ACIAR WP1 mid-term report – Ethiopia
World Agroforestry Centre
October 2014 Miyuki Iiyama, Ermias Ayenkulu, Roeland Kindt, Anne Kuria, Joseph Tanui, Evelyn Kiptot,
Ruth Kinuthia, Kiros Hadgu, Abayneh Derero, Kaleb Kelemu, Catherine Muthuri
Contents
1: Introduction..................................................................................................................................... 3
1.1: Background............................................................................................................................... 3
1.2: Research Objectives ................................................................................................................. 3
1.3: Site Selection ............................................................................................................................ 4
1.4: Materials and Methodologies .................................................................................................. 6
2: Socioeconomic and Biophysical Baseline (1.1)................................................................................ 8
2.1: Land health baseline (1.1.1) ..................................................................................................... 8
2.2: Tree density, cover, seed/seedling delivery systems (1.1.2) ................................................. 10
2.3: Food security, health/nutrition and income (1.1.3) ............................................................... 10
2.4: Opportunities and barriers (1.1.4) ......................................................................................... 12
3: Match Species and Management Options (1.2) ............................................................................ 17
3.1: Spatially explicit guidelines on initial best-bet tree species and management options (1.2.1)
....................................................................................................................................................... 17
3.2: Burundi vegetation map (1.2.2) ............................................................................................. 20
3.3: Decision support tools for matching species and management options to sites and
circumstances (1.2.3) .................................................................................................................... 22
4: Target and Priority Sites and Farmer Circumstances (1.3) ............................................................ 25
4.1: National dialogues on priotisation criteria (1.3.1) ................................................................. 25
4.2: Frameworks for priotising landscapes and farming systems for AF interventions (1.3.2) .... 27
5: Conclusions and recommendations .............................................................................................. 29
1: Introduction
1.1: Background
During the last century, the expansion of agricultural land to support the growing population in Ethiopia was in general achieved by massive deforestation with soil degradation. This land use change has transformed forests and woodlands into agricultural land with trees scattered as the forest-woodland remnants.
Scattered trees on farm characterize a large part of the Ethiopian agricultural landscape today, while tree species differ depending on their agro-ecological suitability such as rainfall, altitude, and soil and natural distribution patterns. In central Ethiopia, for example, Faidherbia albida trees are dominant in tef-wheat zones of central and eastern highlands. Diverse Acacia species such as A. nilotica, A. seyal, A. senegal and A. tortilis are abundant in low lying dry savanna regions. Cordia africana is commonly observed intercropped with maize in sub-humid parts of western highlands.
Through either farmers managed natural regeneration of trees (FMNR) or active planting, a massive-scale adoption of trees on farmlands can play an important role to enhance tree diversity and cover- at landscape level, then potentially contribute to enhancing food security of resource-poor smallholders through the provision of ecosystem goods and service.
In turn, across landscapes, patterns of trees on farm are extremely varied in terms of mixes of species types, multiple utilities and functions, management intensities, niches, and so on, under specific biophysical and socio-economic circumstances. Even given a similar biophysical conditions, the socio-economic needs of the households as well as institutional factors makes the mode of management and the intensity of adoption of trees more dynamic. In this context, recommending a single tree species with a single utility or function is not likely to be effective to address rather complex systems in which farmers decide to integrate and manage trees on farms.
The success of scaling up and out of trees on agricultural landscapes to enhance food security is dependent on characterizing heterogeneous patterns of trees on farm in local specific contexts, and identifying factors affecting optimal mixes of species and management.
1.2: Research Objectives
The project- Trees for Food Security- targets semi-arid and sub-humid agro-ecologies of Ethiopia and Rwanda respectively to understand the roles of trees on farm in relation to crop/livestock management and overall livelihood goals of farm households. The main objectives of the WP1 include to characterize the agro-ecologies in order to target intervention areas with matching tree species and management systems and to identify biophysical, socio-economic, and institutional barriers to adoption, so as to develop effective strategies for scaling up/out the packages of
agroforestry technology to wider but similar agro-ecologies of the countries and beyond (Uganda, Burundi). The sound analyses of understanding underlying conditions to enhance trees on farm and landscapes will contribute to promoting tree cover and diversity, which then will lead to resilient agricultural systems for farm household as well as at landscape levels. This mid-term report summarizes key findings from different baseline tools used to collect various socio-economic, institutional biophysical data, to consolidate findings to guide recommendations for targeting interventions, while more detailed separate reports are available on the project website (http://www.worldagroforestry.org/aciar/resources ).
1.3: Site Selection
The study was conducted in Oromia region of Ethiopia. Oromia region accounts for 34.3% of a total area of the country with the population of over thirty million people, making it the largest state in terms of size and population in the country. Oromia is a region of great physiographic diversity whose landscape includes high and rugged mountain ranges, undulating plateaus, panoramic gorges and deep incised river valleys, and rolling plains. About 30 percent of the lowlands of eastern sub-region have arid climate. Over 35 percent of the intermediate highlands of central and western Oromia have hot tropical rainy climate, while the highlands have warm temperate rainy, tropical climate. The rainfall pattern of the region is bimodal, receiving the greatest share of rainfall in summer and the smallest portion in spring. The distribution of mean annual rainfall varies from place to place and from year to year, decreasing in all directions from the western highlands (1600-2400 mm) towards the eastern and south eastern arid lowlands (less than 400 mm).
In selecting the sites for the project, the diversity of crop/trees (and livestock) systems within an agro-ecology (semi-arid or sub-humid) are to be identified to characterize the complexities of the tree/crop systems in relation to intra-agro-ecology heterogeneity in biophysical (altitude, soil, rainfall etc.) and socio-economic (cultural, market/infrastructural access, institutional arrangements) conditions. This project was conducted in East Shewa Zone chosen to represent diverse trees on farm patterns in the semi-arid agroecology, and East Wollega and West Shewa Zones to represent those in sub-humid agroecology (Figures 1.3.1, 1.3.2).
The semi-arid sites mostly fall in the lowlands of the Central Rift Valley with the altitude less 1500 m, whereas it rises up to 2300 m at the mountain fringes of the Rift. Grain crop and livestock farming are dominant in the study area. Some diversity is observed in terms of the combination and management of crop-tree systems; from tef-wheat plus Faidherbia albida to maize-beans-sorghum plus Acacia ssp. across the north-south transect, and from tef-wheat plus Faidherbia albida to tef-maize-sorghum plus Acacia tortilis and Ziziphus spina-christi across the west-east transects, while the livestock system is dominantly communal grazing of cattle and goats on farmland.
The sub-humid sites are characterized by rugged landscapes, with hills and valleys. , Dominant crops are maize, sorghum, tef, nug (oil crop), while relative importance may be vary depending on
altitudes and micro-climates. Wheat is also observed in relatively higher altitudes and sugarcane in swampy areas in lower altitudes. In contrast, types of trees species are observed to be heterogeneous across the north-south and east-south transects as well as by altitudes. Croton macrostachyus is dominant on home compound and farm boundaries in southwestern tepid/mid-highland, Ficus spp. in eastern tepid/mid-highland, and Cordia africana in eastern warm lowland, while Eucalyptus spp. small woodlots and Acacia abysinica on farm are commonly observed. Across the east-west transect, fruits trees such as pawpaw, mangoes, bananas and coffee are observed in home compound, along rivers or localities with irrigation structures available.
Figure 1.3.1: Agroeclogical map of Ethiopia with locations of selected sites
BofaMojo
Alem TenaMekiZeway
TibeAnoUke
Arjo
Addis Abeba
2a: F. albida on tef farm in semi-arid Oromia
2c: C. africana on maize farm in sub-humid Oromia
2b: Ziziphus, Acacia on maize farm in semi-arid Oromia 2d: Various trees on homestead, in sub-humid Oromia
Figure1.3. 2: Selected trees on farm patterns observed in semi-arid and sub-humid zones of Oromia
1.4: Materials and Methodologies
Several surveys were undertaken to understand and characterize trees on farms following the nested research design (Box 1.4.1). These were: a socio-economic survey, tree inventory, stakeholder interviews respectively on marketing of tree products, existing extension systems, seed-seedling systems, local knowledge study, as well as a land-health survey through soil sample analysis using the land degradation surveillance framework (LDSF) developed by the World Agroforestry Centre (UNEP 2012).
Box 1.4.1: Nested Research Design
socio-economic (300 HHs from 4-5 sites x 2 zones)
land health (1 representative site x 2 zones)
tree inventory (100HHs x 2 zones )
marketing survey (4-5 sites x 2 zones) *
extension system (4-5 sites x 2 zones) *
seed-seedling system (4-5 sites x 2 zones)*
local knowledge study (1 site x 2 zones)
national dialogues
*These surveys will be reported by WP3.
Figure 1.4.1: Sampling in Ethiopia
This report is organized corresponding to the deliverables of the Log-frame of the proposal, followed by concluding remarks and recommendation derived from consolidated findings of the baseline.
Ano sentinel site (10 km x 10 km): Socio-economy (red points) ( and biophysical (green points) sampling points
Socio-economy sampling points
Sampling in Ethiopia
Land health survey points
Marketing, extension, seed-seedling system sampling points
Tree inventory survey points
2: Socioeconomic and Biophysical Baseline (1.1)
2.1: Land health baseline (1.1.1)
The Land Degradation Surveillance Framework (LDSF) was applied to characterize two sentinel sites in semi- arid (Alem Tena) and sub-humid (Ano) agroecologies in Ethiopia. A sentinel site is a 10 km x 10 km block stratified into 16 clusters of 100 ha. Each cluster then is represented by 10 randomly selected plots of 0.1 ha area. Within each sampling, plot there are 4 subplots of 0.01 ha where data on vegetation cover, structure, floristic composition, land use, topography, visible signs of soil erosion and soil physical characteristics were collected. Data on vegetation cover, structure, floristic composition, and specific tree attributes such as breast height diameter (dbh) and height as well as land use, topography, visible signs of soil erosion and soil physical characteristics have been collected using a hierarchical stratified, random sampling approach.
The study sites of both zones are dominantly under cultivation in which crop lands cover about 70% of the land, while the semi-arid site had higher shrub (< 3m) density than the sub-humid site (103.0 vs. 33.4 shrubs/ha respectively) but lower tree (> 3m) density than in sub-humid site (49.9 vs. 65.8 trees/ha). Uncultivated plots had higher shrub/tree densities than cultivated plots both in the semi-arid and sub-humid sites, while few but some observations on cultivated plots of shrubs in the semi-arid and trees in the sub-humid sites.
Figure 2.1.1: Shrub and tree densities in cultivated vs. uncultivated plots
In terms of species diversity, Leguminosae is the dominant family in both sites. Acacia tortillis (relative density =30%) and croton macrostachyus (relative density= 16%) are the dominant species in the semi-arid and sub-humid sites respectively. The sub-humid (46 species) was more species rich than the semi-arid (28 species) site.
Figure 2.1.2: Species relative densities in Alem Tena semi-arid (left) and Ano sub-humid (right) sentinel sites.
Figure 2.1.3: Species-accumulation curve
The topsoil (0-20 cm) has significantly higher carbon and nitrogen content than the subsoil (20-50 cm). The topsoil for the humid trip humid-tropics site has significantly higher carbon and nitrogen content than the semi-arid site for all depths classes (0-20 cm, 20-50 cm, 50-0 cm and 80-110 cm). The mean (standard deviation) for carbon for the top soils in the sub humid and semi-arid sites were 3.05(0.81) and 1.18(0.49), respectively. 93% of the plots in the semi- arid site and 8% in the sub-humid site had low (< 2%) carbon content which could be to the inherent soil properties and land management practices. The soils in the humid-tropics site, for instance, are richer in clay content than the soils in the semi-arid site. Soil carbon and nitrogen content also varies with land use types in which cultivated lands have less Carbon and Nitrogen than semi-natural vegetation. We found 66 and 55 trees per ha in the sub-humid and semi-arid sites, respectively. Sheet erosion is the major cause of soil loss. No root depth restriction was recorded in both sites.
The semi-arid site has low carbon and nitrogen content that needs an integrated soil management practices to management that enhances the soil carbon and prevents further soil degradation.
2.2: Tree density, cover, seed/seedling delivery systems (1.1.2)
Data collection and data entry for tree diversity studies was completed in the first quarter of 2014 (Y2 M7-M12) by three MSc students (two in Ethiopia, one in Rwanda). Complete inventories were made on a subset of households included in the baseline surveys, as planned (selecting 200 households per country, stratified by geographical locations). Data analysis is underway, whereas two students have prepared early drafts of their thesis (one student has already submitted the draft. A map showing the locations and species diversity at the different locations will be created by the time of mid-term review meeting.
2.3: Food security, health/nutrition and income (1.1.3)
The socio-economic survey focused on collecting categorical/qualitative information on farmers’ perceptions on the status of trees on farm systems as well as income/asset parameters of households, while quantitative information on trees on farm was captured through the tree inventory survey for about a third of the households surveyed. This is due to the assumption that there would be significant bias in households’ recollection to present quantitative data such as on the number of trees on farm, income levels and asset status.
Figure 2.3.1 presents the mean values of the diverse indicators for assets and income categories by the different agroecological zones. In terms of assets (livestock as well as different asset categories), the semi-arid zone households scored higher than those of the sub-humid zones. For farm income scores, the semi-arid zone households had higher scores for cereal & pulse and animal incomes, while the sub-humid zone households had higher scores for cash crops, roots & tubers and tree-based income. For off-farm income scores, the semi-arid households had higher scores for casual employment and rent, while the sub-humid zones had slightly higher scores for regular income and loan.
Figure 2.3.2, created by ILRI team, presents food security ratio indicators. They indicate that 18% of the households in the semi-arid zone and 24% of the sub-humid zone were food insecure. Probably this is due to the fact income/asset strategies would be determined somehow geographically. According to the theory of development domain, agricultural potentials, market access and population density, matter for production indicators and resource use practices (cropping pattern and livestock use). For our study area, East Shewa Zone for the semi-arid zone is less agro-ecologically favored especially in terms of rainfall and soil, but better accessed infrastructure and markets than West Shewa/East Wollega Zones for the sub-humid zone at the time of the survey. With ILRI, we will work to see the association between food security indicators and adoption of tree diversity pattern indicators, controlling zonal effects, to understand the role of trees on food security.
Figure 2.3.1: Livestock, assets, farm/off-farm income diversification indicators by the agroecological zones
Figure 2.3.2: Food security ratios for semi-arid (left) and sub-humid (right) zones by ILRI
-
1.0
2.0
3.0
4.0
5.0
6.0
7.0
Semi-arid zone Sub-humid zone
FAO
Liv
est
ock
Un
it
Livestock - local animals
Livestock - exotic animals
-
50
100
150
200
250
300
Semi-arid zone Sub-humid zone
Esti
mat
ed
val
ue
s in
USD Asset - domestic asset
value
Asset - communicationasset value
Asset - transport assetvalue
Asset - farm asset value
-
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
Semi-arid zone Sub-humidzone
aggr
ega
ted
sco
res
du
rin
g th
e la
st 1
2
mo
nth
s; 0
=no
t p
rod
uce
d,
1=p
rod
uce
d&
con
sum
ed
, 2=i
f p
rod
uce
d&
sold
fo
r ca
sh
Farm income- cereal &pulse
Farm income - cash crops
Farm income - roots &tubers
Farm income - animal
Farm income - tree-based
-
0.1
0.2
0.3
0.4
0.5
0.6
Semi-arid zone Sub-humid zone
aggr
ega
ted
sco
res
du
rin
g th
e la
st 1
2
mo
nth
s; 0
=no
t e
nga
ged
, 1=
en
age
d a
nd
e
arn
ed
inco
me
Off-farm income - regularbusiness
Off-farm income - casual
Off-farm income -remmitance & gift
Off-farm income - loan
Off-farm income - rent
Fo
od
se
cu
rity
ra
tio
01
23
45
60
12
34
56
01
23
45
60
12
34
56
01
23
45
60
12
34
56 off farm
sold lvst
sold oth crop
sold cash crop
cons livestock
cons crop
18 % food insecure hh
Food Security ratio
Fo
od
se
cu
rity
ra
tio
01
23
45
01
23
45
01
23
45
01
23
45
01
23
45
01
23
45
off farm
sold lvst
sold oth crop
sold cash crop
cons livestock
cons crop
24 % food insecure hh
Food Security ratio
2.4: Opportunities and barriers (1.1.4)
A farm household may adopt a portfolio of diverse tree species in specific management intensity, to maximize utilities derived from goods and functions from trees, while bounded by not only physical but also economic scarcity of its resources. On the other hand, patterns of trees on farm are complex and extremely heterogeneous in terms of mixes of species types, multiple utilities and functions attached, management intensities, niches occupied, and so on, under specific biophysical and socio-economic circumstances. Many farmers in Ethiopia include those who never practice an intensive growing of commercial tree species, indeed implement FMNR to manage a set of indigenous species scattered on farm for multiple utilities. In turn, if few farmers ever invest in intensive agroforestry, it is of great interest to understand the motivation behind it as well as its enabling vs. constraining factors to promote such practices.
As Tables 2.4.1 and 2.4.2 below present, in the semi-arid zone, Acacia species, especially A. tortilis and A. senegal were most commonly mentioned along with Ziziphus spp, Faidherbia A, and Balanties aegyptiaca, while in the sub-humid zone, Cordia africana, Croton macrostachyus, Vernonia amygdalina, Mangifera indica and Eucalyptus spp. were commonly mentioned. At the same time there was high variability of the proportion of households adopting these tree species across sites/households within each agroecological zone. For example, the adoption rate of Acacia tortilis ranged between 35% and 78% across sites of the semi-arid sites, while that of Cordia africana varied between 38% and 67% across the sub-humid sites. On the other hand, while 105 and 126 different tree species were counted in the semi-arid zone and sub-humid zone respectively, some species names reported in local languages were difficult to be identified and verified with scientific names by enumerators/Ethiopian botanists. At the same time, farmers often mentioned family names only (ex. Acacia, Eucalyptus) without specifying species. Thus there could be significant problems of counting species diversity by recalling famers’ memory, which need to be verified with the tree inventory survey.
Then we assume that different levels of the intensity in agroforestry adoption at household level– raise seedlings, plant trees, protect naturally regenerated trees – may be not completely independent from types of utilities derived from those trees on farm. For example, tree species with high commercial values, such as fruit and timber, would be more intensively managed through raising and planting seedlings, while those with subsistence and ecosystem service utilities may be managed through natural regeneration. This study thus employed Principal Component Analysis (PCA) to derive independent component factors (i.e., each component has zero correlation) reflecting the diversity of trees on farm patterns, accounting the potential correlations between utilities derived and management intensities.
Four derived components (Table 2.4.3) were interpreted as following. Component 1 represents farmers managed natural regeneration of trees on farm, esp. subsistence, woodfuel, environmental services and fences (FMNR). Component 2 represents active implementation of high value agroforestry species (HVAF). Component 4 indicates agroforestry practices with focus on environmental services but without fence, while Component 3 seemingly captured cases in which farmers did not specify utilities derived from specific tree species. Figure 2.4.1 presents the distribution of the derived principal component scores for households by study sites. Despite
variances, the sub-humid sites in general have higher scores for Component 2 (HVAF) than those in semi-arid sites, while the intra-site variability seems larger than the inter-site difference for Component 1 (FMNR) score. The inter-site trends or differences are even less clear for Components 3 (OTHERS) and 4 (ENV-FENCE) with many outliers observed.
Table 2.4.1: Proportion of households with the most common species in the semi-arid sites
Table 2.4.2: Proportion of households with the most common species in the sub-humid sites
Once proxy variables for distinctive trees on farm patterns were derived, regression analyses were employed to determine which agroecological and socio-economic factors would have enabling vs. constraining effects on each of the patterns of particular interests. The especially interesting indicators for trees on farm patterns were Components 1 (FMNR) & 2 (HVAF) (Table 2.4.4).
Site name/code/ no. of households
Bofa semiarid1
Mojo semiarid2
Alem Tena semiarid3
Meki semiarid4
Zeway semiarid5 Total
rank. Tree species name 120 37 55 48 80 340
1 Acacia tortilis 78 70 58 52 35 60
2 Zizyphus spp 68 24 31 13 5 35
3 Faidherbia albida 13 73 36 40 8 26
4 Acacia senegal 37 5 24 17 23 25
5 Balanites aegyptiaca 18 32 16 21 28 22
6 Croton macrostachyus - 14 38 42 1 14
6 Melia azedarach 19 3 20 13 6 14
8 Acacia etbaica 16 43 - - - 10
8 Eucalyptus spp 13 16 13 4 3 10
10 Ajo_loc - - 2 - 39 9
Site name/code/ no. of households
Arjo subhumid1
Ano subhumid2
Bako subhumid3
Uke subhumid4 Total
rank. Tree species name 76 104 75 92 347
1 Cordia africana 38 63 67 48 54
2 Eucalyptus spp 36 46 55 28 41
3 Croton macrostachyus 46 51 36 14 37
4 Vernonia amygdalina 22 22 35 33 28
5 Mangifera indica 3 20 17 20 16
6 Albizia gummifera 33 15 5 4 14
6 Syzygium guineense 12 12 7 25 14
8 Ficus spp 11 17 3 11 11
9 Acacia sieberiana 13 13 5 - 8
10 Calpurnia aurea - 16 9 - 7
10 Vernonia spp 16 6 5 2 7
10 Ficus vasta 1 13 8 3 7
Table 2.4.3: Derived components representing utilities and management intensities
Figure 2.4.1: Derived principal component factor scores for each household by study sites
Principal Component
Normalized scores Extraction 1 2 3 4
No. of species whose seeds/seedlings were raised during the last 3 years .700 -.192 .689 -.167 .402
No. of species which were planted during the last 3 years .642 -.203 .737 -.129 .203
No. of species which were naturally regenerated and protected during the last 3 years .588 .734 .102 .189 .060
Scores for species types by utility for high commercial value .445 .257 .485 .158 -.344
Scores for species types by utility for subsistence use .367 .518 .281 .014 .142
Scores for species types by utility for woodfuel .453 .439 -.246 -.422 .150
Scores for species types by utility for environmental services .626 .558 -.169 -.037 .533
Scores for species types by utility for fence .654 .456 .308 -.048 -.591
Scores for species types by utility other than above .823 -.020 .017 .881 .215
Total Variance Explained (% of variance) 18.48 16.95 11.82 11.64
Culminative (%) 18.48 35.43 47.25 58.89
Agro-ecological conditions have a determining effect on the agroforestry patterns. The sub-humid zone has higher likelihood to adopt HVAF. In contrast, FMNR in the form of deliberately protecting trees and shrubs scattered in farm is highly associated with the semi-arid zone, probably because harsh climatic conditions may discourage farmers from investing in active tree planting provided low survival rates of seedlings. The significant positive sign of the mid-highland dummy, which applies to two out of four sites in the sub-humid zone, for FMNR, however indicates that FMNR is also practiced in the sub-humid zone especially in the mid-highland parts.
Trees on farm patterns have rather great variability among households even within the same agroecological zones. Controlling zonal factors, socio-economic and institutional factors are observed to have in general contrasting effects on the adoption of FMNR and HVAF.
Labor availability, especially the proportion of male labor, and land availability, are significant factors for FMNR but not constraints for HVAF, while the gender and education of the household heads are found not significant for both. As FMNR requires extensive management of trees on farm though occasionally pollarding and lopping, a large land size with a large number of family labor may provide an advantage. In turn, HVAF which is associated with homestead and boundaries as niches can be managed without being constrained by land and family labor sizes.
Land access conditions and management/policy experiences also affected the likelihoods of FMNR and HVAF differently. Land ownership provides positive incentives to adopt HVAF, even without formal certificates. In turn, the ownership does not significantly affect FMNR, which still has a positive association with famer’s experience of tenure upgrading. It is interesting to note that HVAF had negative association with communal grazing, while FMNR had positive. Thus promoting intensive agroforestry of actively planting tree species for timber, fruit and income crops require not only favorable agroecological conditions, but also institutional/policy arrangements to set up physical fencing or institutional arrangement of social fencing to protect planted young trees.
Asset/income strategies again have contrasting patterns of associations with FMNR and HVAF. HVAF is associated positively with farm asset value, as well as tree-based income and regular business off-farm income while negatively associated with animal farm income, roots and tuber farm income and casual off-farm income. This implies intensive agroforestry requires some livelihood capitals to invest in tree planting while somehow to divert their livelihood strategies from traditional livestock or low-paying casual/farming activities. In contrast FMNR has no significant associations with assets, but positive association with cereal and pulse farm income and rent off-farm income, thus is consistent with traditional land-based livelihoods based on staples.
While accesses to training and training centres were found positive for FMNR and HVAF, there seems room for improving extension modes for promoting agroforestry adoption, as attendance to community meetings had negative associations with both, and field day negative for FMNR. It is rather odd to find that longer distances to markets, roads and services are found positive for the adoption not only for FMNR of subsistence nature but also HVAF. Though its significant positive association with tree-based income, opportunities of HVAF may not have been fully exploited in the sub-humid sites which were in general rather isolated from market and infrastructure accesses during the time of the survey. Thus the result might not necessarily mean that improved market access would discourage the adoption of HVAF.
Table 2.4.4: Regression analysis result
Component 1-FMNR
Component 2-HVAF
B Std. Error
B Std. Error
(Constant) -.877 .277 *** .247 .261
Zonal factors
Zone dummy (semi-arid=1, sub-humid=0) .432 .112 *** -.538 .106 *** Mid-highland dummy (mid-highland=1, otherwise=0) .375 .132 *** .165 .124
Household composition
Head gender dummy (male=1, female=0) -.187 .170
-.076 .160
Head education level -.045 .072
.044 .068
Total family size .038 .020 * -.009 .019
Male ratio .624 .284 ** .190 .267
Working age member ratio .421 .252 * -.041 .237
Land/parcel access, characteristics, policy experiences
Total land size .121 .052 ** -.053 .049
Parcel fragmentation (Simpson index) .036 .050
.017 .047
Mean distance of parcels from homesteads .069 .053
.074 .050
Proportion of parcels owned with certificate .236 .274
.419 .259
Proportion of parcels owned wit yet certificate issued .272 .206
.368 .194 * Proportion of parcels rented from others .084 .185
.185 .175
Proportion of parcels on sloped/steep land .099 .049 ** .018 .046
Experience of free communal livestock grazing .242 .048 *** -.111 .046 ** Experience of land resettlement -.062 .040
-.056 .038
Experience of tenure upgrading .118 .050 ** -.011 .047
Asset/income diversification
Livestock - local animals .056 .056
-.017 .052
Livestock - exotic animals -.075 .050
-.048 .047
Asset - domestic asset value -.019 .044
.012 .041
Asset - communication asset value -.079 .053
-.058 .050
Asset - transport asset value .051 .048
-.139 .045 *** Asset - farm asset value .000 .060
.361 .057 *** Farm income- cereal & pulse .112 .050 ** .006 .047
Farm income - cash crops -.012 .051
-.021 .048
Farm income - roots & tubers .023 .046
-.156 .044 *** Farm income - animal .072 .050
-.082 .047 * Farm income - tree-based .006 .050
.165 .047 *** Off-farm income - regular business .049 .047
.102 .044 ** Off-farm income - casual -.055 .050
-.103 .047 ** Off-farm income - remmitance & gift .014 .044
-.042 .041
Off-farm income - loan .044 .045
-.108 .042 ** Off-farm income - rent .143 .047 *** .052 .044
Access to extension services, markets, infrastructure
Participation in field school .062 .062
.043 .058
Participation in training .126 .051 ** -.009 .048
Participation in field day -.094 .053 * .245 .050 *** Participation in field visit -.031 .054
.027 .051
Participation in demonstration farm .026 .052
.011 .049
Participation in interaction -.068 .051
-.008 .048
Participation in community meetings -.116 .049 ** -.142 .046 *** Participation in training centres .121 .059 ** .050 .056
Distances to markets, roads & services .135 .046 *** .077 .043 *
F-value 4.135 .000b
6.837 .000b
Adjusted R Square .219
.343
(Note): *** significant at 1%; ** significant at 5%; * significant at 1%.
3: Match Species and Management Options (1.2)
3.1: Spatially explicit guidelines on initial best-bet tree species and management options (1.2.1)
In both semi-arid and sub-humid sites of Ethiopia, local knowledge studies conducted using AKT5 tool and methodology between 2013 and 2014 elicited spatially-explicit knowledge regarding the smallholder farming systems. The main objectives included: 1) to characterize farming systems and wider landscapes in order to identify drivers of land use/ tree cover change; 2) identify constraints and opportunities for integration of trees on farms; 3) elicit farmers’ knowledge about tree species and their utilities to understand the roles trees play in meeting livelihood and conservation needs; 4) identify farmers’ knowledge and practice of tree management on fields in order to assessing farmers capacity to sustain trees on their fields. Tables below summarize the key results from the semi-arid and sub-humid sites.
Table 3.1.1: Drivers of tree cover/ land use change
East Wollega (sub-humid) East Shewa (semi-arid)
Government’s land redistribution and villagization policies after 1974
Government’s land redistribution and villagization policies after 1974
Massive deforestation due to demand for tree products such as timber and firewood
Massive deforestation due to demand for tree products such as timber and firewood
Government’s introduction of Eucalyptus plantations resulted into drying up of water bodies and change in rainfall from bimodal to unimodal
Drought, which in most cases inhibit tree survival
Unfavourable free grazing policy of livestock management that threaten tree survival
Unfavourable free grazing policy of livestock management that threatens tree survival
Farmers practice of native tree regeneration Farmers practice of native tree regeneration
Decreasing soil fertility due to unsustainable farming practices such as continuous cropping with no fallows
Decreasing soil fertility due to unsustainable farming practices such as continuous cropping with no fallows
Soil erosion due to low tree cover on sloped nature of the terrain, shallow soils and high rainfall
In both sites, government land policies, deforestation to meet the demand for tree products, unfavorable free grazing of livestock management, farmers practice of native tree regenerations, declining soil fertility due to continuous cropping without fallows were mentioned as drivers of land use change. Additionally, drought was identified as an inhibiting factor for tree survival in the semi-arid site, while changing water systems due to introduction of Eucalyptus and soil erosion due to low tree cover in sloped lands were noted in the sub-humid lands.
Farmers prioritized tree utility based on important roles they felt trees could play on their farming systems and landscapes. Tree products include: fruits, other food products, fodder, firewood, charcoal, timber for construction and furniture, farm implements, household tools, medicine, fibre, live fence, dead fence. Services valued by farmers included: shade, soil erosion control, soil fertility improvement through the roots, soil fertility through leaves, apiculture, streambank stabilization, and biodiversity (wildlife) conservation. There was a contrasting observation in the two zones. Commercially valued species for income such as fruit, timber, firewood, coffee, fodder species were more preferred in the sub-humid zones, while species with subsistence needs (fence, farm implements, firewood, ropes/fibre) as well as ecosystem services (shade) were named in the semi-arid zone.
Table 3.1.2: Priority trees that farmers were interested in retaining/ planting
East Wollega (sub-humid) East Shewa (semi-arid)
Income: from fruit trees, Rhamnus prinoides (gesho), timber, firewood, coffee
Live and dead fence- was ranked as the highest priority tree product sought after by farmers, mainly for keeping livestock away from homesteads where also fodder straws were stored
Improved fruit trees- The farmers requested for improved varieties of mangoes, avocadoes, pawpaw and other fruits
Farm implements (Balanites aeyptiaca and C.macrostachyus) and household tools (Faidherbia albida)
Soil fertility improvement trees: legumes as Sesbania sesban, Leucaena leucocephala, Cordia africana, Croton macrostachyus
Ropes/ fibre for construction of houses (Acacia tortilis)
Coffee shade trees: Albizia schimperiana, Acacia abyssinica, Sesbania sesban, Grevillea robusta
Fuelwood: there is acute fuel shortage and farmers result to using cow dung ‘kubet’
Timber for construction: Cordia africana, Grevillea robusta, Eucalyptus camaldulensis
Shade for livestock , crops and farmers
Fodder: Vernonia amygdalina, Sesbania sesban, Croton macrostachyus
Fruits
Fodder for livestock- Leaves, fruits, bark and seedlings
Overall, in the study sites, more native species were encountered compared to exotic species. Farmers practice farmers manage natural regeneration (FMNR) on fields by maintaining priority species naturally, pruning-coppicing and pollarding, while very little planting is done, especially in the semi-arid zone. Tree management, especially pruning frequency depended on various factors namely: frequency of demand for tree products, tree harvesting policy restrictions, tree canopy size, ownership of ploughing land (owned or rented), tree regeneration growth rate after previous pruning, prevailing weather conditions, rate of drought risk anticipation, maturity age/ rate of trees and tree density on the land.
Farmers in the different agroecological zones acknowledge various factors which affect negatively the adoption of trees/agroforestry, as listed in the below Table, while tenure security also in general affect tree management, i.e., farmers who rented land not allowed to prune trees.
Table 3.1.3: Factors affecting tree adoption / success of agroforestry interventions
East Wollega (sub-humid) East Shewa (semi-arid)
Livestock browsing due to the open grazing system
Livestock browsing
Land shortage Drought
Lack of germplasm Land shortage
Shortage of labour Hired labour- hired labourers are not keen to assist in regenerating trees e.g., they sweep away tree seeds/ seedlings for ease of ploughing
Low knowledge on: tree utilities, tree-crop interaction, tree management
Lack of tree germplasm
Low seedling and trees survival rate due to water scarcity
Theft and destruction by children
Lack of interest because tree a long time to mature and due to insecurity of tenure
Diseases: Mango disease
Low survival rate of trees such as apples (due to low knowledge of farmers on ecological suitability)
3.2: Burundi vegetation map (1.2.2)
A draft of an interactive species selection map has been developed for Burundi, based on an
expansion of a map and application of methodologies developed by an earlier project (the ‘VECEA’
[Vegetation and Climate Change in East Africa] map; URL www.vegetationmap4africa.org ). The
main baseline map selected for Burundi was Pouilloux 1979 (Figure XXa), which was georeferenced
and digitized by the geospatial laboratory at the World Agroforestry Centre (ICRAF). Other baseline
maps included historical soil maps (Van Wambeke 1957) and maps of vegetation belts in western
Burundi (Lewalle 1972; Figure XX.b). Based on a digital elevation model, comparison with
vegetation boundaries of the VECEA map and consultation with national botanical and ecological
experts from Burundi, a final map will be made online by the end of 2014.
(a)
(b)
Figure 3.2.1. Some of the baseline vegetation maps that were consulted when developing the
interactive vegetation map. a: vegetation map developed by Pouilloux for the Atlas of Burundi
(1979); b: vegetation belts described by Lewalle (1972) in eastern Burundi.
Draft documentation sections have been compiled for each of the vegetation types, including lists
of species documented to occur within the vegetation type and a shortlist of characteristic species
for the vegetation type (these characteristic species are used to investigate the correspondences
between the national vegetation types and the regional vegetation types of the VECEA map). The
main reference for vegetation descriptions and species assemblages has been Lewalle (1972) who
provided detailed descriptions and phytosociological tables for vegetation types encountered in
Burundi. However, as Lewalle (1972) only describes the western part of Burundi, other literature
was consulted to expand vegetation types and species lists, especially focusing on references that
documented the eastern part of the country (e.g. Reekmans [1981] describing Miombo woodland
types; or Masharabu [2011] describing floristic composition of Ruvubu national park, the main
national protected area in eastern Burundi). The documentation of the baseline map (Pouilloux
1979) was also consulted and harmonized with vegetation lists and descriptions, as well as a
report on biodiversity strategies for Burundi prepared by the Institut National pour l’
Environnement et la Conservation de la Nature (INECN) in 2000.
In parallel to developing the documentation and species assemblages for the vegetation map, a
master list of 2367 plant species known to occur in Burundi was obtained by compiling information
from Lewalle (1972 ; 1714 current species names), the List of East African Plants (LEAP; Knox &
Vanden Berghe 1996; 895 species) and species listed to occur in Rububu national park (Masharabu
2011; 506 species). 545 of these species were listed to occur in specific vegetation types and have
been included into ‘species selector’ Excel sheets for each vegetation type, including 256 ‘useful
tree species’ compiled for useful tree species interactive maps for eastern Africa (the VECEA map;
URL www.vegetationmap4africa.org) or continental africa (URL
www.worldagroforestrycentre.org/our_products/databases/useful-tree-species-africa ). A
standardized nomenclature was applied to the species lists by combining information from the
Taxonomic Name Resolution Service (http://tnrs.iplantcollaborative.org/TNRSapp.html ) with
information from the African Flowering Plants database (URL http://www.ville-
ge.ch/musinfo/bd/cjb/africa/recherche.php?langue=an ), thereby allowing full referencing
between species lists and tables with documented uses and environmental services of tree species
and full referencing with the Agroforestry Species Switchboard
(http://www.worldagroforestry.org/products/switchboard/index.php/name_like/melia%20volken
sii ).
Currently the vegetation classification system obtained for Burundi is re-confirmed with the VECEA
regional classification system based on documented correspondences from literature (for example
White [1983] referring to Lewalle [1972] as reference for Afromontane rain forest, Lake Victoria
transitional rain forest or Lake Victoria scrub forest), lists of indicator species for national and
regional vegetation types and an analysis of floristic distance. Once the correspondence between
vegetation typologies has been re-confirmed, species known to occur in Burundi (i.e. the
masterlist of 2367 species) and also known to occur in particular vegetation types in other
countries, but that were not listed to the same vegetation type in Burundi will be allocated to the
Burundi vegetation types, using a methodology developed for the VECEA map. This methodology
is particularly useful for the eastern part of Burundi where available references listed only a
relatively small subset of characteristic species. In 2015, the Burundi map will be fully integrated
into the VECEA map and its species selection and distribution tools made available from URL
www.vegetationmap4africa.org
3.3: Decision support tools for matching species and management options to sites and circumstances (1.2.3)
Despite differences in the methodologies and intensity of investigation, there is significant consistency in findings from the local knowledge and tree inventory studies as well as the biophysical and socio-economic surveys. The consolidated recommendation from these findings therefore enable the ACIAR team to reasonably come up with interventions relevant to and customized to local conditions and circumstances of the study sites as summarized in Tables 3.3.1 and 3.3.2.
Across the semi-arid/sub-humid sites, the tree species dominantly adopted by the majority of the households are heterogeneous across the semi-arid sites possibly reflecting fine-scale variations of suitable species. Still their tree-crop-livestock management systems are more or less similar within the agro-ecological zones.
In the semi-arid zone, FMNR is a dominant form through protection of naturally regenerated trees on farm. While farmers rarely adopt intensive agroforestry, farmers derive important utilities from these tree species, esp. for fuel, subsistence needs (construction materials), fence, as well as ecosystem services (especially shade, also important for windbreak, soil fertility and erosion control). Promoting agroforestry there requires technical interventions to improve the survival of seedlings such as rainwater harvesting as well as institutional management of free grazing through physical or social fencing. Marketing potentials and logistics of tree-based products need to well explored.
In the sub-humid sites, HVAF through raising seedlings and planting trees for utilities such as timber and fruits are practiced especially with secure land rights, while FMNR is also practiced especially in mid-highlands and sloped plots. Farmers practicing HVAF has association with higher tree-based income and farming assets. In turn, due to limited access to markets and infrastructure than those counterparts in the semi-arid zones, proportionally more populations in the semi-humid zone are food insecure. There is need to improve enabling institutional/policy environment for promoting agroforestry.
Table 3.3.1: Matching species and management options for semi-arid sites
Bofa Mojo Alem Tena Meki Zeway
Existing tree species
Acacia t, Zyzyphus, Acacia s
Faidherbia a, Acacia t/a, Balanites a.
Acacia t, Croton m, Faidherbia a, Zizyphus spp.
Acacia t, Feiderbia a, Croton m.
Ajo_loc, Acacia t, Balanites a, Acacia s.
Crop tef-maize-sorghum, oil crop
Teff, wheat Teff, wheat, maize
maize-tef-wheat, vegetables
maize-bean-wheat
Management protection of naturally regenerated trees on farm (FMNR) under the arrangement of communal grazing of livestock
Opportunities *access to infrastructure/tarmac roads which connects to Addis Ababa
Constraints to be overcome
*low survival of trees due to drought and communal grazing
* lack of agroforestry marketing experiences
Table 3.3.2: Matching species and management options for sub-humid sites
Arjo Anno Bako Uke
Existing tree species
Croton m, Cordia a, Eucalyptus, Albizia
Cordia a, Croton m, Eucalyptus, mango
Cordia a, Eucalyptus, Croton m, Veronia, mango
Cordia a, Veronia, Eucalyptus, Syzium, mango
Crop Sorghum-Maize, tef, wheat, oil crop (nug)
Maize, tef, sorghum, finger millet, nug
Maize, tef, sorghum, nug, sugarcane
Maize-sorghum, nug
Management HVAF + some FMNR on slopes
HVAF + some FMNR on slopes
HVAF HVAF, less AF experience
Opportunities *relatively favorite climates
Constraints to be overcome
*market potentials of HVAF yet fully exploited to improve income/assets
*quality germplasm/AF knowledge lacking to control pest/disease
In turn, It is interesting to note, for example in the sub-humid zone, that some farmers consider a specific tree species, such as Cordia Africana, most dominant tree species there, as primarily a timber species to earn income and as secondarily a fence/shade thus plant in homesteads or along external boundaries, while others primarily see it as a fuelwood species for subsistence but also as environmental services, thus protect its naturally regenerated trees on farm. In either way, the
tree contributes to enhancing livelihoods and food security, yet preferred management modes and intensities vary depend on farmers’ perceptions and preference. This indicates the critical importance of understanding local knowledge and farmers’ perceptions to make fine-scale recommendations of optimal mixes of species and management options rather than a wide-scale promotion of a few iconic agroforestry practices.
4: Target and Priority Sites and Farmer Circumstances (1.3)
4.1: National dialogues on priotisation criteria (1.3.1)
A larger proportion of agricultural landscapes face multiple challenges related to increasing agricultural production, food security, poverty, climate change, ecosystem degradation and biodiversity loss. These challenges are highly interlinked and exacerbate the pressure on scarce natural resources. Most of these challenges are linked to existing or not sufficient policies, therefore creating a need of coming up with approaches that support multifunctional benefits and multistakeholder engagement processes. The fundamental challenge that remains, given the intricate integrated landscapes, is how to combine the various efforts addressing divergent interest through a more harmonized and collective engagement of the various actors as a means of scaling up initiatives. In essence, for any landscape level initiative to be extensive, it is necessary that both the functional and institutional needs are clearly articulated and prioritized. There is also need to recognize the multiplicity of smallholder landowners across landscapes that dictate multiple interests and needs.
The Australian Centre for International Agricultural Research (ACIAR) Project has been working on enhancing the food security of resource poor rural farmers in Eastern Africa through research that underpins national programmes to scale up the use of trees within farming systems. This includes customizing tools methods and approaches that are ‘best fit’ to the specific contexts for the eastern African region. A crucial step in this regard is defining a process for identifying and facilitating an enabling environment. It also includes the recognition of pertinent policies and institutional arrangements that inhibit the adoption of trees on farm. In identifying these challenges, the ACIAR project in partnership with the Strengthening Rural Institutions (SRI) project began a process to understand the policy and institutional frameworks in the two focus countries of the project in Ethiopia and Rwanda through the Trees for Food Security (TFS) initiative. The TFS initiative seeks to enhance food security for resource-poor rural people in Eastern Africa through research that underpins national programmes.
The format of the participatory workshop was designed on the premise of the scalability of the benefits from the Woreda/ district or site level to the national level. Further, it was with the recognition of constraining policy factors that inhibit the flow of benefits to the various categories of society. This scenario was visually articulated through an hourglass (Figure 4.1.1) whereby the multiplier effects of increased adoption of trees on farm could be realized at Landscape level resulting to streams of benefits to a large number of households. In this regard deliberate steps are articulated to motivate an institutional arrangement that provide enabling environment for addressing constraints that inhibit the scalability of trees on farm.
Figure 4.1.1: Policy scenario
The workshop design is based on a policy dialogue tool and process, which explore the household benefits of having trees on farm, such as food supply, energy supply, shade, fodder for livestock amongst other potential environmental benefits. It also provides a forum for identification of trade offs and opportunity for brokerage. The main objective in the policy dialogue process is to identify strategies for the scalability of benefits and ensuring win-win outcomes such as improved livelihoods, land rehabilitation, water supply systems, carbon sequestration, conservation and capitalization of ecosystem services. The key areas identified from the district policy dialogues were escalated to the national level discussion thereby establishing their relevancy in inhibiting the scaling up adoption of trees on farm. In the case of the ACIAR project in both Ethiopia and Rwanda, initial baselines undertaken formed the basis for the biophysical, socio-economic and institutional barriers to adoption of trees on farm. The baselines also informed the development of effective strategies for scaling up/ out the packages of agroforestry technologies. Table 4.1.1 shows the series of deliverables provided through the use of the policy dialogue tool.
Table 4.1.1: List of workshop reports for policy dialogues in Ethiopia
OUTPUT Title
Workshop report District level policy dialogue Meeting Ethiopia
Workshop report National level policy dialogue Ethiopia
Policy dialogue Process Tool Facilitating policy dialogues
Various issues inhibiting adoption of trees on farm were identified and prioritized based on the level of influence. The prioritized issues in Ethiopia include: market access (opportunities); water stress; grazing management; Allelopathic (knowledge/ awareness; FMNR; land certification; pests and disease. Policy recommendations were also identified in order to facilitate the adoption of trees on farm as summarized in the figure below.
Landscape Benefits
Household
Benefits
Barrier
s
Policy
Response
Figure 4.1.2: Key policy findings (Ethiopia)
4.2: Frameworks for priotising landscapes and farming systems for AF interventions (1.3.2)
The findings/recommendations of the baseline surveys are referenced to the vegetation map to guide the priotization of landscapes and farming systems for agroforestry interventions.
Figure 4.2.1 indicates that the semi-arid sites falls mostly in one vegetation zone category – “We” (Upland Acacia wooded grassland) typical of the central rift valley. Therefore, considering the species variations within the agroecology/vegetation zone, the central rift valley region and its farming system could be priotised for AF interventions.
On the other hand, the sub-humid sites are more diverse in vegetation types – Wcd (Dry
combretum wooded grassland) for lowland, FA (Afromontane rain forest) for highlands, while
mixed vegetation types for transition areas reflecting diversity of tree species suitability due to
rugged landscapes (Fb: Afromontane undifferentiated forest; wd:Edaphic wooded grassland on
drainage-impeded or seasonally flooded soils; and we: Upland Acacia wooded grassland). Still
these vegetation types cover most parts of western Ethiopia. Therefore, considering potential
variations in socio-economic and institutional conditions, the recommendations derived from the
finding from this baseline could be quite applicable to large number of smallholders in western
sub-humid parts of Ethiopia.
Figure 4.2.1: Vegetation map of Ethiopia, with locations of study sites
5: Conclusions and recommendations
Ethiopia has faced food security concern due to severe soil degradation of agricultural landscape under population growth and deforestation trends coupled with minimal application of organic and mineral fertilizers by resource poor farmers. In turn, the system of scattered trees on farm, as a dominant landscape feature in Ethiopia, has been long recognized for its multi-dimensionality to provide not only ecosystem goods and but also services to enhance system productivity and resilience which contribute to food security. Scattered trees on farm not only provide fuel, fruits, timber, etc., for direct consumption and income for farm households but also serve as local nutrient reserves to modify micro-macro climates as well as soil properties, such as soil organic carbon, N, and P in agricultural landscapes which are otherwise uniform, thus to increase grain yields under tree canopies.
Despite this recognition, multi-dimensionality of utilities from farmers’ perspective has never been formally studied. The baseline confirmed that FMNR is a dominant agroforestry system not only in the semi-arid zone but also in the sub-humid zone, and consistent with supporting subsistence land-/staple-crop based livelihoods through provision of ecosystem goods and services. In contrast, HVAF associated with tree-based farm income, assets and off-farm enterprises is practiced more in the sub-humid zone.
Agroecological conditions however are not the only determinants to promote these agroforestry systems. Policy/institutional arrangements to ensure tenure security and collective actions to handle externalities of communal grazing are critical. At the same time there is room for improving modes of extension services and the role of markets to promote agroforestry thus to enhance tree-based incomes and enterprises.
The success of scaling up and out of trees on farm systems throughout agricultural landscapes is dependent on understanding heterogeneous patterns of trees on farm in fine-scale local specific contexts. Optimal tree species and management intensities of crop-tree mixes are of course differ by agroecological zones but even within agroecological zones. These patterns are not solely defined by ecological conditions alone but also by farmers’ preferences. For example, some farmers consider a specific tree species as primarily a timber species to earn income and as secondarily a fence/shade thus plant in homesteads or along external boundaries, while others primarily see it as a fuelwood species for subsistence but also as environmental services, thus protect its naturally regenerated trees on farm. In either way, the tree contributes to enhancing livelihoods and food security, yet preferred management modes and intensities vary depend on farmers’ perceptions and preference. This indicates the critical importance of understanding local knowledge and farmers’ perceptions to make fine-scale recommendations of optimal mixes of species and management options rather than a wide-scale promotion of a few iconic agroforestry practices. The success of such approaches to scale up and out enhancing tree covers and diversity across landscape level need to be proven through integrating research in development.
To follow up and to deepen the understanding of the role of trees for food security and livelihoods, we will work with ILRI to see the association between food security indicators and adoption of tree diversity pattern indicators.