statistical Agency (CSA), Addis Ababa Kippie T (2002) Five thousand years of sustainablity? A case CSA (2007) Population and housing census. Analytical report study on Gedeo land use (Southern Ethiopia). Dissertation, for the Southern Nations, Nationalities and People’s Wageningen University, Wageningen Region. Central Statistical Agency (CSA), Addis Ababa Kumar BM, Nair PKR (2004) The enigma of tropical home- Descheemaeker K, Ronner E, Ollenburger M, Franke AC, gardens. Agrofor Syst 61:135–152 Klapwijk CJ, Falconnier GN, Wichern J, Giller KE (2016) Kumar B, Nair P (eds) (2006) Tropical Homegardens: a time- Which options fit best? Operationalizing the socio-eco- tested example of sustainable agroforestry 3. Springer, logical niche concept. Exp Agric:1–22 Dordrecht Dessie G (2013) Current African issues: favouring a demonised Lemessa D (2001) Khat (Catha edulis): botany, distribution, plant khat and Ethiopian smallholder enterprise. No 51. cultivation, usage and economics in Ethiopia. UN-Emer- Nordic Africa Institute, Uppsala gencies Unit for Ethiopia, Addis Ababa Dessie G, Kinlund P (2008) Khat expansion and forest decline in Mekbib F (2009) Farmers as treasure chest of biodiversity: the Wondo Genet, Ethiopia. Geografiska Annaler Ser B Hum case of sorghum (Sorghum bicolor (L.) moench) in Ethio- Geogr 90(2):187–203 pia a five decade analysis. Maydica 54:11–23 Distefano J (1983) An enquiry into the history of Qat. Pro- Moll HAJ (2005) Costs and benefits of livestock systems and the ceedings of the 2nd international conference on Somali role of market and nonmarket relationships. Agric Econ Studies, Hamburg, p 1–6 32:181–193 Dorsey B (2015) Agricultural intensification, diversification, Mollison B, Slay R (1991) Introduction of permaculture. The and commercial production among smallholder coffee Tropical press, Harare growers in Central Kenya. Econ Geogr 75:178–195 NBE (1991–2014) Exchange rate report National bank of Dube D, Yadav R, Dube S (2014) Why a shift from coffee to Ethiopia. Addis Ababa chat? A study of the Kersa Woreda in Jimma Zone of South Rehima M, Belay K, Dawit A, Rashid S (2013) Factors affecting Western Ethiopia. Res Human Soc Sci 4:177–182 farmers’ crops diversification: evidence from SNNPR, Ebanyat P, de Ridder N, de Andre J, Delve R, Bekunda M, Giller Ethiopia. Int J Agric Sci 3:558–565 K (2010) Drivers of land use change and household 123 Agroforest Syst (2018) 92:1579–1595 1595 Samuel M (2014) Livestock feeds and feeding system in enset (2007) Coping strategies in livestock dependent house- (Ensete ventricosum) dominated mixed farming systems of holds in East and southern Africa: a synthesis of four case southern Ethiopia. J Anim Feed Res 4(6):150–158 studies. Hum Ecol 35:461–476. doi:10.1007/s10745-007- Shumbulo A, Gecho Y, Tora M (2012) Diversity, challenges and 9118-5 potentials of Enset (Ensete ventricosum) Production: in Timmer C (1997) Farmers and markets: the political economy of case of Offa Woreda, Wolaita Zone, southern Ethiopia. new paradigms. Am J Agr Econ 79:621–627 Food Sci Qual Manage 7:24–31 Tsegaye A, Struik P (2001) Enset (Ensete ventricosum (Welw.) Solomon Z, Masresha F, Olsson M (2008) Effect of drought/ Cheesman) kocho yield under different crop establishment irrigation on proximate composition and carbohydrate methods as compared to yields of other carbohydrate-rich content of two enset (Enset ventricosum (Welw.) (Chees- food crops. Neth J Agric Sci 49:81–94 man) clones. Ethiop J Sci Technol 31(2):81–88 Von Braun J (1995) Agricultural commercialization: impacts on Tamire C, Argaw M (2015) Role of Enset (Ensete ventricosum income and nutrition and implications for policy. Food (Welw.) Cheesman) in soil rehabilitation in different agro- Policy 20:187–202 ecological zones of Hadiya, Southern Ethiopia. Am J Env Wiggins S, Gem A-K, Jennifer L, Colin P (2011) Small farm Protect 4(6):285–291. doi:10.11648/j.ajep.20150406.14 commercialisation in Africa: reviewing the issues. Tefera TL (2009) Supply response, local reality and livelihood Research Paper, no 23. Brighton, UK sustainability: the policy dilemma of khat (Catha edulis) Woldu Z, Belew D, Benti T (2015) The coffee-khat interface in production in eastern Ethiopia. Int J Agr Sustain Eastern Ethiopia: a controversial land use and livelihood 7(3):176–188. doi:10.3763/ijas.2009.0428 change scenario. J Agri Sci Technol 5:149–169. doi:10. Thornton P, Boone R, Galvin K, Bunsilver S, Waithaka M, 17265/2161-6264/2015.03.001 Kuyiah J, Karanja S, Gonzalez-Estrada E, Herrero M http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Agroforestry Systems Springer Journals

Home garden system dynamics in Southern Ethiopia

Download PDF
17 pages

Loading next page...
 
/lp/springer-journals/home-garden-system-dynamics-in-southern-ethiopia-VY5wVI8y2s

References (45)

Publisher
Springer Journals
Copyright
2017 The Author(s)
ISSN
0167-4366
eISSN
1572-9680
DOI
10.1007/s10457-017-0106-5
Publisher site
See Article on Publisher Site

Abstract

Agroforest Syst (2018) 92:1579–1595 https://doi.org/10.1007/s10457-017-0106-5(0123456789().,-volV)(0123456789().,-volV) . . . Beyene Teklu Mellisse Gerrie W. J. van de Ven Ken E. Giller Katrien Descheemaeker Received: 28 June 2016 / Accepted: 18 July 2017 / Published online: 31 July 2017 The Author(s) 2017. This article is an open access publication Abstract Home gardens in southern Ethiopia are cereal-vegetable systems. In densely populated, mar- regarded as efficient farming systems, allowing inter- ket proximate areas a major trend was expansion of actions and synergies between crop, tree and livestock khat, from 6 to 35% of the area share per farm, while components. However, these age-old traditional home the combined area share of enset and coffee decreased gardens are evolving rapidly in response to changes in from 45 to 25%. Concurrently, the cattle herd size fell both the socio-economic and biophysical environ- from 5.8 TLU to 3.9 TLU per household. In medium ment. Altered cropping patterns, farm size and com- populated, less accessible areas the trend was consol- ponent interactions may affect the systems’ idation of combined production of food and cash sustainability. Home gardens exhibit a huge diversity crops. Enset and coffee together maintained a share of in farms and farming systems, which needs to be over 45%. Easy transport and marketing of the understood in order to design interventions for perishable cash-generating khat compared with tradi- improvement. Dynamics of home gardens were stud- tional crops favoured its cultivation among small- ied over two-decades (1991–2013) based on a survey holders located close to markets. The insights in home of 240 farm households and focus group discussions. garden change in response to increasing population Farms were grouped into five types: Khat-based, pressure, decreasing farm size and market develop- Enset-cereal-vegetable, Enset-based, Enset-coffee and ment may help to design interventions to increase Enset-livestock. Farm trajectories revealed a shift system sustainability. from food-oriented Enset-based and Enset-livestock systems to (1) cash crop oriented khat-based systems, Keywords Home garden types  Cropping patterns and (2) combined food and cash crop oriented Enset- Cash crops  Commercialization  Diversification Catha edulis B. T. Mellisse (&)  G. W. J. van de Ven K. E. Giller  K. Descheemaeker Plant Production Systems, Wageningen University and Introduction Research (WUR), PO Box 430, Wageningen, The Netherlands Home gardens are characterized by multipurpose trees e-mail: beyene.mellisse@wur.nl; beyteklu@gmail.com and shrubs in intimate association with annual and perennial crops and livestock around the homestead B. T. Mellisse (Kumar and Nair 2006). Home gardens ensure a Wondo Genet College of Forestry and Natural Resources, sustained availability of multiple products and Hawassa University, P.O.Box 128, Shashemene, Ethiopia 123 1580 Agroforest Syst (2018) 92:1579–1595 generate income (Kumar and Nair 2004). Despite their the traditional land use systems (Abebe et al. 2010; small size (Kumar and Nair 2006), home gardens fulfil Dessie and Kinlund 2008). Yet, little is known about most of the basic food and nutritional needs of the the rate at which those systems have shifted to mono- households, while the multi-storied configuration and cropping systems, how that differs across the region, high species diversity maintain their structure and and which factors could explain it. Other studies function in the face of external stress (Kumar and Nair (Herrero et al. 2009; Ebanyat et al. 2010) suggest 2004). trends in institutional support, resource endowment, In the Southern Nations Nationalities and Peoples’ prices, social conditions and technology as drivers of Regional State (SNNPRS) of Ethiopia, home gardens farming system change, but these have not yet been are the prevalent land use system covering about investigated for the particular case of home gardens in 576,000 ha, which is 31% of the region’s cultivable southern Ethiopia. land (Abebe 2005). The traditional ‘enset-coffee’ Similar to many other areas in sub-Saharan Africa, home gardens are characterised by the combination of SNNPRS is characterized by a huge diversity in farms two perennial crops: enset (Enset ventricosum and farming systems, even though they are grouped (Welw.) Cheesman) and coffee (Coffea arabica L.). under the common term home gardens. Recognizing Enset is a herbaceous, multipurpose crop and a staple this variability within and among farming systems and food for more than 15 million people (Abebe et al. localities is the first step in the design of new technologies to improve agricultural production (Gil- 2010). Enset leaves are a key livestock feed and used as mulch to reduce soil erosion and runoff (Amede and ler et al. 2011; Descheemaeker et al. 2016). Diro 2005). Hence, enset fulfils both productive and With the wider aim to contribute to the develop- protective functions (Kippie 2002). Coffee is mainly a ment of tailored interventions to improve agricultural cash crop, but is also consumed (Kippie 2002; Abebe production and productivity, this paper aims to: (1) 2005). understand the diversity in home garden systems of Increasing population pressure and subdivision of southern Ethiopia; (2) analyse trends in farm size, farms have led to fragmentation of land, a decline in cropping pattern and livestock population over the the area under coffee and enset (Tsegaye and Struik past two decades; (3) explain the spatial patterns of 2001; Abebe 2005) and gradual replacement of the change in home gardens in relation to market access, age-old diverse system. A major change is the agro-ecological conditions and population density. expansion of a mono-cropping system of khat (Catha We combined quantitative and qualitative methods, edulis Forsk) at the expense of enset-coffee home including secondary data collection on demography gardens (Tsegaye and Struik 2001; Abebe et al. 2010). and farm size, a farm typology based on the area share Khat is grown for its economically important leaves of crops and grazing land, and constructed a 20-year and tender twigs, which are chewed for their stimu- timeline of farming system change based on detailed lating effect. Due to the decline in enset cultivation, household interviews concerning current and past many households have become more dependent on the farm assets and practices. food market (Amede and Diro 2005). Market depen- dency for family food is further exacerbated by decreasing farm area and productivity of food crops Materials and methods (Amede et al. 2001). In countries like Ethiopia, where smallholder farmers have no access to insurance, the The study area market dependency increases the vulnerability to economic or environmental shocks. The replacement Sidama and Gedeo are representative zones of the of enset has also induced a shortage of livestock feed SNNPRS region with a prevalence of enset-coffee with direct repercussions on herd size, herd compo- home gardens. These home gardens are characterised sition, livestock production and hence, nutritional by the production of enset, coffee and multi-purpose quality of human diets (Tsegaye and Struik 2001). trees, accompanied by root and tuber crops, vegeta- Most studies on the home gardens of southern Ethiopia bles, annual cereal crops and livestock keeping claimed that expansion of khat has resulted in (Kippie 2002; Abebe et al. 2010). Sidama is located 0 0 0 0 homogenization of the structure and composition of within 545 –645 N and 385 –3941 E, covering a 123 Agroforest Syst (2018) 92:1579–1595 1581 total area of 7672 km (Abebe 2005) with 3.50 million 2010) and (3) three classes of population density, i.e. inhabitants (CSA 2007). Gedeo is located within high ([800 person/km ), medium (500–800 person/ 2 2 0 0 0 0 550 –612 N, 3803 –3818 E, covering a total area km ) and low (\500 person/km ). of 1347 km (Kippie 2002) with 0.84 million inhab- On average 10 households per PA, 240 households itants (CSA 2007). Most kebeles (smallest Ethiopian in total, were selected for the household survey. The administrative unit, also called peasant association or districts’ Bureau of Agriculture categorized farms as PA) in both zones are classified as ‘rural’ (CSA 2007). small, medium and large based on the area of their land The rainfall distribution is bimodal with a long (June holding and the number of livestock. Farms with to September) and short (March to May) rainy season. [1.25 ha of land and [5 TLU were categorized as Both Sidama and Gedeo straddle two agro-ecological large, 0.5–1.25 ha and 1–5 TLU as medium, and zones, the moist mid-altitude (In Amharic: woina- \0.5 ha and 0–1 TLU as small. Stratified random dega) and the moist highland (In Amharic: dega). The sampling ensured that the number of selected house- moist mid-altitude zone ranges in elevation from 1500 holds in each group was in line with their share in each to 2300 m a.s.l. and receives 1200–1600 mm rainfall of the PAs. annually; the average annual temperature ranges from 16 to 22 C. The moist highland zone comprises an Data collection elevation range of 2300–3200 m a.s.l.; mean annual rainfall amounts to 1600–2000 mm and the average We combined qualitative and quantitative data col- annual temperature ranges from 15 to 19 C. The soils lection approaches to triangulate the relevance and in the mid-altitude and highland agro-ecologies are accuracy of the information gathered (Ebanyat et al. mainly characterized by clay-loam to silt-loam and 2010). Eight focus group discussions (FGDs) were sandy-loam to sandy-clay textural classes organized, four in midland PAs (Ywo, Kochow, respectively. Soyama and Dero) and four in highland PAs (Gikeatoye, Guguma, Duba and Elalcha). In each Selection of study sites and farm households FGD eight to ten farmers participated representing different age groups, both sexes and each farm size Within Sidama and Gedeo zones, four districts (In category. The discussion focused on positive and Amharic: woreda) were selected (Wondo Genet, negative aspects of khat expansion, constraints to Melga, Dale and Bule) and within each district, 6 maintain enset and coffee, population growth and its PAs, giving 24 PAs in total (Fig. 1). The multistage impact on resource availability (land), change in sampling approach for districts and PAs was designed income sources, emergence of new activities, eco- to encompass differences in distance to markets, agro- nomic infrastructure (roads and transport service), ecological conditions and population density market development and price trends. (Table 1). These factors were hypothesised to affect A detailed household survey was administered farmers’ crop allocation decisions through their influ- between August and November 2013 with 240 farm ence on the ease of transportation of agricultural households. We used a semi-structured questionnaire products, crop suitability and farm size, respectively. to collect data on changes, trends and events in the past Distance from a major market and altitude were two decades with the aim to construct a timeline. To measured using a global positioning system (GPS) capture the changes in cropping and livestock systems, following roads and foot paths. The population density household heads were asked to recall the year of khat of each PA was collected from the Ethiopian Central introduction, land allocation to various annual and Statistics Agency (CSA 1991–2013). The PAs were perennial crops (food crops, cash crops), total land categorized into (1) three classes of distance to market, holding and herd size from 1991 to 2013 in a stepwise i.e. near (\36 km), medium (36–70 km) and far fashion. The year of the regime change in Ethiopia, ([70 km), roughly corresponding to travelling times 1991, and the following national election years (1994, of 1.5, 3 and more than 3 h by car, respectively, (2) 2000, 2005 and 2010) were used as benchmarks to two elevation classes, based on agro-ecological zon- help respondents’ recall. Household heads were fur- ing, i.e. high ([2300 m a.s.l.) and medium ther asked about demographic characteristics (family (1500–2300 m a.s.l.) (Kippie 2002; Abebe et al. size, level of education), production objectives, 123 1582 Agroforest Syst (2018) 92:1579–1595 Table 1 Peasant associations (PA), numbers of surveyed households in each of the three farm size categories, distance to market, elevation and population density. Source: Ethiopian Central Statistical Authority (CSA 1991–2013) No. Site (PA) District Sample Sampled farms by farm size Distance to Elevation Population (Woreda) size major market (m a.s.l.) density Large Medium Small 2 place (km) (person/km ) Mid-land agro-ecology (1500–2300 m a.s.l.) 1 Ywo Wondo Genet 11 1 4 6 N 12.4 1679 H 958 2 Chuko Wondo Genet 12 1 3 8 N 26.4 1756 H 1551 3 Weterakechema Wondo Genet 7 1 2 4 M 36.4 1755 M 714 4 Kochow Melga 10 2 4 4 N 24.4 2145 H 834 5 Manicho Melga 11 1 8 2 N 25.4 2190 H 858 6 Berana Melga 9 1 6 2 N 30.9 2216 H 915 7 Soyama Dale 9 1 6 2 M 36.3 1787 H 852 8 Megera Dale 10 1 4 5 M 47.8 1746 M 566 9 Awada Dale 11 2 2 7 M 44.7 1753 L 433 10 Dero Bule 11 1 4 6 F 126.8 2245 M 792 11 Bassura Bule 11 3 3 5 F 114.8 2253 M 773 12 Osselemajo Bule 8 2 2 4 F 118.5 2250 M 737 Sub total 120 17 (14%) 48 (40%) 55 (46%) Highland agro-ecology (2300–3200 m a.s.l.) 13 Weteragedo Wondo Genet 8 3 3 2 M 38.1 2597 H 945 14 BaboChorora Wondo Genet 9 2 4 3 M 40.5 2624 M 646 15 Gikeatoye Wondo Genet 13 3 5 5 M 44.9 2599 H 1565 16 Fitoketemuna Melga 12 4 4 4 M 41.1 2561 H 950 17 Gerewe Melga 9 3 4 2 N 34.6 2634 M 508 18 Guguma Melga 9 3 2 4 M 47.1 2677 M 583 19 Hayello Dale 8 2 2 4 F 70.2 2328 H 1023 20 Duba Dale 10 1 2 7 M 47.4 2338 H 1153 21 Gedamo Dale 12 1 3 8 M 64.7 2339 L 102 22 Elalcha Bule 11 3 4 4 F 126.9 2902 L 469 23 Akollo 1st Bule 11 2 4 5 F 142.9 2787 L 272 24 Suko Bule 8 2 2 4 F 135.1 2709 L 43 Sub-total 120 29 (24%) 39 (33%) 52 (43%) Total 240 46 (19%) 87 (36%) 107 (45%) N near, M medium, F far, H high, M medium, L low sources of income, constraints to crop production and f  j ij j¼1 M ¼ livestock rearing and dependency on the market for i food (number of months per year). Farmers also listed where M = mean score of factor i; f = the percent- i ij and ranked political, cultural, environmental and age of respondents that gave a score of j for factor economic factors triggering the changes in land i and n = the total number of respondents that listed allocation and farming practices based on a score of factor i. one (low importance), two (average importance) or Enumerators with a certificate in agriculture and three (high importance). The mean score of each factor speaking the local language were recruited from the was calculated using the formula 123 Agroforest Syst (2018) 92:1579–1595 1583 localities and trained in data collection. Retail prices of southern Ethiopia. Five different home garden types coffee, khat, kocho (the processed edible part of enset), were identified and characterized by the area share of tef (Eragrostis tef (Zucc.) Trotter) and maize (Zea the dominant crops and grazing land (Table 2; Fig. 2). mays L.) were collected from CSA. The price in Our decision tree (Fig. 2) dictated that throughout the Ethiopian birr was converted to US$ at a rate of 2.5, study period, farms with an area share of khat over 6.3, 8.6, 17.5 and 19.1 Eth. birr to one US$, the average 30% were classified as Khat-based systems. With less exchange rate in 1991–1994, 1995–1999, 2000–2008 than 30% khat, home gardens were described as Enset- and 2009–2012 and 2013–2014 respectively (NBE coffee if at least 35% of their area was covered with 1991–2014). The revenue per hectare was calculated coffee. Subsequently, Enset-livestock systems were by multiplying the price of each crop with the average identified if grazing land covered more than 20%. Of yield (assumed to be constant over the study period). the remaining two types, Enset-based and Enset- cereal-vegetable systems were distinguished based on Farm typology an enset area share of at least and less than 35% enset, respectively. We constructed a farm typology based on crop In Khat-based and Enset-based home gardens khat allocation data from the household survey after and enset covered 51 and 58% of the farm land exploring the data using Ward’s minimum variance respectively (Table 2). The combined area of enset, method (Joe and Ward 1963), followed by a K-means cereals and vegetables covered more than 60% of the clustering (Blazy et al. 2009). The area shares of nine area in the Enset-cereal-vegetable system, while 89% crops (enset, coffee, khat, maize, other annual cereals, of the area in Enset-coffee home gardens was covered beans, root and tuber crops, vegetables and sugarcane) by enset and coffee. In Enset-livestock home gardens and grazing land were considered. Each of the the combined area of enset and grazing land covered identified farm types was named after the crops with more than 50% of the farm area. The highest the major area shares in 2013. In order to classify farms proportion of cash crops was observed in Khat-based based on the same criteria over time, thresholds of area home gardens (60%) while the highest proportion of shares were derived based on the variation within each food crops was observed in Enset-based home gardens farm type, following Falconnier et al. (2015). Consid- (74%). The largest proportion of grazing land was ering these thresholds, we developed a decision tree to documented in Enset-livestock systems (34%). The classify each farm into a distinct farm type for the six average family size in Khat-based home gardens was benchmarks in the timeline 1991–2013. significantly larger than that in Enset-based and Enset- livestock home gardens (Table 2). The average farm Data analysis size showed no significant difference between home garden types, and the Khat-based home gardens had A one way analysis of variance (ANOVA) was used to the smallest land to labour ratio. test the difference between the means of five home garden types in family size, farm size, land to person Distribution of home garden types across districts ratio, herd size, income and market dependency. and socio-economic and biophysical environments Differences were deemed significant at P \ 0.05. Analyses were performed using the Statistical Package The four districts of the study area showed a different prevalence of the five home garden types. Wondo for Social Sciences (SPSS) version 20. Qualitative data was used to support the discussion of results. Genet and Melga were characterized by the domi- nance of two home garden types, namely Khat-based and Enset-cereal-vegetable in Wondo Genet and Results Khat-based and Enset-livestock in Melga. In both districts Enset-coffee systems were absent. Dale and Characteristics of the home garden types Bule showed a more even representation of four home garden types. The Khat-based home garden The variability in land use indicated a large diversity type was not present in Bule and nearly absent in in farming systems in Sidama and Gedeo zones of Dale. 123 1584 Agroforest Syst (2018) 92:1579–1595 Fig. 1 Map of the study districts (woredas) within Sidama and Gedeo zones, SNNPRS, Ethiopia Thepresenceofthe homegardentypesdiffered Dynamics in cropping pattern and livestock holding in the four districts in relation to market access, population density and agro-ecology (Fig. 3). Close to markets, 56% of the The overall trend in cropping patterns in Wondo Genet farms were Khat-based (Fig. 3a). At medium and far distance their share was 17 and 0% respectively. and Melga indicated a shift from food to cash crop production over time, while in Dale farmers continued The shares of Enset-cereal-vegetable, Enset-based and Enset-coffee home garden types all decreased to produce a combination of food and cash crops and in with proximity to the market. The distribution of Bule they remained food crop oriented (Fig. 4). In home garden types in relation to population density 1991 the combined area share of enset and coffee showed a similar pattern (Fig. 3b), with Khat-based amounted to more than 45% in all districts of the study systems gaining importance with increasing popu- area. In 2013 it had fallen below 25% in Wondo Genet lationdensity andanopposite trend for Enset- and Melga, whereas the area share of khat increased from about 5% in 1991 to about 35% in 2013 (Fig. 4). cereal-vegetable, Enset-based and Enset-coffee home garden types. Population density had no clear Even though khat cultivation was observed in 1991 also in Dale, it expanded less there from 0.9% in 1991 effect on the presence of Enset-livestock home gardens. In the mid-altitude zone Khat-based and to 8% in 2013. In Bule, khat has not played any role worth mentioning in the past two decades (Fig. 4). Enset-coffee home gardens were more prevalent, while in the high-altitude zone Enset-based and Unlike the cropping pattern, the grazing land area was Enset-livestock systems were more common not noticeably influenced by the expansion of khat and (Fig. 3c). was maintained at 8–12% in Wondo Genet and 123 Agroforest Syst (2018) 92:1579–1595 1585 Table 2 Average (±standard error of the mean) crop alloca- P \ 0.05 are indicated with different letters within a row; tion and household characteristics for the five home garden dominant crops and grazing land that lend their name to the types; significant differences in household characteristics at type are indicated in bold Description Home garden types Khat-based Enset-cereal- Enset-based Enset-coffee Enset- (K-b) vegetable (E- (E-b) (E-c) livestock c-v) (E-l) Number of farms 58 51 43 30 58 Share in the farm area (%) Food crops Enset 26.9 ± 1.7 18.8 – 1.3 57.6 – 2.2 39.9 – 2.2 21.7 – 1.4 Annual cereals 5.9 ± 1.4 24.3 – 1.9 11.6 ± 2.2 1.8 ± 1.0 13.5 ± 1.5 Other food crops 0.4 ± 0.2 11.8 ± 1.3 4.8 ± 1.6 2.3 ± 1.1 7.1 ± 1.5 Total food crops 33.2 ± 2.1 54.9 ± 2.5 74.0 ± 2.6 44.0 ± 2.4 42.3 ± 2.2 Cash crops Coffee 3.0 ± 0.6 5.6 ± 1.3 10.8 ± 1.8 49.4 – 2.0 3.4 ± 1.0 Khat 51.2 – 2.0 11.0 ± 1.7 4.1 ± 1.4 2.0 ± 1.3 14.7 ± 1.6 Vegetables 3.8 ± 1.0 18.7 – 2.2 5.8 ± 1.8 1.4 ± 0.8 4.8 ± 1.0 Other cash crops 1.7 ± 0.8 3.4 ± 1.1 1.8 ± 0.9 2.1 ± 0.8 0.8 ± 0.2 Total cash crops 59.7 ± 2.2 37.8 ± 2.6 21.8 ± 2.6 54.9 ± 2.5 23.7 ± 1.8 Grazing land 7.1 ± 1.2 7.3 ± 1.1 4.2 ± 1.3 1.1 ± 0.5 34.0 – 1.7 Household characteristics Family size (#) 9.0 ± 0.5a 8.8 ± 0.4a 7.0 ± 0.5b 8.8 ± 0.8ab 7.3 ± 0.5b Farm size (ha) 0.8 ± 0.1a 0.8 ± 0.1a 1.0 ± 0.1a 0.80 ± 0.2a 1.0 ± 0.1a Land: labour ratio 0.13 ± 0.0b 0.18 ± 0.0ab 0.27 ± 0.1a 0.25 ± 0.0a 0.25 ± 0.0a Cattle (TLU) 3.8 ± 0.4a 3.8 ± 0.4a 2.8 ± 0.5b 1.6 ± 0.2b 4.5 ± 0.4a Small ruminant 0.4 ± 0.0b 0.7 ± 0.1a 1.0 ± 0.2a 1.5 ± 0.3a 1.0 ± 0.1a (TLU) Income (10 Eth. 28.0 ± 8.0a 18.0 ± 2.0ab 7.0 ± 1.0c 12.0 ± 3.0b 16.0 ± 3.0b Birr) Market dependency 5.3 ± 0.4a 2.9 ± 0.4b 3.9 ± 0.5a 3.1. ± 0.6ab 3.2 ± 0.4b for food (months) Number of household heads Illiterate 10 (17%) 7 (14%) 13 (30%) 2 (7%) 21 (36%) according to educational level Write and read 7 (12%) 3 (6%) 7 (16%) 5 (16%) 14 (24%) Elementary 33 (57%) 27 (53%) 17 (40%) 17 (57%) 20 (34%) Secondary 8 (14%) 14 (27%) 6 (14%) 6 (20%) 3 (5%) TLU tropical livestock unit Other food crops include: beans, root and tuber crops Other cash crops include: sugarcane 20–23% in Melga for the last two decades. Grazing Bule from 4.8 to 3.4 TLU over the same time. In land area remained below 10% in Dale, and in Bule it Wondo Genet the cattle holding steadily decreased varied between 8 and 17%, with a share of 13% in from 5.8 TLU in 1991 to 3.9 TLU in 2013. The number 2013. of small ruminants varied between districts, but In the period 1991–2013 the cattle and small changed little over time. Only in Wondo Genet it ruminant population dynamics varied among the four tended to decline from 0.9 TLU in 1991 to 0.5 TLU in districts. In three districts the cattle holding decreased 2013, but also here it varied between years. In Melga and only in Melga it increased from 5.1 TLU in the number of small ruminants per farm represented 1994–2000 to 5.9 TLU in 2005–2013. In Dale the 0.5 TLU, in Dale it varied between 0.1 and 0.3 and in cattle holding decreased from 2.7 to 2.1 TLU and in Bule between 1.0 and 1.3 TLU. 123 1586 Agroforest Syst (2018) 92:1579–1595 Fig. 2 Decision tree and threshold values for area share used for farm classification in five home garden types. Number of farms (n) refers to the year 1991 Fig. 3 Share of home garden types in 2013 classified according to a market access, b population density and c elevation 123 Agroforest Syst (2018) 92:1579–1595 1587 Dale Wondo Genet 1991 1994 2000 2005 2010 2013 1991 1994 2000 2005 2010 2013 Bule Melga 80 80 20 20 1991 1994 2000 2005 2010 2013 1991 1994 2000 2005 2010 2013 Fig. 4 Land use dynamics in the four districts of the study area in the period 1991–2013 Dynamics in the prevalence of home garden types half of the Enset-livestock home gardens. Also in in the period 1991–2013 Melga the major emerging type was the Khat-based home garden, increasing from 3 to 39% of the farms at In 1991, about 30% of the sampled farms were the expense of Enset-based and Enset-coffee home engaged in Enset-based and 30% in Enset-livestock gardens, which fully disappeared by 2013. farming, while only 3% was engaged in Khat-based The farm transitions were different in Dale and farming (Fig. 5). The proportion of Enset-livestock Bule districts. In Dale the emergence of the Khat- and Enset-based home gardens decreased to 24% and based home garden type was restricted to only 3% of less than 20% of the farms respectively in 2013 the farms, but an increase in the proportion of the (Fig. 5). The share of Enset-coffee home gardens Enset-coffee system was noticed (Fig. 6). In Bule the declined from 19% in 1991 to 13% in 2013. In Enset-based and Enset-cereal-vegetable home gardens contrast, the share of khat-based farms increased from gained some importance at the expense of the Enset- 3% in 1991 to 24% in 2013. The percentage of farms in livestock and Enset-coffee home gardens (Fig. 6). the Enset-cereal-vegetable system slightly increased from 17% in 1991 to 21% in 2013 (Fig. 5). Determinants of farming system dynamics The transitions in home garden systems varied across the four districts of the study area (Fig. 6). In Respondents identified population pressure, policy, Wondo Genet most Enset-based home gardens in 1991 infrastructure, profitability, decline in soil fertility, transitioned to Khat-based or Enset-cereal-veg- cultural change, lack of input, labour availability and etable home gardens in 2013, and the proportion of harvesting frequency of khat as major determinants of Enset-based farms dropped from 44% in 1991 to 7% in land use change in the area. Among these nine factors, 2013. The farms practicing Enset-coffee farming in population pressure was ranked as the major driver of 1991 have all transitioned to either Khat-based or land use change. In all four districts population Enset-cereal-vegetable home gardens in 2013, as have pressure increased (Fig. 7a) and contributed to the Area share (%) 1588 Agroforest Syst (2018) 92:1579–1595 Fig. 5 Proportion of each farm type in 1991 and 2013 and their transition during this period in Sidama and Gedeo zones (n = 240 farms). The stacked bar in the middle represents the transition of farms from a given type in 1991 to the types in 2013; the height of each stacked cell and bar is representative of the number of farms Fig. 6 Changes in the 1991 khat-based prevalence of home garden 10% 10% 20% 7% 8% types in the four districts of Enset-cereal-vegetable the study area over the Wondo Enset-based 55% period 1991–2013 18% Genet 28% Enset-coffee 44% Enset-livestock 3% 14% 39% 45% Melga 50% 30% 8% 3% 8% 20% 0% 20% 20% 20% Dale 27% 27% 33% 33% 20% 27% 33% 33% Bule 17% 17% 30% 23% 123 Agroforest Syst (2018) 92:1579–1595 1589 Fig. 7 Population density (a) and change in land holding per household (b)in the four districts between 1991 and 2013 1991 1994 2000 2005 2010 2013 1992/94 1995/00 2001/05 2006/10 2011/13 0.0 -1.0 -2.0 -3.0 -4.0 -5.0 -6.0 decline in farm size and land fragmentation. In Wondo The price of staple foods (maize, kocho and teff) Genet where population density doubled between rose gradually over the two decades (Fig. 8a). Among 1991 and 2013 the rate of decline in land holding per the food crops, the teff price showed the largest household increased from 0.5% per year in the period increase, from 0.4 US$ per kg in 1990/1 to 0.8 US$ in 1992/94 to 5.1% per year in the period 2011/13 2013/14. The price of maize was very close to that of (Fig. 7). In Melga, Dale and Bule districts, where the kocho, increasing from about 0.2 to 0.3 US$ per kg increase in population density was less pronounced, over the whole period. The price of khat fluctuated the rate of decline in land holding did not exceed 3%. around 3.5 US$ per kg of fresh weight in 1990/1 to A market liberalization policy implemented in 1991 2013/14, and for coffee, it fluctuated between 1.3 and and constitutionalized in 1995, which eliminated crop 4.3 US$ per kg in the same period (Fig. 8a). and livestock sale quota and price controls, was ranked Besides the price fluctuation, the greater income as a second determinant. As a third factor farmers per hectare from khat may have further discouraged mentioned the cultivation of khat, which became production of coffee. The average revenue per popular thanks to its frequent harvesting and high hectare from khat by far exceeded the average return per unit area. According to the respondents, the revenues from food crops (maize, teff and kocho) shift to khat cultivation was further fuelled by the (Fig. 8b). As a result of the continuous fragmenta- development of infrastructure (road, mobile phone) tion of land, the inability to produce sufficient and urban expansion. amount of family food also discouraged smallhold- Declining soil fertility coupled with a lack of farm ers to allocate land to food crops. Based on the inputs (seed, fertilizer and manure) also influenced the assumption that 50% of the energy demand is land use dynamics of the study area. Farmers further provided by enset and 50% by maize, the area referred to cultural change, and in particular to required to produce year-round food for an average changing diets. Labour availability was ranked last household of three adults and five children was in terms of its influence on land use change. 0.46 ha. A yield of 6.2 ton per ha for enset and Change in land Population density holding/household/year (%) (Person/km2) 1590 Agroforest Syst (2018) 92:1579–1595 Years Fig. 8 Annual retail prices of food and cash crops (a) and the average maize, teff and coffee yield per ha summarized from revenue per hectare (b) over the period 1991–2014. (Source: the CSA statistical bulletin (2000–2013); average yield for Retailer prices calculated and summarized from price lists for kocho and khat per ha from ATA (2012) the period 1991–2014 from CSA and Regional Custom Offices; 2.3 ton per ha for maize were considered (ATA where khat was introduced and expanded over the last 2012). If these households shift to khat and buy all two decades (1991-2013). This trend was observed in their food, they need only 0.05 ha, assuming a khat particular in Wondo Genet and Melga (Fig. 4) and revenue of 9368 US$ per ha (Fig. 8b) and a price simplified the cropping pattern with khat monoculture 0.3 US$ per kg for enset (kocho) and maize now occupying more than one-third of the farm area. (Fig. 8a). A similar expansion of khat was observed in the eastern highlands of Ethiopia, where it replaced sorghum (Feyisa and Aune 2003; Mekbib 2009) and Discussion in the Jima zone of southern Ethiopia where it replaced coffee (Dube et al. 2014). Expansion of khat cultivation Khat’s popularity with smallholders is related to its profitability, its regular harvesting (two to three times Cropping patterns in Sidama and Gedeo zones of per year) and its quick establishment (one to two southern Ethiopia changed most strongly in areas years). On the one hand, the process of -1 Revenue per hectare (thousand US$) Price (US$ kg ) 1990/1 1990/1 1992/3 1992/3 1994/5 1994/5 1996/7 1996/7 1998/9 1998/9 2000/01 2000/01 2002/03 2002/03 2004/05 2004/05 2006/07 2006/07 2008/09 2008/09 2010/11 2010/11 2012/13 2012/13 Agroforest Syst (2018) 92:1579–1595 1591 commercialization and specialization in cash crop Similar diversification with French beans or tomatoes monoculture may expose smallholder farmers to was reported among smallholder coffee growers in unpredictable market fluctuations, disease outbreaks central Kenya (Dorsey 2015). The degree of crop- and other adverse shocks, with a risk to become food livestock integration varied between farming systems, aid dependent (Amede and Diro 2005). Farmers are and the prevalence of competitive versus complemen- particularly vulnerable if institutions, policies, and tary roles of crops and livestock partly explains markets are non-supportive (Von Braun 1995), as in dynamics in livestock herd sizes. The advantage of the case of the study area, where e.g. crop insurance combining crop and livestock activities in the tradi- schemes are non-existent. Similarly, other studies tional systems was related to the availability of cheap reported the incapability of small commercialized fodder for livestock, such as enset leaves, and of farms to continue cash crop cultivation following a animal manure, contributing to enset and coffee decline in output prices, after which they reverted back growth. In the traditional home gardens, livestock to subsistence cropping (Wiggins et al. 2011). On the also played a critical role in providing a protein and other hand, the increased income from cash crops may nutrient rich diet to the household, given the low enable smallholders to buy food crops rather than nutritious value of enset (Brandt et al. 1997; Tsegaye having to produce everything themselves (Timmer and Struik 2001). The transition towards khat cultiva- 1997). tion reduced herd sizes due to decreased fodder Khat’s popularity with smallholders is further availability, especially from enset, and also lowered driven by disadvantages of the traditional crops, such the interest in livestock rearing as a source of cash and as the long time before enset plants can be harvested manure. For example, in Wondo Genet the khat (3–4 years after transplanting) and the unreliable price expansion went hand in hand with a decline in small and long establishment period of coffee. Thus, in the ruminant numbers. If the demand for cash is met by context of the study area, the incorporation of khat in khat, the need for keeping small ruminants recedes. the production system may be viewed as an innovation Moreover, khat is typically fertilized with mineral to take advantage of market opportunities towards fertilizer, reducing the need for animal manure. In achieving food security and improving household addition, the feeding behaviour of small ruminants, income (Rehima et al. 2013). browsing leaves and twigs, is incompatible with khat, as leaves and twigs are the economic product. Farming system transitions Similarly, less interest in keeping draught animals was reported where khat, which does not require The historical analysis of farming system transitions ploughing, replaced annual cropping systems in east- indicated a shift away from the traditional home ern Ethiopia (Kandari et al. 2014). Contrastingly, gardens based on enset, coffee and livestock (Fig. 5), small ruminant herd sizes did not decrease in Dale and illustrating the inability of these systems to accom- Bule, where the share of enset and coffee remained modate the increasing population. Whereas the tradi- fairly constant (Fig. 4). This could be related to the tional systems are known to support very dense availability of browse tree species for goats and -2 populations of up to 500 persons km (Kippie understory grazing for sheep in coffee fields. 2002; Abebe 2005), the population density of all The decrease in cattle holding per household in the study districts, except Bule, has surpassed this density khat-dominated district of Wondo Genet occurred together with the replacement of enset (Fig. 4). Enset since 2000 (Fig. 7). In areas far away from major roads and markets leaves have a high protein content (Brandt et al. 1997; farmers were less inclined to shift to khat, because of Tsegaye and Struik 2001; Solomon et al. 2008) and a less stringent land constraints on the one hand and feeding value that is comparable to a good quality difficulties related to the marketing of the perishable grass (Fekadu and Ledin 1997). Similarly, Kandari khat twigs on the other hand. Hence, farms transi- et al. (2014) reported a decline in per capita herd size tioned to Enset-based systems or modified the tradi- following complete replacement of major fodder crops tional systems by incorporating annual crops like by khat in Harar region of eastern Ethiopia. In Melga cereals and vegetables, leading to the expansion of we observed the opposite trend of a slight increase in Enset-cereal-vegetable systems in Bule (Fig. 6). cattle ownership, which could be related to the 123 1592 Agroforest Syst (2018) 92:1579–1595 absence of feed shortage in this district, allowing was present. The district is far from the market farmers to expand their herd size. This hypothesis is ([100 km), population density in 2013 was medium -2 supported by the relatively large grazing land area per (750 people km ) in the mid-altitude zone and low in -2 household and the fact that smallholders move their the highlands (250 people km ). Furthermore, the cattle to nearby lowland areas during the rainy season highland altitude of 2700–2900 m a.s.l. is above the and conserve their own pasture land for dry season suitable range for khat (Lemessa 2001). Melga closely grazing (focus group discussion). The smallest herd resembled the situation in Wondo Genet, with the sizes were observed in Dale, where the area of grazing difference that population growth and also land land was particularly small. Indeed, livestock feed fragmentation where somewhat lower. As a result, scarcity due to lack of grazing land is often reported as the area share of khat-based systems and the area the major constraint of livestock production also in under cash crops were smaller than in Wondo Genet in other districts of Sidama zone (Samuel 2014). 2013. Dale also took an intermediate position. It is at medium distance from the market, leading to a risk of The main drivers of changes in land use and farm quality loss for khat during transport (Distefano 1983). type transition It had a medium population density, similar to Melga, and the altitude did not exceed 2350 m a.s.l., posing no The main land use change, the replacement of Enset- agro-ecological constraint to khat. The improved road oriented systems by Khat-based systems, was observed networks around Hawassa also benefitted the districts in areas close to markets (Fig. 3). The improved road further from the market, such as Dale. networks and the proximity to the regional capital city, An increase in population also results in a growing Hawassa, opened up market opportunities for the labour force. Khat requires about three times more farmers, especially in Wondo Genet. The importance labour than cereal crops on a per hectare basis of a nearby market is related to khat’s perishable (Getahun and Krikorian 1973). However, in the nature, with a maximum shelf-life of four days after densely populated study area there are no labour picking (Distefano 1983). Also in the Hararghe shortages that could have constrained khat expansion. highlands of eastern Ethiopia, the proximity of a Some authors argue that khat may expand even market and easy transportation were determinants of further (Dessie 2013). On the other hand, the khat expansion (Woldu et al. 2015). In addition, the controversy and uncertainty about a possible prohi- high income potential motivated smallholders to bition of its cultivation seem to restrain smallholders allocate more land to khat cultivation. The revenue from allocating all of their land to khat (Dessie and per ha from khat exceeded that of maize and teff 16 Kinlund 2008). times and of coffee three times (Fig. 8b). Dessie (2013) reported similar values, fifteen times for cereals and Implication and options for future developments four times for coffee. In Wondo Genet, the expansion of khat was further stimulated by the growing popu- The traditional home garden systems in the SNNPRS lation density (Fig. 7). The resulting land fragmenta- can no longer absorb the increasing population. Part of tion, particularly after 2005 when population density them have transitioned to intensive khat-based sys- -2 exceeded 800 persons km (Figs. 4, 7), left khat tems, replacing the cultivation of enset and coffee by cultivation as one of the few alternatives to make a khat and hence producing less food. This development living from the tiny farm areas. The strong positive is associated with various risks such as an increased relationship between area share of khat and population dependency on one crop, khat, and hence vulnerability density was also confirmed in eastern Ethiopia (Tefera to harvest failure, price fluctuations, and pests and 2009). In Wondo Genet the total area allocated to cash diseases. This study has shown that the trends in land crops increased over time to almost 50% (Fig. 4). The use and farm types differed between districts depend- inclination of smallholders towards cultivating cash ing on market access, local population density and crops when smaller farm sizes impair food self- elevation. Taking into account this observed hetero- sufficiency was also described by Woldu et al. (2015). geneity is crucial in order to identify adapted inter- In contrast to Wondo Genet, in Bule none of the ventions to achieve social and economic development identified drivers for khat introduction and expansion of smallholder farmers. We have identified several 123 Agroforest Syst (2018) 92:1579–1595 1593 Conclusion options for locally adapted interventions. For the khat- based home gardens, diversifying agricultural produce We found a huge diversity of home garden systems in by intercropping khat with annual crops such as cereals and legumes could reduce the risk of depend- southern Ethiopia and their dynamics over the last two decades. Five home garden types were distinguished ing primarily on one crop. Khat is the main crop receiving artificial fertilizer, which in intercropping based on cropping patterns. Smallholders specialized in khat production typically allocate about half of their systems will also benefit the food crops, contributing to food security. Crop residues can be left in the field to farm land to khat, as such depending on the market for improve soil fertility. The integration with livestock is about half of the year for family food requirements, but difficult as not much feed becomes available from this earning a good income. Contrastingly, farmers special- intercropping system. ized in food production allocated about 75% of their farm to food crops, thus generating significantly less Enset-oriented home gardens offer scope in areas with lower population densities as enset takes several income. Over the past two decades cropping patterns had changed considerably, showing: (1) a shift from years to yield. A more intense coupling of enset with crossbred (i.e. Boran-Holstein–Friesian) milking cows food crop to cash crop production in densely populated areas close to markets; (2) a continuation of combined or small ruminants could be of mutual benefit, as the enset leaves can be used as animal feed and enset food and cash crops in medium populated, less acces- sible areas. In line with changes in land use and livestock needs manure for a good production. The time required for enset to reach maturity is up to eight populations, two main trends in home garden systems years without manure application, compared to three were observed: (1) a transition to cash crop oriented to four years with manure application (Shumbulo et al. home gardens, and (2) development of combined food 2012). Milk and meat could improve the nutritional and cash crop oriented home gardens. These dynamic changes were influenced by changes in population quality of family diets, whereas other livestock functions like banking functions also contribute to density and market access, changes in prices, a decline in soil fertility and a policy of market liberalization. Our rural livelihoods (Moll 2005;Thorntonetal. 2007). In areas where grazing land is scarce the integration with insights on how enset and coffee based home gardens of southern Ethiopia have responded to increasing popu- livestock is not feasible. Here, introducing locally adapted improved coffee and enset varieties could lation pressure and commercialization provide insights contribute to enhance the income and food security of that can inform the design of alternative options for smallholder farmers through improving the productiv- sustainable development. ity. The shortage of manure to maintain soil fertility Acknowledgements This research was funded by the can be addressed by the use of compost, prepared government of the Netherland through the project capacity from coffee husk waste (Abebe 2013) and the inedible building for scaling up of evidence-based best practices in portions of enset leaves and, pseudostem sheath and agricultural production in Ethiopia (CASCAPE). We thank the corm (Tamire and Argaw 2015). There is also scope to farmers who participated in this research. Our thanks are extended to experts from Bureau of Agriculture in Wondo integrate high value and productive fruit trees with the Genet, Melga, Dale and Bule districts for making available aim to maintain the diversified nature of the system secondary data and for their cooperation during the survey data and smallholders’ sources of income. A continuous collection process. We also wish to thank Mr. Azmach Tensaye soil cover with diversified living plants facilitates the for his help in translating questionnaire into local language and supervising enumerators during data collection process. capture and infiltration of rainwater and protects the soil, besides being a source of organic matter through Open Access This article is distributed under the terms of the litter accumulation (Mollison and Slay 1991). How- Creative Commons Attribution 4.0 International License (http:// creativecommons.org/licenses/by/4.0/), which permits unre- ever, in-depth research is required to quantify the stricted use, distribution, and reproduction in any medium, effects of intercropping of the khat-based systems on provided you give appropriate credit to the original productivity and of further integration of enset and author(s) and the source, provide a link to the Creative Com- crossbred animals. mons license, and indicate if changes were made. 123 1594 Agroforest Syst (2018) 92:1579–1595 References determinants of sustainability in smallholder farming sys- tems of Eastern Uganda. Popul Environ 31:474–506. doi:10.1007/s11111-010-0104-2 Abebe T (2005) Diversity in home garden agroforestry systems Falconnier GN, Descheemaeker K, Van Mourik T, Sanogo O, of Southern Ethiopia. Dissertation, Wageningen Univer- Giller K (2015) Understanding farm trajectories and sity, Wageningen development pathways: two decades of change in southern Abebe T (2013) Determinants of crop diversity and composition Mali. Agric Syst 139:210–222. doi:10.1016/j.agsy.2015. in Enet-coffee home gardens of southern Ethiopia. J Agr 07.005 Rural Dev Trop Subtrop 114(1):29–38 Fekadu D, Ledin I (1997) Weight and chemical composition of Abebe T, Wiersum K, Bongers F (2010) Spatial and temporal the plant parts of enset (Ensete ventricosum) and the intake variation in crop diversity in agroforestry home gardens of and degradability of enset by cattle. Livest Prod Sci southern Ethiopia. Agrofor Syst 78:309–322 49:249–257 Amede T, Diro M (2005) Optimizing soil fertility gradients in Feyisa T, Aune J (2003) Khat expansion in the Ethiopian the enset (Ensete ventricosum) systems of the Ethiopian highlands. Effects on the farming system in Habro District. highlands: trade-offs and local innovations. In: Bationo A, Mt Res Dev 23:185–189 Kimetu J, Kihara J (eds) Improving human welfare and Getahun A, Krikorian A (1973) Chat: coffee’s rival from Harar, environmental conservation by empowering farmers to Ethiopia, botany cultivation and use. Econ Bot 27:353–377 combat soil fertility degradation. Tropical Soils Biology Giller KE, Corbeels M, Nyamangara J, Triomphe B, Affholder and Fertility Institute of CIAT, Addis Ababa F, Scopel E, Tittonell P (2011) A research agenda to Amede T, Takele B, Endrias G (2001) Reversing the degrada- explore the role of conservation agriculture in African tion of arable land in Ethiopian highlands. Managing smallholder farming systems. Field Crops Res African Soils No 23 124(3):468–472. doi:10.1016/j.fcr.2011.04.010 ATA (2012) Report on results of crop yield sample survey. Herrero M, Thornton P, Notenbaert A, Msangi S, Wood S, Agricultural Transformation Agency (ATA), Addis Ababa Kruska R, Dixon J, Bossio DA, van de Steeg J, Freeman H, Blazy J, Ozier-Lafontaine H, Dore` T, Thomas A, Wery J (2009) Li X, Parthasarathy R (2009) Drivers of change in crop- A methodological framework that accounts for farm livestock systems and their potential impacts on agro- diversity in the prototyping of crop management systems. ecosystems services and human well-being to 2030. Application to banana-based systems in Guadeloupe. Agric CGIAR Systemwide Livestock Programme (SLP). ILRI, Syst 101:30–41. doi:10.1016/j.agsy.2009.02.004 Nairobi Brandt SA, Spring A, Hiebsch C, McCabe G, Taboge E, Diro M, Joe H, Ward J (1963) Hierarchical grouping to optimize an Wolde-Michael G, Ynitso G, Shigeta M, Tesfaye S (1997) objective function. J Am Stat Assoc. doi:10.1080/ The ‘tree against hunger’: Enset-based agricultural systems 01621459.1963.10500845 in Ethiopia. AAAS, Washington, DC Kandari L, Yadav H, Thakur A, Kandari T (2014) Chat (Catha CSA (1991–2013) The population estimates derived from pro- edulis): a socio-economic crop in Harar region, eastern jecting the 1994 and 2007 census, analytical report for Ethiopia. Springer Plus 3:579. doi:10.1186/2193-1801-3- Southern Nations, Nationalities and People’s Region. Central Statistical Agency (CSA), Addis Ababa Kippie T (2002) Five thousand years of sustainablity? A case CSA (2007) Population and housing census. Analytical report study on Gedeo land use (Southern Ethiopia). Dissertation, for the Southern Nations, Nationalities and People’s Wageningen University, Wageningen Region. Central Statistical Agency (CSA), Addis Ababa Kumar BM, Nair PKR (2004) The enigma of tropical home- Descheemaeker K, Ronner E, Ollenburger M, Franke AC, gardens. Agrofor Syst 61:135–152 Klapwijk CJ, Falconnier GN, Wichern J, Giller KE (2016) Kumar B, Nair P (eds) (2006) Tropical Homegardens: a time- Which options fit best? Operationalizing the socio-eco- tested example of sustainable agroforestry 3. Springer, logical niche concept. Exp Agric:1–22 Dordrecht Dessie G (2013) Current African issues: favouring a demonised Lemessa D (2001) Khat (Catha edulis): botany, distribution, plant khat and Ethiopian smallholder enterprise. No 51. cultivation, usage and economics in Ethiopia. UN-Emer- Nordic Africa Institute, Uppsala gencies Unit for Ethiopia, Addis Ababa Dessie G, Kinlund P (2008) Khat expansion and forest decline in Mekbib F (2009) Farmers as treasure chest of biodiversity: the Wondo Genet, Ethiopia. Geografiska Annaler Ser B Hum case of sorghum (Sorghum bicolor (L.) moench) in Ethio- Geogr 90(2):187–203 pia a five decade analysis. Maydica 54:11–23 Distefano J (1983) An enquiry into the history of Qat. Pro- Moll HAJ (2005) Costs and benefits of livestock systems and the ceedings of the 2nd international conference on Somali role of market and nonmarket relationships. Agric Econ Studies, Hamburg, p 1–6 32:181–193 Dorsey B (2015) Agricultural intensification, diversification, Mollison B, Slay R (1991) Introduction of permaculture. The and commercial production among smallholder coffee Tropical press, Harare growers in Central Kenya. Econ Geogr 75:178–195 NBE (1991–2014) Exchange rate report National bank of Dube D, Yadav R, Dube S (2014) Why a shift from coffee to Ethiopia. Addis Ababa chat? A study of the Kersa Woreda in Jimma Zone of South Rehima M, Belay K, Dawit A, Rashid S (2013) Factors affecting Western Ethiopia. Res Human Soc Sci 4:177–182 farmers’ crops diversification: evidence from SNNPR, Ebanyat P, de Ridder N, de Andre J, Delve R, Bekunda M, Giller Ethiopia. Int J Agric Sci 3:558–565 K (2010) Drivers of land use change and household 123 Agroforest Syst (2018) 92:1579–1595 1595 Samuel M (2014) Livestock feeds and feeding system in enset (2007) Coping strategies in livestock dependent house- (Ensete ventricosum) dominated mixed farming systems of holds in East and southern Africa: a synthesis of four case southern Ethiopia. J Anim Feed Res 4(6):150–158 studies. Hum Ecol 35:461–476. doi:10.1007/s10745-007- Shumbulo A, Gecho Y, Tora M (2012) Diversity, challenges and 9118-5 potentials of Enset (Ensete ventricosum) Production: in Timmer C (1997) Farmers and markets: the political economy of case of Offa Woreda, Wolaita Zone, southern Ethiopia. new paradigms. Am J Agr Econ 79:621–627 Food Sci Qual Manage 7:24–31 Tsegaye A, Struik P (2001) Enset (Ensete ventricosum (Welw.) Solomon Z, Masresha F, Olsson M (2008) Effect of drought/ Cheesman) kocho yield under different crop establishment irrigation on proximate composition and carbohydrate methods as compared to yields of other carbohydrate-rich content of two enset (Enset ventricosum (Welw.) (Chees- food crops. Neth J Agric Sci 49:81–94 man) clones. Ethiop J Sci Technol 31(2):81–88 Von Braun J (1995) Agricultural commercialization: impacts on Tamire C, Argaw M (2015) Role of Enset (Ensete ventricosum income and nutrition and implications for policy. Food (Welw.) Cheesman) in soil rehabilitation in different agro- Policy 20:187–202 ecological zones of Hadiya, Southern Ethiopia. Am J Env Wiggins S, Gem A-K, Jennifer L, Colin P (2011) Small farm Protect 4(6):285–291. doi:10.11648/j.ajep.20150406.14 commercialisation in Africa: reviewing the issues. Tefera TL (2009) Supply response, local reality and livelihood Research Paper, no 23. Brighton, UK sustainability: the policy dilemma of khat (Catha edulis) Woldu Z, Belew D, Benti T (2015) The coffee-khat interface in production in eastern Ethiopia. Int J Agr Sustain Eastern Ethiopia: a controversial land use and livelihood 7(3):176–188. doi:10.3763/ijas.2009.0428 change scenario. J Agri Sci Technol 5:149–169. doi:10. Thornton P, Boone R, Galvin K, Bunsilver S, Waithaka M, 17265/2161-6264/2015.03.001 Kuyiah J, Karanja S, Gonzalez-Estrada E, Herrero M

Journal

Agroforestry Systems – Springer Journals

Published: Dec 1, 2018

There are no references for this article.