Exploring farmers' agrobiodiversity management practices and knowledge in clove agroforests of Madagascar

A free Plain Language Summary can be found within the Supporting Information of this article. Research on local ecological knowledge, defined as ‘a cumulative body practice and belief, evolving by adaptive processes handed down through generations cultural transmission’ (Berkes, 1999), shows that communities have a specific holistic understanding their environment which contributes to capacity manage natural resources sustainably in an way cope with contingencies its spatial temporal heterogeneity (Berkes et al., 2000; Folke, 2004). Knowledge plant species, uses management practices been widely investigated because they concern socio-cultural, economic environmental issues, such biodiversity conservation (Garibaldi & Turner, 2004; Rafidison, 2020), adaptation agriculture climate change (Labeyrie 2021) or food security rural (Quave Pieroni, 2015). Multiscale studies revealed role indigenous perceptions beliefs control soil erosion fertility, land use forest regeneration secondary fallows (Barrera-Bassols 2006; Carrière, 2002). All these emphasise how studying knowledge plants species helps understand human/nature interactions, why play critical resilience socio-ecological systems (Congretel Pinton, 2020; In agroecosystems, interactions between farmers agrobiodiversity, is, living organisms contribute agroecosystem functioning (Altieri, 2002; Jackson 2007; Vandermeer Perfecto, 1995), are key dimension interactions. particular, it is acknowledged agroecosystems maintain high level diversity more resilient socio-economic perturbations (Isbell 2017; 2010; Lin, 2011). The agroecology literature particularly emphasises importance farmers' organise different space depending ways adapt biophysical conditions, associate account for positive negative However, aspects farmers/agrobiodiversity still poorly understood, although improved agrobiodiversity associated crucial if agricultural development initiatives match realities meet global challenges (Altieri Nicholls, Labeyrie 2021). Agroforestry illustrative type agrobiodiversity-based cropping system tropics subtropics where interact daily (Nair, 1991). ‘where woody perennials (trees, shrubs, palms, bamboos etc.) deliberately used same unit crops and/or animals, either some form arrangement sequence’ (Zomer 2014). recognised agroecological which, under certain ensures sustainable production (Mercer, Schroth McNeely, effects agroforestry productivity (Pumariño 2015), (Bhagwat 2008; Jose, 2012) ecosystem services provides (Jose, 2009) well-being populations (Reed 2017), increase (Mbow 2014) help restore degraded landscapes (Hillbrand 2017). rarely focusses famers' drive organisation diversity, domain pivotal efficient resources. few described mainly home gardens (Abebe 2013; Méndez 2001) relation fertility (Junqueira, 2015) altitude (Sahoo Rocky, but none take into consideration observed practices. Inversely, light shade (Graefe Smith Dumont 2019), invertebrates influence (Pauli other types affect range (Ango 2014; Cerdán 2012; Rigal 2018), analyse influences space. lack linking due methodological challenges, what required articulation methods properly collect them from perspective, allow quantification measurement patterns species. aim present study was fill research gap investigating relate environment, based case clove-based agroforests north-east coast Madagascar (Figure 1). An emic perceive own actions related (Olivier de Sardan, 1998), participatory interviews were data. First, current functions characterised along socio-cultural drivers diversity. Second, analysed at two levels: association sub-units distribution overall agroforest plot. Third, documented focusing clove trees, agroforests. Finally, discussed process shapes organisation, function carried out Vavatenina district Analanjirofo region 2). This experienced massive planting cash crops, especially coffee trees spicy imposed colonial government beginning 20th century satisfy needs French metropolis achieve colony's financial autonomy flourishing export-based (Dandoy, 1973; Isnard, 1951; Ruf Blanc-Pamard, 1992). Driven factors, (literally ‘Forest trees’) gradually became main zone plantations Syzygium aromaticum instead emblematic shifting rice cultivation called tavy Danthu products provided tree, cloves essential oil become major sources income, currently, many Malagasy people east base strategy both self-production income (Droy surveys conducted village Vohibary located 50 km 5 town Vavatenina, district's administrative centre. chosen in-depth agrarian diagnosis 1966 enabled diachronic approach 1973). Most population belongs Betsimisaraka ethnic group. Between 2019, increased 186 inhabitants (fewer than households) over 120 households, leading creation second hamlet hillside next original 1973 survey data 2016). humid tropical 3,600 mm rainfall per year average annual temperature 24℃ (Jury, 2003). Local differentiate three seasons: cool wet winter season June August, hot summer rest very period December March during cyclones occur. coastal severely regularly affected cyclones, last ones Honorinina (1986) Bonita (1996) Ivan (2008). landscape shaped tanety (hills) exceeding 400 m height slopes gullied rainfall. hydrographical network structured around river, Sahameloka, fed numerous streams flowing springs irrigates rice-growing lowlands. Soil characteristics closely correlated topography. Soils lowlands hydromorphic accumulate organic inorganic sediments. occupied eroded ferralitic soils overlying humus-bearing horizon Along topological gradient, soil's texture, matter content depth variable create heterogeneous conditions author's measurements made auger). enables perennial crops. cultivate paddy valley irrigable flat lands. On tanety, association, tsabo, including (clove, coffee, fruit various palms others), lianas (vanilla pepper) herbaceous (sugarcane, cassava, maize, sweet potato others). mixed several ways, time, densities levels stratification. cultivated tsabo are: presence least crop also naturally occurring non-cultivated shrubby bamboo typically compose fallow vegetation (Rafidison 2020). These correspond definition given Torquebiau (2000), agro-silvicultural productive combine commercial subsistence products. One concerns agroforests, (Arimalala 2019). CIRAD ethics committee does not deliver ethical approval involving humans. To carry study, Code Ethics ISE (International Society Ethnobiology) followed. work collaboration University Antananarivo frame ‘Forêt Biodiversité à Madagascar’ platform (CIRAD-FOFIFA-University). permit issued Secretary General ‘Direction Générale l'Environnement des Forêts’ before undertaken. Before holding village, meeting organised leader fully inform purpose would conducted. methodology collected explained so knew participating involve. future expected benefits inhabitants. Written consent leader. participated voluntarily coercion, had right withdraw any time. conducting interviews, gave prior informed verbally. No activities without consent. involvement interpreter, native region, helped ensure rules customs respected well rights interviewed. semi-directive sample 17 voluntary corresponding household head (14% households 1 woman 16 men). gender disparity linked division tasks genders, men being responsible most cases. following criteria: only farmer began cover age categories location distributed geographical area village. interviewed asked list on-site all present, using common names. considered important them, functions. It included planted spontaneously tolerated, even encouraged farmers. correspondence names Latin established interpreter specialists Antananarivo. document each cited, focus groups organised: one hamlet, 10 villagers, comprised two-thirds one-third women groups. method draw map visited, al. (2001). maps describe mentioned farmers: (i.e. according slope) sub-unit farmers, micro-zone 3). (2001) adapted context perceptions, assemblage differs Hence, micro-zones units homogeneous perspective. whether there pattern managed tsabo. If so, outline Some throughout corresponded micro-zone. way, total 28 (with tsabo) obtained. surface measured GPS. size 0.5 ± 0.3 ha (min = 0.10 ha/max 0.96 ha) 0.29 0.23 0.03 0.87 slope 17.7 8.9° 9°/max 29°). Farmers' estimations abundance vanilla four frequent (avocado, breadfruit, jackfruit lychee) recorded number and, category. Indeed, exploratory surveys, chief elders, tree started producing structuring factor managing need after becoming productive. According usually years old, may younger older growing conditions. coded enable comparison Two variables selected it: position topographical gradient top, centre bottom; Figure 3) spatially compared random, edge aggregated; For variable, times counted. Fisher's exact test applied occurrence significantly organisation. characterise typology presence/absence trees. Thus, analysis, separated group, differentiation similarity calculated Jaccard index: pair micro-zones, index corresponds ratio excluding those absent al.,l., Borcard analysis R (version 3.6.2) VEGDIST vegan package 2.5.6). similarity, grouped clusters representation agglomerative complete linkage principle (Borcard identified first discrimination cluster. Each computing mean standard deviation, 60% belonging type, six Semi-structured about hours resulting questions centred tree. explain interaction described, negative. discussion represented, deepen information gathered semi-structured summarised transcribed map. mental model framework represent construction actors perception general phenomena (Özesmi Özesmi, our case, intended bring occur built nature involved (competition growth, fertilisation, uptake water soil, application physical element (soil, light). process) (out 17) who interaction. cited sampled consisted 51 36 seven plants, bamboos, (Appendix Plant families, Myrtaceae represented five followed Fabaceae, Moraceae Rutaceae, reports varied sample. average, listed 15 (±4; 8–22). Only 22% half Artocarpus heterophyllus lychee Litchi chinensis, breadfruit altilis banana (Musa sp.). Coffee Coffea canephora Vanilla planifolia studied. 54% fewer eight among these, 28% each. villagers nine azahoam bola category ‘gives money’) bômafana (literally: ‘what heated up day eaten morning’) whose carbohydrate-rich fruits tubers consumed breakfast afternoon (breadfruit; yam, Dioscorea spp.; Manihot esculenta). dessert contained wide snacks nutritional foods (dessert voankazo non-woody species). make vegetable (menaka): palm Elaeis guineensis cooking coconut Cocos nucifera haircare. rô-mazava ‘clear broth’) prepare broths accompany rice. supply wood: kitay firewood (distilling cooking) kakazo trano ‘house tree’) carpentry (poles, beams, fencing, roofs boards) sometimes firewood. assigned uses, bola. Lastly, did named (Albizia stipulata, A. lebbeck), support (Pachira aquatica, Ficus sp., Gliricidia sepium) Raphia farinifera, leaves string woven stool seats. results indicated significant dependence describing organisation: topography (p value < 0.05) 0.0005). Analysis (aggregated, random) showed tended distribute randomly across silky oak, entire 4a). large lychee, border Aggregated appeared apply particular vanilla, banana, (38% ≤ 60%). Concerning prefer bottom conserve (‘be’ species) location: 50% individuals 4b). demanding terms clove, sepium, equally frequently positions. seemed let oak Grevillea banksii traveller's Ravenala madagascariensis grow top (occurrence ≥ 50%): since, when locations, removed them. absence old produced cloves) differentiated micro-zones. Group (tsabo trees) 13 82% previously occurred. B young no 74% clustering dendrograms led identification group: 2 (Ma1, Ma2), (Y1, Y2) 5). (Ma1), represented. (mean: 2, They mature jackfruit, avocado density (161 177 trees/ha, 6) (84 65 trees/ha). (17 18 trees/ha) rich highest (132 183 plants/ha), (Ma2 56 139, Y1 108 215, Y2 4 12 plants/ha). less diversified Ma1. accounted richness 3, (frequency 100%). palm, lychee. Unlike Ma1, Ma2 (194 93 low (39 old. clearly distinguishable composition: 72% whereas ascribed 38%. Type (mean 3 stood species: (present micro-zones), micro-zones). composition highly Y1, shared (230 199 6), (18 24 (29 42 higher much shown deviations. Half reached 632 plants/ha. therefore displayed variability structure variations present. lower richness, common. (268 211 occurred (58 31 another 7): 21 effects. referred 19 eucalyptus roots water, growth. Lychee canopy too shade. 11, interacted positively improving providing adequate underlined central place ones. tree's factor, 11 pointed suitable During specifying involved. instance, taller (banana, jackfruit). Conversely, tall cast Our confirmed nearby areas limited differences (lists vs. floristic inventory). findings agreement showing (food, timber, stake, tree),