Content uploaded by Sven Goenster-Jordan
Author content
All content in this area was uploaded by Sven Goenster-Jordan on May 03, 2017
Content may be subject to copyright.
Effects of transformation processes on plant species richness
and diversity in homegardens of the Nuba Mountains, Sudan
Martin Wiehle •Sven Goenster •Jens Gebauer •
Seifeldin Ali Mohamed •Andreas Buerkert •
Katja Kehlenbeck
Received: 24 December 2013 / Accepted: 15 May 2014 / Published online: 3 June 2014
ÓSpringer Science+Business Media Dordrecht 2014
Abstract Traditional homegardens (HGs) are con-
sidered to harbor high levels of plant diversity and
have been therefore characterized as sustainable agro-
ecosystems suitable for on-farm (incl. circa situm)
conservation of plant genetic resources. While the
functional structure of traditional HGs is poorly
understood specifically for semi-arid and arid regions,
their plant species richness and diversity is increas-
ingly threatened by recent and fast evolving agricul-
tural transformation processes. This has been
particularly claimed for traditional jubraka HG sys-
tems of Sudan. Therefore, sixty-one HGs in four
villages of the Nuba Mountains, South-Kordofan
Province, Sudan, were randomly selected, geograph-
ically recorded and plant richness and abundance
determined and plant diversity parameters calculated.
In addition, socio-economic household data were
assessed by interviews and soil samples taken to allow
a comprehensive analysis of putative factors affecting
HG plant diversity across different villages, levels of
commercialization and plant species composition
based clusters. A total of 110 species from 35 plant
families were grown in the HGs along with 71
ornamentals. Perennial species accounted for 57 %
including 12 indigenous fruit tree (IFT) species and six
exotic fruit tree species. Mean species richness of
useful plant species (excluding ornamentals) per HG
was 23 (range 6–46). On average, 41 % of the 23
species per HG were of exotic origin, however, with a
large range (21–83 %) among locations. Mean diver-
sity and evenness indices were 1.46 (range 0.49–2.42)
and 0.48 (0.15–0.87), respectively. The level of
commercialization of HGs only marginally affected
species diversity measures although the species rich-
ness was significantly higher for commercial than
subsistence HGs. Species richness was higher on
lower (6.6–7.2) pH soils. IFT richness was highly
variable, but non-significantly different across the four
locations. Plant species richness and diversity was
high in comparison with other HG systems in semi-
arid regions. Cluster analysis was found to be a
Electronic supplementary material The online version of
this article (doi:10.1007/s10457-014-9717-2) contains supple-
mentary material, which is available to authorized users.
M. Wiehle S. Goenster A. Buerkert (&)
Organic Plant Production and Agroecosystems Research
in the Tropics and Subtropics (OPATS), University of
Kassel, Steinstr. 19, 37213 Witzenhausen, Germany
e-mail: tropcrops@uni-kassel.de
J. Gebauer
Sustainable Agricultural Production Systems with Special
Focus on Horticulture, Faculty of Life Sciences, Rhine-
Waal University of Applied Sciences, Marie-Curie-Straße
1, 47533 Kleve, Germany
S. A. Mohamed
Department of Horticulture, University of Khartoum,
P.O. Box 321, Shambat, Khartoum North, Sudan
K. Kehlenbeck
Tree Diversity, Domestication and Delivery, World
Agroforestry Centre (ICRAF), United Nations Avenue,
Gigiri, P.O. Box 30677, Nairobi 00100, Kenya
123
Agroforest Syst (2014) 88:539–562
DOI 10.1007/s10457-014-9717-2
valuable tool to classify HGs and to extract homoge-
neous HG types with low, intermediate and high
richness and diversity. In addition, the share of exotic
and ornamental species in HGs indicated a trend
towards the loss of traditional farming practices,
particularly in areas with good market access. The data
did not indicate the hypothesized loss of inter-specific
diversity due to commercialization and species rich-
ness was numerically even higher for market-oriented
HGs compared to subsistence ones.
Keywords Agroforestry Circa situm
conservation Commercialization Jubraka
Shannon index Subsistence gardening
Introduction
Traditional homegardens (HG) are known to harbor
high levels of plant diversity and have therefore been
claimed to play a pivotal role in circa situm (i.e. on-
farm) conservation of plant genetic resources (Atta-
Krah et al. 2004; Bardhan et al. 2012; Galluzzi et al.
2010; Hughes 1998). Structurally complex and species
diverse agro-ecosystems are reported to reduce the
risk of total crop failure, increase the utilization of
limited resources and provide several ecological
service functions (Abdoellah et al. 2006; Eyzaguirre
and Linares 2004; Vlkova et al. 2011). Multi-strata
vegetation structures in HGs are particularly important
in hot, semi-arid regions where they provide shade for
understory plants (Blanckaert et al. 2004) and may
protect soils from degradation, leaching, and erosion
during the rainy season (Soemarwoto et al. 1985).
Particularly in rural settings diverse HGs can enhance
a family’s nutritional status and food security by
producing a range of fruits, vegetables, spices, med-
icine, forage and fuel (Kabir and Webb 2009; Maroyi
2009; Sunwar et al. 2006). In addition, surplus produce
may be sold to contribute to the family’s cash income
(Abdoellah et al. 2006; Maroyi 2009; Mendez et al.
2001). Despite their importance throughout the tropics
and subtropics (Fernandes and Nair 1986; Soemarw-
oto 1987), the functional structure of traditional HGs is
poorly understood (Pandey et al. 2006) specifically for
semi-arid and arid regions (Bernholt et al. 2009).
This holds also true for the traditional jubraka
agroforestry systems in the Nuba Mountains of Sudan
that are an important source of food and partly income
for local communities throughout the year, but partic-
ularly at the end of the dry and onset of the rainy
season, the so called ‘hungry periods’ (Obeidalla and
Riley 1984). Jubraka represent the most common type
of HGs within the small-scale farming systems in the
semi-arid zone of Sudan and they are distributed from
Darfur in the western part of the country to the South
Kordofan province in the south (Elsiddig 2007;
Harragin 2003). These agroforestry HGs are also seen
as controlled/protected habitats for first domestication
efforts of wild species, including indigenous fruit trees
(IFTs) such as Ziziphus spina-cristi and Adanso-
nia digitata (Wiehle et al. 2014; Wiehle et al.
submitted).
After decades of civil war which hindered eco-
nomic and agricultural development, rapid transfor-
mation processes arose with the Comprehensive Peace
Agreement between the warring parties in 2005. This
has allowed the opening of regional markets, the
development of infrastructure, an influx of external
inputs for agriculture including plant genetic resources
and easy access to comparatively cheap imported food
(USAID 2011). The introduction of exotic species
(including ornamentals) and improved varieties e.g. of
vegetable species into this region started in the late
nineteenth century (Bedri 1984; Mahmoud et al. 1996)
and is still on-going. However, reliable data and
historical documentation is lacking and remain par-
ticularly vague for the province of South Kordofan
(Abdalla 2007). These changes may affect richness
and diversity of useful plants cultivated in agro-
ecosystems including HGs to different extents and
complexity as described by Shackleton et al. (2008),
Scales and Marsden (2008), and Kabir and Webb
(2009) for similar systems in Africa and Asia. The
frequently reported reduced agrobiodiversity of com-
mercialized gardens may increase the risk of pest and
disease outbreaks (Abdoellah et al. 2006), may
negatively impact year-round availability of food
products for subsistence farming and imply a reduced
resilience to match the challenges of changing human
demands (Atta-Krah et al. 2004) and climate change
(Albrecht and Kandji 2003).
Transformation processes in HGs are often highly
time- and region-specific (Kehlenbeck et al. 2007) and
have been described as reflecting intensification
(Scales and Marsden 2008), homogenization (Peyre
et al. 2006), commercialization of production (Abdo-
ellah et al. 2006; El Tahir and Gebauer 2004), and
540 Agroforest Syst (2014) 88:539–562
123
urbanization (Kehlenbeck et al. 2007). Proximity to
markets for instance can strongly affect species
richness in HGs (Christanty et al. 1986), whereby
richness and abundance of useful plants can be both
enhanced and hampered for different crop groups
(Kehlenbeck et al. 2007; Mendez et al. 2001; Wezel
and Ohl 2005). The wealth status of gardeners’
families and duration of HGs being used for cultiva-
tion were important determinants for diversity patterns
in Ethiopian HGs (Coomes and Ban 2004; Tolera et al.
2008).
Using the Nuba Mountains as a model zone for
transformation processes in HGs of East Africa the
aim of the present study was (i) to analyze plant
species richness and diversity in HGs, focusing
particularly on exotic and IFTs and their role for food
and nutrition security of the gardeners’ families.
Further objectives were (ii) to determine socio-
economic and bio-physical factors affecting plant
species richness and diversity and (iii) to evaluate the
suitability of HGs for circa situm conservation
purposes of plant genetic resources, particularly of
IFT species.
Materials and methods
Natural environment and socio-economic
characteristics of the research area
The study was conducted between June and October
2010 in the Nuba Mountains, South-Kordofan Prov-
ince, Sudan, ranging from 11°570N, 29°430Eto
10°500N, 30°590E (Fig. 1). The region belongs to the
semi-arid Sudano-Sahelian climate zone and receives
a mean annual precipitation of 500–800 mm decreas-
ing from the south to the north (Bedigian and Harlan
1983). Rainfall is distributed uni-modally from May to
October with a pronounced inter-annual variation.
Three climatic periods exist within 1 year: the cold dry
season from November to February (no precipitation),
hot and dry conditions from March to April (no
Fig. 1 Hill shaded map of
the research area in the Nuba
Mountains, Sudan (2010)
with the locations of the four
surveyed villages as well as
mentioned settlements in
text. Source: modified after
Centre for Development and
Environment (CDE),
University of Bern,
Switzerland (2005)
Agroforest Syst (2014) 88:539–562 541
123
precipitation), and the cooler rainy season from May
to October. The mean annual temperature is 30 °C
ranging from 31 °C in April to 24 °C in January
(Ismail and Elsheikh 2007).
Between the mountain ranges, large plains, which
are characterized by deep vertisols (the so called
‘black cotton soils’). Settlements are scattered along
the drained piedmonts, locally called ‘gardud’, on
shallow, sandy, weathered granitic soils classified as
Ustalf (United States Soil Taxonomy). These are also
used to establish HGs where staples [Sorghum bicolor
(L.), Pennisetum glaucum (L.) R. Br., Sesamum
indicum L.], vegetables and fruits for daily consump-
tion and/or to gain cash income are cultivated. The
natural vegetation consists of a woodland savannah
dominated by tall grasses (Antropogon spp.), Acacias
(Acacia spp.) and Balanites aegyptiaca trees.
Ethnically the Nuba Mountains are characterized
by diverse tribal communities with vague early history
(Bedigian and Harlan 1983), which can be classified
into three main groups: the Nuba (likely prehistoric
inhabitants, small groups of diverse origins), the Arabs
(originated from North Africa, settled and partly
mixed with Nuba tribes), and other African tribes
(mainly pastoralists) migrated to the area from Central
and West Africa.
Data collection
Four villages—Kauda, Kalogi, Habila and Sama—
were chosen along gradients of rainfall, altitudes,
ethnicity and accesses to main markets (Table 1;
Fig. 1). Sixty-one households with a HG [15 per
village, except of Sama (n =16)] were selected
within an area previously circulated and mapped with
a handheld GPS device (Vista HCx eTrex, accuracy
±2 m, GARMIN
Ò
Ltd., Ireland). To select a HG, a
previously GIS-generated random point was visited
and the nearest HG within a 30 m radius from that
random point was identified. The household head
managing the identified HG was visited and asked for
permission to conduct the survey. In case these
conditions were not met, the next random point was
taken and the occurrence of a HG within the given
radius evaluated.
Geographic location, altitude, HG size and size of
cultivable area of each HG were determined by GPS
and measuring tapes. Basic socio-economic farm and
household data such as total farm size, household
possessions, including livestock [calculated as tropical
livestock unit (TLU)], number of household members
and their ages, education level, ethnic affiliation and
main occupation of the gardener as well as information
about the age and management of the HG and the
proportion of sold produce from HGs was gathered
through individual interviews with the household head
and the household member mainly responsible for
gardening using a semi-structured standardized ques-
tionnaire modified from Kehlenbeck et al. (2007). All
questions and replies were translated from English
into Arabic or local languages and vice versa by a
native bi-lingual assistant. HGs of those households
selling any HG produce were considered as ‘market-
oriented’, and as ‘subsistence-oriented’ if no produce
was sold.
Soil characteristics were determined in all HGs
after collecting topsoil samples (0–20 cm) from the
vegetable and cereal plots of each HG. In each
separate plot, three sub-samples of 100 cm
3
each were
taken from randomly chosen points with a 5 cm inner
diameter steel tube, bulked, and about 150 g air-dried
for analysis after sieving to\2 mm. Effective cation-
exchange-capacity (CEC
eff
) and exchangeable alu-
minium (Al
3?
), calcium (Ca
2?
), potassium (K
?
),
sodium (Na
?
), magnesium (Mg
2?
) as well as available
P (Bray-P1), organic carbon (C
org
) and total nitrogen
(N
total
) were determined by standard methods (Houba
et al. 1995; van Reeuwijk 1993) at the Charles Renard
Analytical Laboratory, ICRISAT, Niamey, Niger.
Extractable Al was below detection limits and thus
not further considered in analyses. Soil pH (1:2.5,
0.01 M KCl) was determined by a pH-meter (WTW
GmbH, Weinheim, Germany). Except for Na
?
, the
measurement of all parameters revealed non-signifi-
cant differences between the two plot types (vegetable
and cereal plots). Thus, results of the two plots were
averaged to obtain one value per HG for each of the
soil parameters.
Most soil quality parameters differed significantly
among villages (Table 2), but not between the type of
plots (except Na which showed higher contents on
vegetable plots). Mean soil pH per village varied from
6.6 to 7.3 (Table 2). Bray-P levels were three times
lower in Habila than in Sama. Concentrations of Ca,
Mg and CEC
eff
were highest in Habila, while K was
highest in Sama. N
total
was highest in Kauda, while
C
org
contents did not differ among villages. When
comparing soil quality parameters of subsistence and
542 Agroforest Syst (2014) 88:539–562
123
market-oriented HGs, N was higher and pH lower in
soils of subsistence HGs (Table 2).
In each HG a botanical inventory was conducted.
Scientific names of the species, their potential uses,
particularly as food and geographical origins (indig-
enous to the study area or exotic) were determined
using various field guides (Andrews 1948; Bebawi and
Neugebohrn 1991; Braun et al. 1991; El Amin 1990).
The occurrence, local name, abundance and use of
each individual plant species were recorded, excluding
species regarded as ‘not useful’ or ‘weeds’ by the
respective gardener. We are well aware that terms
Table 1 Geographical and general socio-economic characteristics of the four surveyed villages in the Nuba Mountains, Sudan
(2010)
Coordinates Elevation
(m)
Precipitation
rank
(600–800 mm)
Distance
to next
city (km)
Main
ethnic
group
No. of
inhabitants
(estimated)
Socio-economic
characteristics
NE
Sama 10°5803400 29°4402000 514 Medium 5 Shawabna
(Arab)
13,000 Old settling area, close to
main market and
administrative unit
Kadugli
Habila 11°5700900 30°0100700 662 Low 40 Tama
(other
African)
12,000 Youngest village of
surveyed locations, large
agricultural schemes,
village on a vertisol soil
Kauda 11°0505700 30°3301300 743 High 80 Otoro
(Nuba)
6,000 Most remote village and
very old settling area, in
2010 a runway to Kadugli
has been opened
Kalogi 10°5102200 30°5805400 511 Medium 50 Hawasma
(Arab)
14,000 Second most remote
village, established
relatively recently by
settled nomadic tribes
Table 2 Mean soil quality parameters in 61 homegardens (HGs) surveyed in four villages, in the Nuba Mountains, Sudan (2010)
arranged by village, economic orientation of HG production and cluster affiliation
npH Na
?
K
?
Ca
2?
Mg
2?
CEC
eff
Bray-P
(mg kg
-1
)
C
org
(%) N
total
(%) C/N
Kauda 15 6.6
b
0.2 0.8
b
5.4
b
2.2
a
8.7
b
34
ab
1.0 0.11
a
9.9
a
Kalogi 15 7.2
a
0.2 1.2
ab
5.7
ab
1.4
b
8.5
b
60
a
0.9 0.07
b
12.2
b
Habila 15 6.7
b
0.3 1.0
ab
6.2
a
2.3
a
9.8
a
23
b
1.1 0.09
ab
12.3
b
Sama 16 6.9
ab
0.2 1.4
a
5.4
b
1.3
b
8.4
b
69
a
0.9 0.08
b
11.4
b
P <0.001 0.122 0.021 0.013 <0.001 <0.001 0.002 0.187 0.009 <0.001
Market-oriented 19 7.0 0.2 1.2 5.8 1.6 8.9 40 0.9 0.07 11.9
Subsistence 42 6.7 0.2 1.0 5.6 1.9 8.8 51 1.0 0.09 11.3
P 0.009 0.202 0.151 0.266 0.092 0.808 0.263 0.089 0.032 0.243
Cluster 1 15 6.6 0.2 0.8 5.4 2.2
a
8.7
ab
34 1.0 0.11
a
9.9
b
Cluster 2 22 6.9 0.2 1.1 5.7 1.7
b
8.7
ab
44 0.9 0.07
b
11.9
a
Cluster 3 11 7.0 0.2 1.1 5.6 1.5
b
8.4
b
77 1.1 0.09
ab
11.9
a
Cluster 4 13 6.9 0.3 1.4 5.9 1.8
ab
9.5
a
42 1.0 0.08
ab
12.3
a
P0.192 0.377 0.067 0.313 0.007 0.036 0.183 0.243 0.010 <0.001
Total 61 6.8 0.2 1.1 5.7 1.8 8.8 47 1.0 0.1 11.5
Different letters behind means in a column and bold P values indicate significant differences at P \0.05 (Mann–Whitney or Kruskal–
Wallis tests, depending on the data structure)
Agroforest Syst (2014) 88:539–562 543
123
such as ‘useful’ and ‘useless’ are location-specific and
subjective terms and interviewees may have mixed up
theoretical knowledge about a species and its practical
current use. Ornamental plants were also recorded, but
skipped for some subsequent analyses because they do
not contribute to food and nutrition security (Sunwar
et al. 2006). According to Bernholt et al. (2009) each
species was grouped into one of the following nine
categories, based on its main use according to the
gardener: fruit, vegetable, stimulant, condiment, med-
icine, staple, wood/multipurpose use (MPU), orna-
mental and ‘other uses’. The group of ‘other uses’
included cosmetics, living fence, fiber, fodder, biofuel,
household articles and insect repellents. For many
species, respondents mentioned several uses, but for
easier analyses, we focused on the ‘main use’ only.
Since many tree species, however, compile several use
categories, we asked also for the secondary uses.
Regarding fruit tree species, we also included those
species as ‘fruits’, which were not mentioned with that
specific use by the respondents, but are assigned as
having edible fruits in the available literature (see
above) to assess and analyze their potential for family
nutrition. Height strata of the vegetation (0–0.99,
1–1.99, 2–4.99, 5–10 and [10 m) were only deter-
mined for tree species.
Data analysis
To test if minimum size of sampled areas was covered,
species area curves were generated by using the Mao
Tao estimator calculated with EstimateS (Colwell
2011). Such curves are especially useful when com-
paring species richness at a fixed number for subsets of
different sample sizes (Kindt and Coe 2005). An
asymptotic stagnation at a certain ordinate value
would be equal to the maximal possible species
richness at one area and shows that enough sites were
sampled.
Species abundance was transformed to individual
density per 1,000 m
2
HG area to balance out effects of
different HG sizes and used for all subsequent analyses.
For the same reason and also based on 1,000 m
2
,a
modified Arrhenius equation was used to determine
species density (Evans et al. 1955). Species richness
and abundance data were used to calculate Shannon–
Weaver diversity index (H0) and Shannon evenness
index (J0) separately for total plant species excluding
ornamentals and for exotic and indigenous fruit tree
(EFT and IFT, respectively) species using the MS
Ò
Excel based Diversity Add-In Calculator (SSC, Read-
ing, UK). To compare the importance of species in
different use categories for the different levels of
commercialisation and obtained clusters, the summed
dominance ratio (SDR) was calculated by using relative
density and frequency of the species per respective
group and then summing up the values within the
respective use category (McCune and Grace 2002).
All data were subjected to statistical analysis using
SPSS
Ò
19.0 for Windows
Ò
(SPSS Inc., Chicago, IL,
USA), whereby the significance level was set to
P\0.05. As the data were not normally distributed,
non-parametric Mann–Whitney or Kruskal–Wallis-
tests were used to compare parameters between two
or more groups, respectively. Wilcoxon signed-rank
tests were performed for same parameters assessed for
different plant categories such as indigenous and exotic
species richness or soil fertility parameters in vegetable
and staple plots of the same HG. Chi square (v
2
)tests
were applied to test nominal and categorical variables.
Three multivariate techniques were applied to
analyze factors affecting species richness, density,
diversity and composition: stepwise multiple regression
analyses, cluster analysis, and stepwise discriminant
analysis. Stepwise multiple regression analyses was
employed to analyze the influence of the socio-
economic and bio-physical independent variables HG
size and age, elevation, locations (as dummy variables),
level of subsistence-oriented production, household
poverty index (HPI; see below) and the soil parameters
pH and CEC
eff
on the dependent variables richness,
species and individual densities, share of exotic indi-
viduals, as well as diversity and J0indices for total plant
species excluding ornamentals. The influence on IFT
species richness was further evaluated based on the
mentioned parameters. To evaluate the relative wealth
of each household, a poverty index (HPI) was obtained
by the method following Henry et al. (2003)basedon
principal component analysis scores. The following
socio-economic parameters determined during the
household interviews, were included: family size,
number of meals in past 2 days, weeks of stock for
food staple, number of rooms per household member,
quality of dwelling floors and walls, value of owned
livestock species, and total value of assets per house-
hold member. The lower the HPI value,the more severe
is the relative poverty of the respective household in
comparison with the entire interviewed households.
544 Agroforest Syst (2014) 88:539–562
123
To characterize HGs based on their species com-
position and to determine relationships among them,
minimum variance (Ward’s method) cluster analysis
with squared Euclidian distance as a measure of
dissimilarity of ln-transformed plant individual den-
sity data was performed with the software MVSP
(Kovach 2001). A preliminary nearest-neighbor pro-
cedure was conducted to test for outliers; the most
likely number of clusters was assessed by means of the
‘elbow-criterion’ (Leyer and Wesche 2007). To iden-
tify the plant species that were most responsible for the
cluster formation and to assess the strength of the
classification model, stepwise DA was performed by
SPSS, which also determined the derived Wilks’
lambda values. High power of discrimination between
groups is denoted by Wilks’ lambda values near zero.
Results
Socio-economic characteristics of the surveyed
households
Most of the interviewed gardeners belonged to the
ethnicity of Arabs (44 %) or Nuba (41 %) and only
15 % to other African tribes. About 26 % of the
respondents migrated to their current village from
other regions of the Nuba Mountains or even further
parts of Sudan. Gardeners’ average age was 39 years
(range 13–81 years); 13 % of the respondents were
Christians, whereas the remaining were Muslims.
While 92 % of the surveyed households were male-
headed, 90 % of HGs were managed by women.
Family size was on average nine persons (range 2–19),
the ratio of children to adults was 0.9 and illiteracy of
family members[14 years was 52 % with significant
differences between villages (Table 3). Sixty-two
percent of the 61 interviewed gardeners had no formal
school education and were illiterate with lowest
numbers of years in school in Kauda (Table 3). The
mean size of the total landholding (HG and additional
fields) was 7.2 ha however with large differences
among villages (smallest in Kauda and largest in
Habila). Eighty-seven percent of the respondents
owned livestock with significantly higher numbers of
TLUs in Kauda (P =0.024; Table 3). In all cases,
livestock was kept outside the HG or in small corals
within. Cattle were mainly led by herdsman to the
surrounding forests and grasslands, while goats, sheep
and chickens were mostly freely roaming around and
within the villages. Pigs, only present in Kauda, were
kept in small shelters inside HGs during the crop
cultivation period, but were roaming around during the
remaining time. HPI was lowest for Kauda and Habila
(Table 3) and positively correlated with gross cash
income from HGs (r =0.353, P =0.006).
When comparing the economic orientation of HG
production, 69 % of the surveyed HGs were subsis-
tence-oriented, and 31 % market-oriented, with no
differences among villages. Although not statistically
significant, market-oriented gardens were compara-
tively more likely managed by men than women
(P =0.069). When comparing subsistence- and mar-
ket-oriented HGs, gardener’s age (data not shown),
literacy rate of family members [14 years, HPI and
income generated from the HG were significantly
higher and gross income lower in market-oriented
HGs (Table 3), while total farm and HG sizes as well
as education level of the gardener did not differ.
Garden characteristics and management
The total area surveyed in the 61 HGs (cultivatable
area) covered 12.1 ha. Mean total HG size was
1,988 m
2
, ranging from 168 to 7,934 m
2
. Kauda had
by far the oldest HGs (P = 0.001), the largest cultiva-
ble HG areas (P \0.001) and most hilly conditions
(P \0.001) with a mean slope inclination of about
seven degree (Table 3). Terraces for erosion control
were found in all of Kauda’s HGs, while only two
times in the remaining HGs. Fences surrounded most
of the surveyed HGs, except for Kauda where fences
were absent. All 61 HGs were owned by the gardeners.
According to the respondents, the main function of
HGs was growing crops for self-consumption (88 %).
Only 12 % of the respondents mentioned market
production as main function. However, market pro-
duction was the most important secondary function,
mentioned by 57 % of the respondents, followed by
self-consumption (30 %) and pastime/recovering
(13 %). In seven percent of all HGs, laborers were
hired for certain tasks, e.g. to establish fences or to
weed the garden. The use of mineral fertilizers
accounted for three percent of the respondents, while
organic fertilizers were used by 41 % of the respon-
dents. Pesticides (as ash or chemicals) were applied in
48 % of the HGs. The most important needs men-
tioned to improve HG production were fencing
Agroforest Syst (2014) 88:539–562 545
123
Table 3 Mean socio-economic parameters in 61 homegardens (HGs) surveyed in four villages, in the Nuba Mountains, Sudan (2010) arranged by village, economic orientation
of HG production and cluster affiliation
n HG characteristics n Social household characteristics n Economic household characteristics
HG
size
(m
2
)
Slope
(%)
Duration of
being used as
HG (years)
No. of
family
members
Level of formal
education in
school (years)
Literacy of the
household
(members
[14 years) (%)
Total
land
holding
(ha)
Tropical livestock
units (TLU) per
family member
House-hold
poverty
index (HPI)
Gross cash
income from
HG (SDP)
Kauda 15 4,644
a
7.3
a
67
a
14 8.5 0.9
b
22
b
14 0.7 0.4
a
-1.0
b
13
Kalogi 15 1,109
b
0.1
b
13
b
15 9.3 3.7
a
62
a
15 2.1 0.4
a
0.4
a
140
Habila 15 1,084
b
0.0
b
14
b
15 8.7 1.9
a
50
ab
15 21.7 0.2
ab
-0.6
b
33
Sama 16 1,168
b
4.1
b
21
b
16 7.3 3.9
a
70
a
16 4.1 0.1
b
1.0
a
178
P <0.001 <0.001 0.001 0.800 0.033 0.001 0.057 0.024 <0.001 0.234
Market-
oriented
19 1,899 1.7 45 19 9.2 2.7 69 19 5.2 0.1 0.5 296
Subsistence 42 1,904 3.5 37 41 8.3 2.6 44 41 8.1 0.4 -0.2 0
P0.510 0.225 0.629 0.278 0.799 0.005 0.110 0.053 0.011 <0.001
Cluster 1 15 4,644
a
7.3
a
67
a
14 8.5 0.9 22
b
14 0.7 0.4
a
-1.0
b
13
b
Cluster 2 22 1,331
b
2.3
b
16
b
22 8.7 2.6 54
ab
22 6.5 0.2
b
0.1
a
36
ab
Cluster 3 11 808
b
1.1
b
18
b
11 7.5 4.4 64
a
11 5.3 0.5
ab
0.9
a
3
b
Cluster 4 13 1,032
b
0.4
b
14
b
13 8.7 3.2 70
a
13 17.0 0.2
ab
0.2
a
356
a
P <0.001 <0.001 0.002 0.887 0.058 <0.001 0.053 0.021 <0.001 0.003
Total 61 1,988 2.9 40 60 8.6 2.7 52 60 7.2 0.3 0.0 94
One of the 15 households in Kauda could not be interviewed resulting in missing information on socio-economic characteristics
Different letters behind means in a column and bold P values indicate significant differences at P \0.05 (Mann–Whitney or Kruskal–Wallis tests, depending on the data
structure). SDP Sudanese pound; 1 SDP equal to 0.324 €(based on the mean exchange rate during the study period between 01 June and 01 October 2010, www.oanda.com,
accessed 13 June 2013)
546 Agroforest Syst (2014) 88:539–562
123
(15 %), extending the cultivation area and soil enrich-
ment by manure (each 10 %), crop rotation and use of
improved seeds/varieties (each 7 %). About 93 % of
the interviewed 61 HG owners claimed to have fertile
soils. When asked about changes over time, 38 % of the
gardeners reported degradation, while 10 % observed
improvements in soil fertility. Twenty percent of all
respondents attributed changes in soil fertility to
changes in rainfall, but no one mentioned that use of
fertilizers may be influential. None of the respondents
ever had contact with governmental or non-govern-
mental agricultural extension services. Planting mate-
rial was mostly obtained through using own seeds from
the previous harvest or by exchange with neighbors.
Only in Kauda gardeners had access to a nursery
(managed by a NGO) where they had purchased
planting material, foremost ornamental trees.
Total plant species richness, diversity, and use
A total of 110 useful plant species from 33 plant
families were grown in the HGs plus 71 ornamental
species. Out of the overall 181 species, 105 (58 %)
were of exotic origin. Sixty-three species (89 %) of the
ornamental species were exotics, markedly higher
than the 42 exotics (38 %) of the useful plant species.
Fifty-seven percent of the 181 species were perennial
species including 12 IFT and six EFT species. Many
other tree species also had edible fruits, although not
of primarily importance, resulting in a total of 23 IFT
and nine EFT species if the secondary use as fruit was
included (Table 4). Most of the 110 useful species
were used as source of wood/MPU (25 %), vegetable
(20 %), fruit (17 %), ‘other uses’ (15 %), and staple
and medicinal (each 8 %). Species with their main
uses as stimulants and condiments (each 3 %) were
negligible. The five most frequent species were
Abelmoschus esculentus (occurring in 95 % of the
surveyed HGs), Zea mays and Solanum lycopersicum
(each 90 %), S. bicolor (89 %) and Cucumis melo ssp.
(84 %). The five most abundant species were S.
bicolor (48 % of all useful plant individuals without
ornamentals), S. indicum (22 %), Arachis hypogaea
(8 %), Z. mays (4 %), and Corchorus fascicularis
(3 %). Species accumulation curves for exotic and
indigenous species based on sampled HGs showed that
the minimum sample size was partly not covered
(Fig. 2). For example, overall indigenous species
numbers would increase slightly if more HGs were
sampled (Fig. 2a). When comparing fruit tree species
accumulation curves total saturation was reached for
EFTs, but not at all for IFTs (Fig. 2b).
All five vegetation strata of woody species were
recorded in the surveyed HGs, showing a continuous
increase of tree abundance from the highest (1.7 %,
[10 m) to the lowest strata (64.8 %; 0–0.99 m).
Regarding fruit trees, the most frequent IFT species
(1st use category, Table 4)wereZ. spina-christi (found
in 61 % of the surveyed HGs), A. digitata (46 %), B.
aegyptiaca (43 %, not present in Kauda), and Sclero-
carya birrea (26 %). The most frequent EFT species
were Phoenix dactylifera (23 %), Annona squamosa
(18 %) and Mangifera indica (16 %). Within the fruit
use group, Z. spina-christi was the most abundant (201
out of a total of 553 fruit tree individuals), followed by
B. aegyptiaca (107) and A. digitata (74).
Plant species richness, density, diversity, and use
among villages
Mean species richness of all observed plant species
was 30 species per HG, with lowest richness in Kauda
(28) and highest one in Sama (33, P =0.001). By
excluding ornamentals, 23 species per HG were found,
(range 6–46), of which 41 % were of exotic origin
(range 21–83 %, Table 5). The richness was lowest in
Kalogi but highest in Kauda, while share of exotic
species was highest in Kalogi, but lowest in Kauda and
Sama (Table 5). Species density was lowest in Kalogi
and similar for the other three villages (Table 5). A
mean of 12,097 individuals of useful plants were
documented per HG (Table 5) in addition to a median
of 100 ornamental plant individuals. Kauda showed
highest abundance of useful species, where also HG
sizes were largest (Table 3). The mean individual
density was 4,137 plants per 1,000 m
2
, of which 46 %
were exotics (Table 5). Individual density was highest
in Kauda and lowest in Kalogi, while the share of
exotic individuals was lowest in Kauda and highest in
Kalogi (Table 5). Kalogi was highest in mean number
of ornamental species per HG (13), with the largest
difference to Kauda (one species, data not shown).
Mean H0and J0in the surveyed 61 HGs were 1.46
(range 0.49–2.42) and 0.48 (0.15–0.87), respectively
(Table 5). Both H0and J0were significantly lower in
Kauda than in Kalogi, Habila and Sama (P \0.001).
Per HG, a mean of four staple, three fruit, eight
vegetable, two wood/MPU, one condiment, medicinal
Agroforest Syst (2014) 88:539–562 547
123
Table 4 Fruit tree species found in 61 homegardens (HGs) surveyed in four villages in the Nuba Mountains, Sudan (2010)
Scientific name Author Family Vernacular name Use Abundance
(individuals)
Frequency
(% of HGs)
English Arabic 1st 2nd
Indigenous fruit trees (IFT)
Adansonia digitata L. Malvaceae Baobab tree Tabaldi/humeir fr v 74 45.9
Balanites aegyptiaca (L.) Del. Zygophyllaceae Desert date Heglig/lalub fr w 107 42.6
Borassus aethiopum Mart. Arecaceae African fan palm Deleb fr w 6 1.6
Capparis decidua (Forssk.) Edgew. Capparaceae – Doumduneidii w fr 2 1.6
Commiphora africana (A. Rich.) Engl. Burseraceae African myrrh – w – 2 1.6
Commiphora pedunculata (Kotschy & Peyr.) Engl. Burseraceae – Gureng w – 1 1.6
Cordia Africana Lam. Boraginaceae Large-leaved cordia San w o 14 16.4
Ficus sp. L. Moraceae – – w fr 1 1.6
Ficus sycomorus L. Moraceae Sycomore fig Gumeiz fr m 3 3.3
Gardenia ternifolia Schumach. & Thonn. Rubiaceae – – w o 3 3.3
Grewia bicolour Juss. Malvaceae Bastard brandy bush Basham fr – 2 3.3
Grewia tenax (Forsk.) Fiori Malvaceae White cross-berry Geduem fr – 8 4.9
Grewia villosa Willd. Malvaceae Mallow raisin – fr – 6 6.6
Hyphaene thebaica L. Arecaceae Gingerbread tree Doum other fr 136 41.0
Lannea acida A. Rich. Anacardiaceae – Duoam v fr 64 14.8
Lannea microcarpa Engl. & K. Krause Anacardiaceae – – w fr 2 1.6
Nauclea latifolia S. M. Rubiaceae African peach Karmadoda fr w 2 3.3
Piliostigma thonningii (Schum.) Milne-Redh. Fabaceae Camel’s foot Kharub w fr 58 16.4
Salvadora persica Wall. Salvadoraceae Toothbrush tree Arak m fr 7 9.8
Sclerocarya birrea A. Rich. Anacardiaceae Marula Homeid fr – 28 26.2
Tamarindus indica L. Fabaceae Tamarind tree Ardeb fr w 11 11.5
Vangueria venosa (Hochst.) Sond. Rubiaceae – Kirkir fr – 2 1.6
Ziziphus spina-christi (L.) Desf. Rhamnaceae Christ’s thorn jujube Sidr/nabak fr other 201 60.7
Exotic fruit trees (EFT)
Annona squamosa L. Annonaceae Sugar-apple Gishta fr – 23 18.0
Azadirachta indica A. Juss. Meliaceae Neem tree Neem w m 125 49.2
Carica papaya L. Caricaceae Papaya Pawpaw fr – 7 8.2
Citrus 9aurantiifolia (Christm.) Swingle Rubiaceae Lemon Lemon fr – 8 9.8
Mangifera indica L. Anacardiaceae Mango Manga fr – 13 16.4
Melia azedarach L. Meliaceae White cedar – w o 10 6.6
548 Agroforest Syst (2014) 88:539–562
123
and stimulant species each as well as three species
with other uses were cultivated. Mean number of
vegetable (6, P =0.018), fruit (4, P =0.134) and
medicinal species (2, P \0.001) were highest in
Sama. Habila exhibited highest numbers of staple (5,
P=0.001) and condiment species (1, P =0.014),
while Kauda harbored most of wood/MPU (5,
P=0.001), stimulant (1, P \0.001) and ‘other use’
species (4, P =0.001). Calculations of the SDR per
plant use category showed some similarities among
villages (only Kauda had a higher dominance of staple
crops and wood/MPU species than the other villages,
data not shown).
Regarding fruit tree species richness, significant
differences among villages were only found for IFT
richness, which was lowest in Kalogi (mean =1) and
highest in Sama (3, Table 6). Several IFT species were
exclusively found in Kauda such as Nauclea latifolia,
Grewia villosa,Grewia bicolor,Lannea acida or
Commiphora africana, the two latter only secondarily
regarded as fruit trees (Table 4). Highest abundance
and individual density of IFTs was found in Sama
(Table 6). Mean IFT H0and J0were 0.56 and 0.65,
but non-significantly different among villages
(Table 6). EFT abundance was significantly highest
in Sama, while differences among villages for the
other assessed EFT parameters were non-significant
(Table 6). Mean richness, abundance, species density,
individual density, H0and J0were significantly lower
for EFT than for IFT species (each P \0.001, two
latter: P =0.001 and 0.002, respectively, Wilcoxon
signed-rank test). However, the fruit tree species
accumulation curves at village level (Fig. 2c, d)
indicated that not enough HGs were inventoried to
cover the total species richness, both for IFTs and
EFTs. Regarding IFTs, Habila showed still increasing
levels of species, while Kalogi reached saturation after
about six sampled HGs (Fig. 2c). Less pronounced
differences among villages were found for EFT
species, though more species can be expected in all
villages, except for Kalogi (Fig. 2d).
Plant species richness, diversity and use
between market-oriented and subsistence HGs
Market-oriented HGs showed a statistically signifi-
cantly higher mean of useful species than subsistence
gardens (Table 5). Further richness and diversity
Table 4 continued
Scientific name Author Family Vernacular name Use Abundance
(individuals)
Frequency
(% of HGs)
English Arabic 1st 2nd
Parkinsonia aculeate L. Fabaceae Jerusalem thorn Seisaban w o 38 8.2
Phoenix dactylifera L. Arecaceae Date palm Ballah fr – 47 23.0
Psidium guajava L. Myrtaceae Common guava Guava fr – 5 6.6
Note: Not only species mentioned as fruits by the respondents (either as primary (1st use) or secondary (2nd) use), but also those with edible fruits according to the literature are
listed below
Use categories: fr fruit tree, mmedicinal plant, oornamental, other other uses (fiber and fencing material), vvegetable, wwood/multipurpose use
Agroforest Syst (2014) 88:539–562 549
123
parameters (incl. those assessed for IFT and EFT
species) showed no significant differences, though
market-orientated HGs exhibited slightly higher val-
ues for some of the parameters (Tables 5,6). Subsis-
tence and market-oriented HGs differed, however, in
the dominance of plant use categories. Subsistence
HGs showed a higher dominance of staple crops, while
vegetable species dominated market-oriented HGs
(Fig. 4a). Although generally lower in proportion, also
medicinal and fruit species had a higher SDR in
market-oriented HGs. Market- and subsistence-ori-
ented HGs also slightly differed in their species
composition. Of the 110 plant species found, 79 were
present in both subsistence- and market-oriented HGs,
while 21 and 11 species were exclusively found in
subsistence and market-oriented gardens, respec-
tively. Species uniquely found in subsistence gardens
were for instance Sorghum 9drumondii and Cit-
rus 9aurantiifolia, whereas Acacia nubica and Phys-
alis angulata were only found in market oriented
gardens.
Determinants of richness, density and diversity
of useful plant species
Gardens of the indigenous Nuba people contained
higher richness of useful plant species (excluding
ornamentals) than those of non-Nuba households (21
vs 25; P =0.029). However, H0and J0indices were
lower in gardens of Nuba households (1.26 vs 1.60,
P=0.007, and 0.41 vs 0.54, P =0.003, respec-
tively). The gender of the person mainly managing the
garden did not affect species richness and diversity.
Multiple regression analyses confirmed some of the
effects of the above mentioned factors on species
richness and diversity, although the strength of the
obtained models was rather moderate or weak,
explaining less than 60 % of the variation (Table 7).
The models fit best for individual density, H0, and J0
index (more than 50 % of variation explained;
Table 7). Species richness was positively influenced
by the location Kauda, but negatively influenced by
subsistence-oriented production and increasing soil
Fig. 2 Species accumulation curves using observed richness
(Mao Tao estimator ±standard deviation (SD) for 61 surveyed
homegardens (HGs) in four villages of the Nuba Mountains,
Sudan (2010). For all gardens: aexotic versus indigenous useful
plant species, bexotic fruit tree (EFT) species versus indigenous
fruit tree (IFT) species. Per village: cIFT species, dEFT
species. Graphs per village are slightly displaced to avoid SD-
bar overlaps
550 Agroforest Syst (2014) 88:539–562
123
pH levels. Individual density was positively influ-
enced by the locations Kauda and Habila. Share of
exotic individuals was solely negatively influenced by
the location Kauda. A strong negative influence on H0
as well as J0was caused by the location Kauda, slightly
also by Kalogi on H0. Furthermore, IFT richness was
negatively affected by pH, but positively by HPI.
Putatively affecting variables such as CEC
eff
, culti-
vated HG area, age of the HG as well as elevation did
not contribute to the models.
Classification of gardens according to species
composition
Nearest neighbor cluster procedure did not identify
any outlier and allowed to include all surveyed HGs in
the subsequent minimum variance cluster analysis.
Based on the ‘elbow-criterion’ four different clusters
of HGs were found (Fig. 3). HGs of Kauda clearly
separated from all others and clustered as a whole
together (cluster 1), whereas HGs of the other three
villages were assigned to different clusters each.
Stepwise discriminant analysis confirmed that the four
clusters were correctly classified (98.4 % of cases in
the right cluster) and the two first ordination functions
together explained already 96.3 % of variation. The
low and significant Wilks’ kvalue of 0.1 % unex-
plained variation out of 100 % reflected the goodness
of clustering and the independence between each of
the clusters (P \0.001). Discriminant analysis
extracted 15 plant species with a major predictive
power to separate HGs into the four clusters (in
decreasing order): S. bicolor,Z. mays,A. esculentus,
A. hypogaea,B. aegyptiaca,Solanum melongena,S.
lycopersicum,S. indicum,C. melo,Vigna unguiculata,
Terminalia laxiflora,A. nubica,P. angulata, and
Cajanus cajan. While the first four species were
frequently found in all villages even though differ-
encing individual densities, the remaining 11 species
were rather present in unbalanced densities or were
limited to certain villages.
In addition to the above mentioned differences in
species composition, the four clusters also differed in
various variables including species richness and
Table 5 Mean richness, density, abundance and diversity for
useful plant species (without ornamentals) and for ornamental
species in 61 homegardens (HGs) surveyed in four villages, in
the Nuba Mountains, Sudan (2010) arranged by village,
economic orientation of HG production and cluster affiliation
Plant species without ornamentals
n Richness Share of
exotic
species (%)
Arrhenius
species density
per 1,000 m
2
Abundance Individual
density per
1,000 m
2
Share of
exotic
individuals
(%)
Shannon Evenness
Kauda 15 27.5
a
33.5
b
23.0
a
39,363
a
8,552
a
11.4
b
0.94
b
0.28
b
Kalogi 15 16.5
b
50.0
a
16.7
b
1,822
b
1,851
b
64.0
a
1.45
a
0.53
a
Habila 15 23.1
ab
43.8
ab
24.0
a
4,560
b
3,732
b
55.7
a
1.69
a
0.54
a
Sama 16 23.6
a
36.8
b
23.9
a
3,236
b
2,521
b
51.6
a
1.74
a
0.58
a
P 0.001 <0.001 0.006 <0.001 <0.001 <0.001 <0.001 <0.001
Market-
oriented
19 25.9 45.1 24.8 8,570 3,757 52.6 1.56 0.49
Subsistence 42 21.0 39.2 20.6 13,179 4,294 42.2 1.43 0.49
P 0.032 0.107 0.058 0.313 0.617 0.174 0.305 0.830
Cluster 1 15 27.5
a
33.5
c
23.0
ab
39,363
a
8,552
a
11.4
b
0.94
b
0.28
b
Cluster 2 22 20.7
b
38.3
bc
20.7
bc
4,562
b
3,353
b
51.0
a
1.58
a
0.53
a
Cluster 3 11 13.3
b
47.6
ab
14.5
c
197
c
321
c
71.4
a
1.44
a
0.57
a
Cluster 4 13 28.5
a
48.5
a
29.2
a
3,459
b
3,598
b
54.9
a
1.89
a
0.57
a
P 0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001
Total 61 22.7 41.1 21.9 12,097 4,137 45.8 1.46 0.48
Different letters behind means in a column and bold P values indicate significant differences at P \0.05 (Mann–Whitney or Kruskal–
Wallis tests, depending on the data structure)
Agroforest Syst (2014) 88:539–562 551
123
Table 6 Mean richness, density, abundance and diversity for indigenous and exotic fruit tree species in 61 homegardens (HGs) surveyed in four villages in the Nuba Mountains,
Sudan (2010) arranged by village, economic orientation of HG production and cluster affiliation
n Indigenous fruit trees (IFT) Exotic fruit tree species (EFT)
Richness Arrhenius
species
density per
1,000 m
2
Abundance Individual
density per
1,000 m
2
Shannon Evenness Richness Arrhenius
species
density per
1,000 m
2
Abundance Individual
density per
1,000 m
2
Shannon Evenness
Kauda 15 2.4
ab
2.1
ab
6
b
1
b
0.68 0.66 0.8 0.7 1.3
b
0.5 0.17 0.18
Kalogi 15 1.3
b
1.6
b
3
b
5
ab
0.29 0.38 0.9 0.9 1.5
ab
1.9 0.41 0.48
Habila 15 1.9
ab
2.2
ab
5
b
8
ab
0.52 0.55 0.7 0.5 1.6
ab
2.1 0.22 0.20
Sama 16 2.8
a
3.0
a
15
a
19
a
0.70 0.62 1.4 1.1 3.8
a
4.6 0.26 0.34
P 0.030 0.031 0.020 0.001 0.055 0.491 0.131 0.188 0.014 0.136 0.537 0.474
Market-
oriented
19 2.5 2.6 10 8 0.70 0.63 1.1 0.9 2.8 2.6 0.23 0.25
Subsistence 42 1.9 2.1 6 9 0.49 0.52 0.8 0.7 1.6 2.2 0.25 0.31
P0.249 0.178 0.182 0.295 0.119 0.481 0.293 0.569 0.131 0.513 0.803 0.687
Cluster 1 15 2.4 2.1 6 1
b
0.68 0.66 0.8 0.7 1.3 0.5
c
0.17 0.18
Cluster 2 22 1.9 2.1 7 9
ab
0.46 0.49 0.8 0.5 2.1 2.2
bc
0.35 0.44
Cluster 3 11 1.6 2.0 5 12
ab
0.40 0.48 1.1 1.1 1.6 3.4
ab
0.26 0.38
Cluster 4 13 2.5 2.8 12 13
a
0.69 0.61 1.3 1.2 3.1 3.5
a
0.31 0.28
P0.354 0.474 0.097 0.001 0.242 0.594 0.505 0.109 0.343 0.047 0.700 0.653
Total 61 2.1 2.2 7 9 0.56 0.65 1.0 0.8 2.0 2.3 0.27 0.31
Different letters behind means in a column and bold P values indicate significant differences at P \0.05 (Mann–Whitney or Kruskal–Wallis tests, depending on the data
structure)
552 Agroforest Syst (2014) 88:539–562
123
Table 7 Results of stepwise multiple regression analyses of selected socio-economic and bio-physical factors affecting species
richness and diversity parameters in 61 homegardens (HGs) surveyed in four villages in the Nuba Mountains, Sudan (2010)
Useful plant species richness (excluding ornamentals) Indigenous
fruit trees
(IFT)
Richness Indiviudal density
per 1,000 m
2
Share of exotic
individuals
Shannon
diversity
Shannon
evenness
Richness
Adjusted R
2
0.417** 0.573*** 0.467** 0.506*** 0.568*** 0.278**
Location Kauda (0 =no; 1 =yes) 0.335* 0.813*** -0.693*** -0.768*** -0.760***
Location Habila (0 =no; 1 =yes) 0.282*
Location Kalogi (0 =no; 1 =yes) -0.267*
Level of subsistence (0 =selling
produce, 1 =subsistence)
-0.586***
pH -0.463** -0.581***
CEC
eff
Household poverty index (HPI) 0.443**
Cultivated HG area (m
2
)
HG age (years)
Elevation (m)
For each explanatory variable, the standardized regression coefficient (b) and its significance levels are given
ns not significant
*PB0.05; ** PB0.01; *** P\0.001, respectively
Fig. 3 Dendrogram resulting from minimum variance (squared
Euclidean distances) cluster analysis based on ln-transformed
density data (individuals per 1,000 m
2
HG area) of 110 useful
plant species (without ornamentals), in 61 homegardens (HGs)
of the Nuba Mountains, Sudan (2010). The dashed line indicates
the separation into four clusters according to the ‘elbow
criterion’. Brief cluster description (left side) gives parameters
in the order: assigned HG type, percentage of market-orientation
and share of the main village present as well as a rough diversity
assignment
Agroforest Syst (2014) 88:539–562 553
123
diversity parameters (Tables 5,6) as well as socio-
economic and bio-physical variables (Tables 2,3).
HGs of cluster 1 most clearly separated from all
other HGs as revealed by the deepest dendrogram
node position (Fig. 3). These HGs were all located in
Kauda and managed by young female Nuba (Otoro
tribe), native to the location. HGs of this cluster were
old and had the largest sizes, managed by rather poor
households with very little additional land holdings
and of low literacy rate (Table 3). Commercialization
of production was low. In HGs of cluster 1, highest
total soil N and Mg values were recorded (P B0.01),
while overall mineral content, P and C/N was lowest
(Table 2). Plant species richness was high, however,
diversity and J0indices as well as share of exotic
species and individuals were lowest (Table 5). In
contrast, IFT richness and diversity were highest
(P \0.05, Table 6). Regarding SDR values, staple
crops dominated this cluster while the importance of
vegetable and condiments species was lowest (Fig. 4).
Fiber plants were largely cultivated in HGs of
cluster 1. The five most important and dominant
species were S. bicolor,S. indicum,A. hypogaea,
C. fascicularis and V. unguiculata. Only in this clus-
ter, the two latter species were among the five most
dominant species. According to the above mentioned
characteristics, HGs of cluster 1 were named as
‘traditional-staple’ HGs.
HGs of cluster 2 were of rather young mean age and
small size (Table 3). These HGs were mostly for
subsistence purpose (73 %, Fig. 3); about 50 % of the
gardeners were illiterate and non-native to the partic-
ular village. TLUs per household member were
lowest, but the rate of commercialization of HG
production was the second highest of our study.
Regarding soil quality, N was low in HGs of cluster 2,
while soil mineral values were rather moderate
(Table 2). Richness and diversity measures of both
total useful plant species and IFTs were intermediate
or low (Tables 5,6). The most balanced mixture of
plant use categories was found in this cluster that was
relatively similar to cluster 1, but with a higher
importance of vegetables at the expense of staples
(Fig. 4b). HGs of cluster 2 had the highest number of
medicinal plants (data not shown). Only in this cluster
the staple P. glaucum was among the five most
dominant crop species, the other four were S. bicolor,
Z. mays,A. hypogaea and S. indicum. Since house-
hold as well as plant diversity characteristics were
both at intermediate stages and the importance of plant
use categories relatively balanced, HGs in this cluster
were named as ‘transitional-staple’ HGs.
Cluster 3 comprised rather young and small HGs
(Table 3). The share of female gardeners was low and
literacy rate high (Table 3). Households managing
HGs of cluster 3 were wealthy as expressed by the high
HPI and TLUs per household member (Table 3). As
much as 91 % of HGs in this cluster were subsistence-
oriented. Income gained from the HG produce was by
far the lowest recorded, while employment rate was
highest (both P \0.05; Table 3). Soils in these HGs
had lowest CEC
eff
and intermediate N concentrations
(Table 2). Species richness was low, but share of
exotic species high (Table 5). J0was highest, similar to
cluster 4. IFT richness and diversity were relatively
low, whereas the corresponding EFT parameters were
intermediate (Table 6). HGs of cluster 3 were domi-
nated by vegetables, but the categories fruits and
‘other uses’ were also important. The five most
Fig. 4 Summed dominance ratios for eight plant use groups
(without ornamentals), separately for athe two types of
economic orientation of HG production and bfour clusters of
61 homegardens surveyed in four villages of the Nuba
Mountains, Sudan (2010). MPU multipurpose use
554 Agroforest Syst (2014) 88:539–562
123
dominant species were A. esculentus,Z. mays,S. ly-
copersicum,C. fascicularis and Ocimum gratissi-
mum. The latter species—used as repellent against
insects, but with a ‘weedy’ behavior—was exclusively
found in HGs of cluster 3. We observed a low
dependence of households on HG produce and lower
maintenance levels in these HGs, which were conse-
quently described as ‘pastime-mixed’ HGs.
Cluster 4 comprised small and young HGs man-
aged by households with relatively large landholdings
and high illiteracy rates (Table 3). Almost 39 % of
these HGs were managed by male gardeners. Fifty-
five percent of the gardeners were of Arabic affiliation.
In line with the high proportion of market-oriented
HGs in cluster 4 (69 %), gross income derived from
HG produce was highest (Table 3). Soils of HGs in
this cluster showed high mineral levels resulting in the
highest CEC
eff
, while total N was moderate (Table 2).
Useful plant species richness and diversity were
comparatively high (Table 5). Also richness and
diversity of IFT and EFT species generally ranged
highest among all clusters (Table 6). The importance
of staple crops was lowest while that of vegetables
highest (Fig. 4). Only in this cluster, a ‘living fence’
made of Jatropha curcas and Xanthium brasilicum,
an invasive weed from South America, was found, the
latter was used as a fuel. The two vegetables,
Eruca sativa and Corchorus olitorius grown as cash
crops were among the five most dominant species. The
three others were A. esculentus,Z. mays and S. lyco-
persicum. Based on the HG features extracted we
named this type as ‘commercial-vegetable’ HG type.
Discussion
The present HG study revealed highly variable levels
of plant diversity and socio-economic household
characteristics of this agroforestry system (Tables 3,
5and 6) as well as clear signs of transitional processes.
The mean HG size with 1,900 m
2
was in range with
other studies from semi-arid regions (Albuquerque
et al. 2005; Bernholt et al. 2009; Okafor and Fernandes
1987). HGs of Kauda, however, were four times larger
compared with the other three villages (Table 3),
laying within the range of 0.4–3.0 ha known for HGs
of East and Central African highlands (Abebe et al.
2006). The dominance of staple crops in HGs of Kauda
(see cluster 1 in Fig. 4) reflects their importance for
subsistence farming as similarly shown for HGs in
Nepal (Gautam et al. 2008).
Plant species richness and diversity
The total of 110 useful plant species (without orna-
mentals) and the mean of 23 species per HG (Table 5)
was comparatively high (and would have even been
higher for total numbers according to the species
accumulation curves, Fig. 2a, b) regarding the region
of Kordofan where Gebauer (2005) documented a total
of only 32 species and a mean of three per HG,
however, in an urban setting. Studies from Niger
(Bernholt et al. 2009) and Yemen (Ceccolini 2002)
also reported lower mean richness ranging from four
to 14 species per garden. Although not a main focus of
our analyses, a closer look on ornamental species
contributed to the observed differences and further
understanding. The mean ornamental species richness
with nine species per HG and a range of 1–13 per
village was comparatively higher to gardens of semi-
arid tropical Niger (Bernholt et al. 2009), where on
average only two ornamental species were cultivated.
As ornamentals mainly have aesthetic function instead
of subsistence food production and a gradual substi-
tution of crops by ornamental plants is known to occur
in HGs of wealthier families (Christanty et al. 1986),
the richness of ornamental species can serve as an
indicator for transformation processes. This was
evident for HGs of clusters 3 and 4 that grouped
gardens managed by better-off families or market-
oriented HGs, which at the same time had highest
richness of ornamentals (mean =12 species, data not
shown) and high dominance of vegetables (Fig. 4a).
The opportunity to gain off-farm incomes (cluster 3)
or cash income from HG produce (cluster 4) may thus
have contributed to shifting HG production away from
growing basic staple crops for self-consumption. In
our study, the mean H0of 1.46 (Table 5) can be
considered as moderate according to Barbour et al.
(1987), who rated H0values of [2 as high. Diversity
was still higher than for urban gardens of Sudan’s
capital Khartoum (H0=1.20, Thompson et al. 2010),
gardens of a rural urban gradient in Zambia
(H0=0.81–1.35, Drescher 1998) as well as for
mostly commercial gardens in Niamey, Niger
(H0=0.77–0.93, Bernholt et al. 2009). A study from
Ethiopia showed a similar range of diversity with a
mean of 1.45 (Abebe et al. 2009). Although species
Agroforest Syst (2014) 88:539–562 555
123
richness was highest for Kauda, diversity was lowest
(Table 5) while at the same time garden sizes were
largest in Kauda. However, HGs in Kauda were
largely dominated by few species, such as S. bicolor
and V. unguiculata. It was already shown in other
studies that H0is influenced by single dominant
species (Bernholt et al. 2009) as well as rare species
(Fentahun and Hager 2009b; McCune and Grace
2002) such as C. africana (Table 4) in our study,
which only occurred in HGs of Kauda. Gardeners of
Kauda rely on cultivating a broad diversity of tradi-
tional and non-exotic species for food and nutrition
security of their families. The very traditional garden-
ing practices in Kauda may also be a result of the
limited migration history of the respondents at this
particular location. This fact, together with the
remoteness and rather homogenous ethnicity of
inhabitants in Kauda may have led to limited exchange
of planting materials with other communities and a
certain dependence on internal seed/seedling
exchange networks though a high useful plant species
richness could be maintained. Such dynamics are
known from Peru (Wezel and Ohl 2005), Iran (Has-
hemi et al. 2013) and Zambia (Drescher 1998) where
poor market proximity resulted in lower richness
through hampered introduction of new species
(including ornamentals). Coomes and Ban (2004)
showed that exchange of plant material contributed to
an increased plant species diversity in HGs of
Amazonian Peru. Better access to markets and
enhanced exchange of plant material through mobility
of people could have resulted in the high species
richness and diversity (particularly of ornamentals) in
the other villages (Table 5) such as Sama (close to
Kadugli, the capital of South Kordofan, Fig. 1) even
though a functional loss of some HG types as a staple
food production system may have occurred (Fig. 3).
Indigenous fruit tree (IFT) diversity
Total IFT richness of our study (12) was similar to
HGs in Kordofan, Sudan (10 species, Gebauer 2005)
as well as Niamey, Niger (14, Bernholt et al. 2009), but
higher to urban gardens of Khartoum, Sudan (2,
Thompson et al. 2010). However, the species accu-
mulation curves showed a still increasing IFT richness
if more HGs had been inventoried, particularly in
Habila (Fig. 2c). The mean H0for IFTs (0.56, Table 6)
in our study was low compared with farmland of
Uganda (Agea et al. 2007), Ethiopia (Fentahun and
Hager 2009a) and Tanzania (Munishi et al. 2008),
where fruit tree H0of 2.2, 1.9 and 2.7, respectively,
were documented, the latter calculated for all tree
species. Higher richness, but lower diversity reflects
an imbalance of tree species abundances (dominated
by Z. spina-christi,B. aegyptiaca and A. digitata)in
our study. J0however, showed a higher value (0.65) in
our study, as compared for instance with Ethiopian
farmland (Fentahun and Hager 2009a), where an J0of
only 0.4 was found. In our study, mean IFT richness
and abundance was highest in Sama (Table 6) indi-
cating the function of wild food resources in these
villages where trees were left for special purposes
when clearing natural vegetation for establishing HGs.
On the other hand, the remote village Kauda also had a
high IFT species richness and some IFT species such
as G. villosa,G. bicolor and N. latifolia were
exclusively found in Kauda, which possibly reflects
also inter-site differences of the current natural tree
species distribution due to climatic differences
(Table 1) and human influence, such as logging and
over-exploitation of forests and woodlands in Kordo-
fan (El Tahir et al. 2010). However, also the higher age
of the HGs in Kauda (Table 3) might have had a
positive influence on IFT richness, as similarly stated
for HGs in Peru (Coomes and Ban 2004).
Market-oriented HGs harbored similar mean num-
bers of fruit tree individuals and species (Table 6), but
showed a slightly higher dominance of the use group
fruits (and wood/MPU tree species) than subsistence
gardens (Fig. 4) which was in contrast to results of
Muneer (2008) in Sudan, who found commercial HGs
with a reduced set of woody species, including fruit
trees. Mendez et al. (2001) on the other hand, showed
that HGs in Nicaragua devoted to fruit tree cultivation
were not particularly for generating cash income, but
rather for self-consumption of fruits. Regarding the
present study, this was in line to HGs of cluster 1, but
in contrast to HGs of cluster 4: although both HG
types harbored rather high IFT richness (Table 6),
density of IFTs was relatively high in the latter with
the largest share of market-oriented HGs. However,
cash income generation from these on-farm fruit tree
resources seemed to play only a minor role in the Nuba
Mountains particularly of IFTs in the area of Sama
(Goenster et al. 2011). Instead, branches of on-farm
IFTs were for example used for fencing the garden like
those from thorny Z. spina-christi and B. aegyptiaca.
556 Agroforest Syst (2014) 88:539–562
123
In addition, wild fruits were mainly collected in the
nearby forests, however, not for sale but rather for
home consumption, similar to results from Tanzania
(Munishi et al. 2008). Other studies, reported higher
contributions of non-timber forest products to house-
hold cash income generation in Kordofan (Adam et al.
2013; El Tahir and Gebauer 2004).
Determinants of species richness and diversity
The medium to low correlations of determinants
revealed by multiple linear regressions (Table 7)
reflected the difficulty to find single or combined key
parameters influencing species richness and diversity.
The location as such had some informative power,
where type of ethnicity, remoteness, level of market
access and mobility of people are likely to boost or
hamper exchange of plant material or level of staple
crop production, among others. The strong influence of
the location Kauda is likely related to the traditional
function of these HGs for subsistence production as
compared to the marked differences in all the remaining
villages (Fig. 3). Latter are obviously more affected by
recent transformation processes, including the decreas-
ing HG sizes that has happened for the last 20 years
according to key informants (Amir Mahmoud el-
Murad, chief of the Shawabna ethnicity in Sama,
personal communication)—a process that will most
probably continue in future. At the same time, the
importance of HGs for subsistence may further
decrease because: (i) additional activities in tertiary
sectors limit the time to work the garden and increase
incomes to purchase food from local markets; (ii)
informal land use regulations prevent families to own at
maximum 400 m
2
of land around their houses including
restrictions to grow crops on unused areas such as in the
front of homesteads; (iii) decreasing knowledge of
traditional gardening practices and motivation to grow
crops, and (iv) changed patterns of husbandry and
increased livestock numbers in settlements resulting
from demographic growth. As a consequence, garden-
ers need to establish fences typically made of branches
of local tree species that are cut in the forests and
installed around the HGs. Collection, transport and
installation of fencing material is labor and thus cost
intensive. The factors listed above and recent changes
through external influences may also threaten the
suitability of the surveyed HGs for circa situm conser-
vation of plant genetic resources (see below).
Little differences in overall as well as IFT species
richness and diversity parameters were detected
between subsistence and market-oriented jubrakas.
Instead, market-oriented gardens harbored higher total
species richness (Table 5). While Bernholt et al.
(2009) documented a similar trend, other studies
reported a loss of vegetation strata, plant species
richness and diversity due to commercialization
(Abdoellah et al. 2006; Peyre et al. 2006). Particularly,
a trend of losing traditional vegetables, often with a
high nutritional value, may occur in commercial HGs
as mentioned by Abdoellah et al. (2006). This was
observed in our study for indigenous vegetable species
such as Amaranthus viridis or Lactuca taraxacifolia
that were more frequently cultivated and used by
subsistence gardeners, L. taraxacifolia even as
malaria preventive.
Classification of HGs
Cluster analysis allowed identifying HG types which
differed not only in species richness and diversity
parameters (Tables 5,6), but also in socio-economy
and soil quality characteristics (Tables 2,3; Fig. 3).
This approach should be applied more frequently to
classify agro-ecosystems as it may help to design more
comprehensive recommendations regarding plant spe-
cies conservation and management (Kehlenbeck et al.
2007; Peyre et al. 2006). In the present study, cluster
analysis identified a gradient along an evolutionary
time-scale of HG development. The most traditional
HGs were grouped in cluster 1 (Fig. 3), located in the
most remote village, managed exclusively by women,
characterized by a high species richness, dominance of
staples for subsistence, low portions of exotic
(Table 5) and few ornamentals. When taking these
characteristics as the traditional jubraka features,
changes in the other clusters are substantial, particu-
larly in cluster 4 with its commercial vegetable
gardens rich in exotic (Table 5) and ornamental
species, and often managed by male gardeners. HGs
of clusters 2 or 3 were on an intermediate state or used
as pastime gardens, respectively (Fig. 3). They are
characterized by a mixed species composition mainly
grown for home consumption and low levels of
commercialization, but including already numerous
ornamental and exotic species (Table 5) as similarly
found in non-commercial HGs of Indonesia (Abdoel-
lah et al. 2006). The large dominance of staples in HGs
Agroforest Syst (2014) 88:539–562 557
123
of cluster 1 may also be explained by less additional
landholdings of the same households (Table 3). Lack
of sufficiently large fields for staple crop production
forces households to grow the needed staples in their
HGs, which was also observed in HGs of migrant
families with little additional farm land in Sulawesi,
Indonesia (Kehlenbeck et al. 2007).
The varying species composition also reflected the
ethnical differences of the surveyed villages (Table 1).
Several plant species were limited to certain villages
and linked to the cultural traditions of the gardener’s
ethnic affiliation. P. glaucum for instance, the most
expensive local grain (2 €per kg
-1
grain) in the study
region was absent in Kauda, but cultivated in Kalogi
and Habila where pastoral ethnicities originally from
Northern and Western Africa are predominantly
present and traditionally cultivate drought-tolerant
millets instead of sorghum. On the other hand,
traditional vegetable species such as C. fascicularis
and Stylochaeton hypogeum occurring in 60 and 18 %
of the surveyed HGs, respectively, were preferably
used by the gardeners grouped in cluster 1 (Kauda),
but, although known, less by the gardeners in the other
villages. Similar influences of ethnicities were
reported from urban gardens in Niamey, Niger (Bern-
holt et al. 2009), and HGs in ethnically diverse villages
of Peru (Wezel and Ohl 2005).
The high levels of plant species richness and
relatively homogeneous species composition in clus-
ter 1 (Tables 5,6) may be further explained by the
limited access of the rather poor farmers of Kauda
(low HPI, Table 3) to external agricultural inputs and
labor force, which is said to conserve plant species,
even those of no or limited utility value (Kaihura et al.
2001). In a summary, our cluster analysis confirmed
the results of the regression analysis that the factors
‘location’ and ‘commercialization level’ had an
important influence on species richness and diversity
parameters, but cluster analysis also revealed addi-
tional influencing factors and differences of species
composition among garden types.
Suitability of HGs for on-farm conservation
of indigenous plant species
Farmers have to define their cultivation goals accord-
ing to the subsistence needs of their families and recent
demands of the market, which may sometimes not
favor indigenous plant species. While the first is likely
to be matched with indigenous and old neophytic crop
species that have been used for centuries, are adapted
to local climatic conditions, and are often strongly
linked with traditional knowledge and cultural values
(Kumar and Nair 2004), the latter will be more related
to newly introduced cash crops (Abdoellah et al. 2006;
Peyre et al. 2006) or improved varieties of traditional
crops. Cash cropping and using improved materials
can be seen as beneficiary for farmers, for example if
low yielding traditional landraces are replaced by
better varieties as reported for instance for the Indian
Himalaya (Bisht et al. 2006). However, often exotic
species and improved varieties may develop high
productivity only under intensive systems using high
levels of external inputs, which might not be acces-
sible by resource-poor subsistence farmers. In our
study, more than 40 % of the useful species were of
exotic origin, although large spatial differences were
observed. Particularly for the food use classes fruits,
vegetables, staples and spices, 47, 55, 56 and 67 % of
the species were exotics, respectively, while for wood
and MPU species only 18 %. Many of the indigenous
species were found only in low frequencies (44 % of
the 68 indigenous useful species were detected in less
than five HGs, 15 species even in only one single HG)
and/or low abundances (almost 50 % of the 68
indigenous species were present with less than 20
individuals, most of them perennial species). Even the
12 IFT species, although relative frequent (five species
occurred in more than 10 % of the surveyed HGs),
were present with only few individuals (seven species
with less than 10 individuals each). Many further IFT
species such as Azanza garckeana,Boscia angustifo-
lia,Celtis integrifolia,Diospyros mespiliformis or
Ximenia americana were found in the natural vege-
tation surrounding the surveyed villages in the Nuba
Mountains (personal observation), but not in the
surveyed HG. Thus, the value of the surveyed HGs
for on-farm conservation of indigenous and traditional
plant genetic resources is questionable, particularly
when considering the current transformation pro-
cesses. Many of the indigenous plant species recorded
in the HGs were either common and very frequently
cultivated food crops (e.g. S. bicolor,C. melo
and S. indicum), weed-like vegetables and fodder
species (e.g. Cleome gynandra,Corchorus tridens
and Commelina sp.) or common fuel wood and
timber species, quite abundant in the surrounding
woodlands (e.g. Albizia amara,Acacia nilotica and
558 Agroforest Syst (2014) 88:539–562
123
Faidherbia albida). Only a few indigenous species
found in the surveyed HGs could be regarded as ‘rare’
in the surrounding natural vegetation, for instance the
IFT Grewia tenax. A similar rather low value of HGs
for the conservation of indigenous plant species was
also reported from Indonesia (Kehlenbeck et al. 2007),
while Bennett-Lartey et al. (2001) stated that HGs in
Ghana are largely suitable for crop species conserva-
tion. However, as the pressure on the remaining
woodlands is still increasing in the Nuba Mountains
and abundance of many indigenous species is said to
decrease (El Tahir et al. 2010), the importance of the
existing HGs for circa situm conservation of plant
genetic resources might increase in the future. Out of
the surveyed HGs, the ones grouped into cluster 1 with
their high species richness (including IFTs) and low
portion of exotics seem to be most promising for
species conservation. This fact is furthermore sup-
ported by the low shares of ornamentals that has been
likely prevented a replacement of indigenous species
and use groups to some extent at this cluster. On the
other hand IFT and EFT species were more abundant
in cluster 4 (commercial vegetable HGs), maybe
because the wealthier families managing these HGs
(Table 3) needed less space for own staple food
production in their gardens than the poorer families in
remote Kauda, where in addition efficient weeding of
tree seedlings was observed. Promotion of agrofor-
estry systems may contribute to enhancing the value of
HGs in the Nuba Mountains for circa situm conser-
vation of tree species including IFTs, e.g. by increas-
ing household incomes from sale of tree products. IFT
species are often neglected by research and develop-
ment, but may fetch high market prices. For instance
fruits of G. tenax were sold in Sudan for about
2.5 €kg
-1
in 2004 at El Obeid market (Gebauer et al.
2007) and 5 €kg
-1
in 2010 at the Omdurman market
near Khartoum (personal observation). The high
potential of IFTs for income generation, but also for
family nutrition (particularly of children, who were
observed to extensively collecting and consuming IFT
fruits from woodlands in the study area) could be
further increased through participatory tree domesti-
cation of the most preferred and valuable IFT species,
which should be performed in a participatory approach
as suggested by Leakey and Simons (1997) and partly
achieved in Southern and Western Africa (Akinnifesi
et al. 2007). However, both on-farm species conser-
vation approaches as well as IFT domestication
programs are still to be developed in Sudan, and
jubraka HG systems could offer a suitable environ-
ment for initiating and testing the mentioned programs
and approaches.
Conclusions
The jubraka HG systems of the Nuba Mountains
harbored a surprisingly high plant species richness and
diversity. However, various constraints of gardening
such as lack of fencing material, small garden sizes as
well as poor access of gardeners to germplasm—both
for traditional species, but alsofor improved varieties—
seem to affect the motivation of households to cultivate
their HGs. Since these HG most are likely to be
increasingly subjected to the introduction of exotic
species as well as to transformations as a result of socio-
economic changes such as commercialization, their
plant species composition will change which may
possibly lead to losses of traditional species. In
addition, the current and future importance of the
surveyed HGs for circa situm conservation of indige-
nous and traditional species might be questionable. It
remains unclear if the HGs in the Nuba Mountains can
still fulfill their current function in contributing to food
and nutrition security of the families managing them in
the future. In contrast to popular beliefs, a modest
commercialization of HG production may contribute to
maintaining and even enhancing species richness and
diversity in these systems. However, looking at the
many factors affecting and possibly threatening species
richness and diversity in HGs, there is the need to
(i) raise awareness of local communities on the
nutritional and ecological advantages of growing
traditional plant species and varieties, (ii) improve
access of gardeners to decentralized distribution sys-
tems of germplasm, and (iii) promote subsistence and
semi-commercial cultivation of diverse plant species
(including IFTs) and varieties for home consumption
and income generation. Species diverse agricultural
production systems are of particular importance for
smallholder farmers to ensure resilience and sustain-
ability of food production in a region which is subject to
climate change, severe health and food security prob-
lems and unstable political conditions.
Acknowledgments We thank the Deutsche Forschungs-
gemeinschaft DFG (German Research Foundation) for
Agroforest Syst (2014) 88:539–562 559
123
funding this research as part of the project ‘Effects of
transformation processes in ‘jubraka’ agroforestry systems of
the Nuba Mountains, Sudan, on plant diversity and nutrient
fluxes’, BU 1308/9-1 & GE 2094/1-1. We would like to highly
acknowledge our field assistants Sabri Abdul Karim from
Lumon, Moza Suleiman from Sama, Ahmed Al Zet,
Mohammad Defallah, Adam Muza and Omar Balandia from
Kalogi, and Kochoro from Habila who translated for us, thereby
enabling fruitful conversations about local knowledge and HG
management in the Nuba Mountains. We are also thankful to all
the respondents who allowed access to their properties and
shared their experiences with us. Moreover, many thanks to all
the families of all villages and the agricultural guesthouse in
Habila for their great hospitality and accommodations provided.
Finally, we would like to extend our great respect to the local
chiefs and authorities of the respective villages for their friendly
cooperation and for providing us with necessary working
permits.
References
Abdalla AI (2007) Country report: the Republic of the Sudan.
Horticultural Crop Production. Korea International Coop-
eration Agency International Cooperation Training Center
and Kyungpook National University International Agri-
cultural Training Center, Seoul, Daegu, Korea, pp 41
Abdoellah OS, Hadikusumah HY, Takeuchi K, Okubo S, Pari-
kesit (2006) Commercialization of homegardens in an
Indonesian village: vegetation composition and functional
changes. Agrofor Syst 68:1–13
Abebe T, Wiersum KF, Bongers F, Sterck FJ (2006) Diversity
and dynamics in homegardens of southern Ethiopia. In:
Kumar BM, Nair PKR (eds) Tropical homegardens: a time-
tested example of sustainable agroforestry. Springer,
Dordrecht, pp 123–142
Abebe T, Wiersum KF, Bongers F (2009) Spatial and temporal
variation in crop diversity in agroforestry homegardens of
southern Ethiopia. Agrofor Syst 78:309–322
Adam YO, Pretzsch J, Pettenella D (2013) Contribution of non-
timber forest products livelihood strategies to rural devel-
opment in drylands of Sudan: potentials and failures. Agric
Syst 117:90–97
Agea JG, Obua J, Kaboggoza JRS, Waiswa D (2007) Diversity
of indigenous fruit trees in the traditional cotton-millet
farming system: the case of Adwari subcounty, Lira dis-
trict, Uganda. Afr J Ecol 45:39–43
Akinnifesi FK, Leakey RRB, Oluyede CA, Silishi G, Tchounjeu
Z, Matacala P, Kwesiga FR (2007) Indigenous fruit trees in
the tropics: domestication, utilization and commercializa-
tion. CABI, Oxfordshire
Albrecht A, Kandji ST (2003) Carbon sequestration in tropical
agroforestry systems. Agric Ecosyst Environ 99:15–27
Albuquerque UP, Andrade LHC, Caballero J (2005) Structure
and floristics of homegardens in Northeastern Brazil. J Arid
Environ 62:491–506
Andrews FW (1948) Vernacular names of plants—flowering
plants of the Anglo-Egyptian Sudan, vol 1. McCorcqodale
& Co. Ltd, Khartoum
Atta-Krah K, Kindt R, Skilton JN, Amaral W (2004) Managing
biological and genetic diversity in tropical agroforestry.
Agrofor Syst 61:183–194
Barbour GM, Burk HJ, Pitts WD (1987) Terrestrial plant ecol-
ogy. Cummings Publishing Company, Redwood City
Bardhan S, Jose S, Biswas S, Kabir K, Rogers W (2012)
Homegarden agroforestry systems: an intermediary for
biodiversity conservation in Bangladesh. Agrofor Syst
85:29–34
Bebawi FF, Neugebohrn L (1991) A review of plants of
Northern Sudan—with special reference to their uses.
Deutsche Gesellschaft fu
¨r Technische Zusammenarbeit
(GTZ), Eschborn
Bedigian D, Harlan JR (1983) Nuba agriculture and ethnobot-
any, with particular reference to sesame and sorghum.
Econ Bot 37:384–395
Bedri AAK (1984) Horticulture production in the Sudan, past,
present and future. Acta Hortic 143:25–29
Bennett-Lartey SO, Ayernor GS, Markwei CM, Asante IK,
Abbiw DK, Boateng SK, Anchirinah VM, Ekpe P (2001)
Contribution of home gardens to in situ conservation of
plant genetic resources farming systems in Ghana. In:
Watson JW, Eyzaguirre PB (eds) Proceedings of the sec-
ond international home gardens workshop, Witzenhausen,
Germany, pp 85–96
Bernholt H, Kehlenbeck K, Gebauer J, Buerkert A (2009) Plant
species richness and diversity in urban and peri-urban
gardens of Niamey, Niger. Agrofor Syst 77:159–179
Bisht IS, Rao KS, Bhandari DC, Nautiyal S, Maikhuri RK,
Dhillon BS (2006) A suitable site for in situ (on-farm)
management of plant diversity in traditional agroecosys-
tems of western Himalaya in Uttaranchal state: a case
study. Genet Resour Crop Evol 53:1333–1350
Blanckaert I, Swennen RL, Flores MP, Lopez RR, Saade RL
(2004) Floristic composition, plant uses and management
practices in homegardens of San Rafael Coxcatlan, Valley
of Tehuacan-Cuicatlan, Mexico. J Arid Environ 57:
179–202
Braun M, Burgstaller H, Hamadoun AM, Walter H (1991)
Common weeds of Central Sudan. Deutsche Gesellschaft
fu
¨r technische Zusammenarbeit (GTZ), Eschborn
Ceccolini L (2002) The homegardens of Soqotra island, Yemen:
an example of agroforestry approach to multiple land-use
in an isolated location. Agrofor Syst 56:107–115
Christanty L, Abdoellah O, Marten G, Iskandar J (1986) Tra-
ditional agroforestry in west Java: the pekerangan (home-
garden) and kebun–talun (annual–perennial rotation)
cropping systems. In: Marten GG (ed) Traditional agri-
culture in Southeast Asia: a human ecology perspective.
Westview Press, Boulder, pp 132–158
Colwell RK (2011) Estimates, CT, USA
Coomes OT, Ban N (2004) Cultivated plant species diversity in
home gardens of an Amazonian peasant village in north-
eastern Peru. Econ Bot 58:420–434
Drescher AW (1998) Hausga
¨rten in afrikanischen Ra
¨umen.
Bewirtschaftung nachhaltiger Produktionssysteme und
Strategien der Erna
¨hrungssicherung in Zambia und Zim-
babwe. Sozioo
¨konomische Prozesse in Asien und Afrika.
Centaurus, Freiburg
El Amin HM (1990) Trees & shrubs of the Sudan. Ithaca Press,
Exeter
560 Agroforest Syst (2014) 88:539–562
123
El Tahir BA, Gebauer J (2004) Non-timber forest products:
opportunities and constraints for poverty reduction in the
Nuba Mountains, South Kordofan, Sudan. Dtsch Tropentag
2004:10
El Tahir BA, Fadl KEM, Fadlalmula AGD (2010) Forest bio-
diversity in Kordofan region, Sudan: effects of climate
change, pests, disease and human activities. Biodiversity
11:34–44
Elsiddig EA (2007) Jebel Marra—The potentials for resources
and rural development in Darfur. University of Khartoum,
Faculty of Forestry, Shambat
Evans FC, Clark PJ, Brand RH (1955) Estimation of the number
of species present in a given area. Ecology 36:342–343
Eyzaguirre PB, Linares OF (2004) Home gardens and agrobi-
odiversity. Smithsonian Books, Washington
Fentahun M, Hager H (2009a) Integration of indigenous wild
woody perennial edible fruit bearing species in the agri-
cultural landscapes of Amhara region, Ethiopia. Agrofor
Syst 78:79–95
Fentahun MT, Hager H (2009b) Exploiting locally available
resources for food and nutritional security enhancement:
wild fruits diversity, potential and state of exploitation in
the Amhara region of Ethiopia. Food Secur 1:207–219
Fernandes ECM, Nair PKR (1986) An evaluation of the structure
and function of tropical homegardens. Agric Syst 21:279–310
Galluzzi G, Eyzaguirre P, Negri V (2010) Home gardens:
neglected hotspots of agro-biodiversity and cultural
diversity. Biodivers Conserv 19:3635–3654
Gautam R, Sthapit B, Subedi A, Poudel D, Shrestha P, Eyzag-
uirre PB (2008) Home gardens management of key species
in Nepal: a way to maximize the use of useful diversity for
the well-being of poor farmers. Plant Genet Resour Charact
Util 7:142–153
Gebauer J (2005) Plant species diversity of home gardens in El
Obeid, Central Sudan. J Agric Rural Dev Trop Subtrop
106:97–103
Gebauer J, El-Siddig K, El-Tahir BA, Salih AA, Ebert G,
Hammer K (2007) Exploiting the potential of indigenous
fruit trees: Grewia tenax in Sudan. Genet Resour Crop Evol
54:1701–1708
Goenster S, Wiehle M, Kehlenbeck K, Jamnadass R, Gebauer J,
Buerkert A (2011) Indigenous fruit trees in homegardens of
the Nuba Mountains, Central Sudan: tree diversity and
potential for improving the nutrition and income of rural
communities. In: Wesonga J, Kahane R (eds) First all
African horticultural congress. ISHS, Leuven, pp 355–364
Harragin S (2003) Nuba Mountains land and natural resources
study. Part I—land study. USAID-USDA PASA, in col-
laboration with University of Missouri, Tuskegee Univer-
sity and the University of Maryland Eastern Shore, www.
postconflict.unep.ch/sudanreport/sudan_website/doccatcher/
data/documents/Nuba Mountains Land and Natural Resour-
ces Study.pdf. Accessed 10 March 2013
Hashemi SF, Khoshbakht K, Mahdavi DA, Veisi H (2013) An
evaluation of agrobiodiversity in home gardens of two
ecogeographically different areas in Gachsaran, south-
western Iran. J Agric Sci 3:71–84
Henry C, Sharma M, Lapenu C, Zeller M (2003) Microfinance
poverty assessment tool. In: Institute IFPR (ed) CGAP/The
international bank for reconstruction and development.
The World Bank, Washington, DC, p 225
Houba VJG, van de Lee JJ, Novozamsky I (1995) Soil and plant
analysis, part 5B, Syllabus. Soil analysis procedures, other
procedures. Department of Soil Science and Plant Nutri-
tion, Wageningen Agricultural University, Wageningen
Hughes CE (1998) Leucaena: A genetic resources handbook.
Forestry Institute, University of Oxford, Oxford
Ismail IM, Elsheikh EA (2007) The woody vegetation of Rashad
district, South Kordofan—Sudan (Eastern Nuba Moun-
tains). Sudan Silva 13:78
Kabir ME, Webb EL (2009) Household and homegarden char-
acteristics in southwestern Bangladesh. Agroforest Syst
75:129–145
Kaihura FBS, Stocking M, Kahembe E (2001) Soil management
and agrodiversity: a case study from Arumeru, Arusha,
Tanzania. In: International symposium on managing bio-
diversity in agricultural systems, Montreal, Canada, p 14,
8–12 November 2001
Kehlenbeck K, Arifin H, Maass B (2007) Plant diversity in
homegardens in a socio-economic and agro-ecological
context. In: Tscharntke T, Leuschner C, Zeller M, Guhar-
dja E, Bidin A (eds) Stability of tropical rainforest margins.
Springer, Berlin, pp 295–317
Kindt R, Coe R (2005) Tree diversity analysis. A manual and
software for common statistical methods for ecological and
biodiversity studies. World Agroforestry Centre, Nairobi
Kovach WL (2001) MVSP: Multivariate statistical package
v3.12f. Kovach Computing Services, Pentraeth, Anglesey,
UK
Kumar BM, Nair PKR (2004) The enigma of tropical home-
gardens. Agrofor Syst 61–2:135–152
Leakey RRB, Simons AJ (1997) The domestication and com-
mercialization of indigenous trees in agroforestry for the
alleviation of poverty. Agrofor Syst 38:165–176
Leyer I, Wesche K (2007) Multivariate Statistik in der O
¨kolo-
gie—Eine Einfu
¨hrung. Springer, Berlin
Mahmoud MA, Khidir MO, Khalifa MA, El Ahmadi AMB,
Musnad HAR, Mohamed ETI (1996) Sudan: country report
to the FAO international technical conference on plant
genetic resources, Leipzig, Germany
Maroyi A (2009) Traditional homegardens and rural livelihoods
in Nhema, Zimbabwe: a sustainable agroforestry system.
Int J Sustain Dev World Ecol 16:1–8
McCune B, Grace JB (2002) Analysis of ecological communi-
ties. MjM Software Design, Gleneden Beach
Mendez VE, Lok R, Somarriba E (2001) Interdisciplinary ana-
lysis of homegardens in Nicaragua: micro-zonation, plant
use and socioeconomic importance. Agrofor Syst 51:85–96
Muneer ES (2008) Factors affecting adoption of agroforestry
farming system as a mean for sustainable agricultural
development and environment conservation in arid areas of
Northern Kordofan state, Sudan. Saudi J Biol Sci 15:
137–145
Munishi PKT, Philipina F, Temu RPC, Pima NE (2008) Tree
species composition and local use in agricultural land-
scapes of west Usambaras Tanzania. Afr J Ecol 46:66–73
Obeidalla AA, Riley JJ (1984) Agricultural Research Corpora-
tion, Western Sudan agricultural project, Development
of the horticultural potential of Kordofan region of
Sudan. Western Savannah Agricultural Research Project
(WSARP) No. 27. http://pdf.usaid.gov/pdf_docs/pnabe017.
pdf. Accessed 28 May 2014
Agroforest Syst (2014) 88:539–562 561
123
Okafor JC, Fernandes ECM (1987) Compound farms of South
Eastern Nigeria: a predominant agroforestry homegarden
system with crops and small livestock. Agrofor Syst 5:
153–168
Pandey CB, Lata K, Venatesh A, Mehdi RP (2006) Diversity
and species structure of home gardens in South Andaman.
Trop Ecol 47:251–258
Peyre A, Guidal A, Wiersum KF, Bongers F (2006) Dynamics of
homegarden structure and function in Kerala, India. Ag-
rofor Syst 66:101–115
Scales BR, Marsden SJ (2008) Biodiversity in small-scale
tropical agroforests: a review of species richness and
abundance shifts and the factors influencing them. Environ
Conserv 35:160–172
Shackleton CM, Paumgarten F, Cocks ML (2008) Household
attributes promote diversity of tree holdings in rural areas,
South Africa. Agrofor Syst 72:221–230
Soemarwoto O (1987) Homegardens: A traditional agroforestry
system with a promising future. In: Steppler HA, Nair PK
(eds) Agroforestry: a decade of development. ICRAF,
Nairobi, pp 157–170
Soemarwoto O, Soemarwoto I, Karyono, Soekartadiredja EM,
Ramlanand A (1985) The Javanese home garden as an
integrated agro-ecosystem. Food Nutr Bull 7:87
Sunwar S, Thornstrom CG, Subedi A, Bystrom M (2006) Home
gardens in western Nepal: opportunities and challenges for
on-farm management of agrobiodiversity. Biodivers Con-
serv 15:4211–4238
Thompson JL, Gebauer J, Hammer K, Buerkert A (2010) The
structure of urban and peri-urban gardens in Khartoum,
Sudan. Genet Resour Crop Evol 57:487–500
Tolera M, Asfaw Z, Lemenih M, Karltun E (2008) Woody
species diversity in a changing landscape in the south-
central highlands of Ethiopia. Agric Ecosyst Environ
128:52–58
USAID (2011) Conflict and expected poor harvest likely to result
in emergency levels of food insecurity. USAID, http://
reliefweb.int/sites/reliefweb.int/files/resources/sudan_ol_
2011_10_final.pdf. Accessed 03 Feb 2013
van Reeuwijk LP (1993) Procedures for soil analyses. Technical
paper, Van Fourth Edition, International Soil Reference
and Information Centre (ISRIC), Wageningen, pp 9
Vlkova M, Polesny Z, Verner V, Banout J, Dvorak M, Havlik J,
Lojka B, Ehl P, Krausova J (2011) Ethnobotanical
knowledge and agrobiodiversity in subsistence farming:
case study of home gardens in Phong My commune, central
Vietnam. Genet Resour Crop Evol 58:629–644
Wezel A, Ohl J (2005) Does remoteness from urban centres
influence plant diversity in homegardens and swidden
fields? A case study from the Matsiguenka in the Amazo-
nian rain forest of Peru. Agrofor Syst 65:241–251
Wiehle M, Prinz K, Kehlenbeck K, Goenster S, Mohamed SA,
Buerkert A, Gebauer J (2014) The role of homegardens and
forest ecosystems for domestication and conservation of
Ziziphus spina-christi (L.) Willd. in the Nuba Mountains,
Sudan. Genet Resour Crop Evol. doi:10.1007/s10722-014-
0124-3
Wiehle M, Prinz K, Kehlenbeck K, Goenster S, Mohamed SA,
Finkeldey R, Buerkert A, Gebauer J (submitted) The
African Baobab (Adansonia digitata L.)—genetic resour-
ces of neglected populations in the Nuba Mountains,
Sudan. Am J Bot
562 Agroforest Syst (2014) 88:539–562
123
