Forest Ecology and Management 233 (2006) 21–35

www.elsevier.com/locate/foreco

Litter production and organic matter accumulation in exclosures

of the Tigray highlands, Ethiopia
Katrien Descheemaeker a,*, Bart Muys b, Jan Nyssen a,c, Jean Poesen d,
Dirk Raes a, Mitiku Haile c, Jozef Deckers a
a
Division Soil and Water Management, Katholieke Universiteit Leuven, Geo-instituut, Celestijnenlaan 200E, BE-3001 Leuven, Belgium
b
Division Forest, Nature and Landscape, Katholieke Universiteit Leuven, Geo-instituut, Celestijnenlaan 200E, BE-3001 Leuven, Belgium
c
Department of Land Resource Management and Environmental Protection, Mekelle University, P.O. Box 231, Mekelle, Ethiopia
d
Physical and Regional Geography Research Group, Katholieke Universiteit Leuven, Geo-instituut,
Celestijnenlaan 200E, BE-3001 Leuven, Belgium
Received 12 January 2006; received in revised form 12 May 2006; accepted 23 May 2006

Abstract
To determine annual litter production of regenerating forest areas in the Tigray highlands of northern Ethiopia monthly litter production was
monitored over a two-year period in areas with varying degree of vegetation cover restoration. Total annual litter production varied from 30 to
425 g m
2 and increased significantly where areas were closed for a longer time. Litter production was depending on vegetation cover through an
exponential relation and was influenced also by soil fertility. Leaf litter typically constituted between 70 and 85% of total litter production, while
contributions of woody and reproductive litter varied according to species composition. Strong seasonality in litterfall was explained by
pronounced seasonal variation in rainfall. Standing crop of litter built up once an area was closed for grazing, increasing from around 20 g m
2 in
degraded grazing lands to nearly 600 g m
2 in an old exclosure. Litter accumulation was mainly determined by litter input, but was also influenced
by litter quality, species composition and microclimate development in the restoring forest areas. A detailed study of nine dominant shrub and tree
species revealed three distinct litter production patterns, corresponding to drought-deciduous species, evergreen species and (semi-)evergreen
Acacia species respectively.
# 2006 Elsevier B.V. All rights reserved.

Keywords: Litter accumulation; Litter quality; Microclimate; Succession; Tropical dry forest; Vegetation restoration

1. Introduction main objective is recovery of the natural vegetation cover with

Land degradation has reached an alarming state in the
highlands of Tigray, Northern Ethiopia, due to the combined
effects of deforestation, overgrazing, expansion of cropland and
unsustainable use of natural resources (Hurni, 1990; Herweg
and Stillhardt, 1999; Nyssen et al., 2004). As vegetation
restoration is conceived a powerful tool for environmental
rehabilitation (Hongo et al., 1995; Gao et al., 2002; Aerts et al.,
2004; Zhang et al., 2004), protected areas where agriculture and
grazing are forbidden were established in Tigray since the
1980s. These exclosures (Kebrom et al., 1997; Le Houérou,
2000; Wisborg et al., 2000; Tenna et al., 2001) are demarcated
in previously degraded areas, commonly on steep slopes. Their
* Corresponding author. Tel.: +32 16 329735; fax: +32 16 329760.
E-mail address: katrien.descheemaeker@biw.kuleuven.be
(K. Descheemaeker).
0378-1127/$ – see front matter # 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.foreco.2006.05.061
concomitant benefits for soil and water conservation (Deschee-
maeker et al., in press). Eweg et al. (1998) noted that the
recovery process in exclosures starts with a rapid increase in
diversity and cover of herbaceous species, quickly followed by
a rise in shrub and tree species abundance. Recently, exclosures
have become a controversial land use, suspect to cause more
pressure on already scarce grazing land in the study region.
Therefore, it is important to obtain more insight into the effects
of the set-aside policy on various aspects of ecosystems
dynamics. Litter production and organic matter accumulation
are part of these aspects as they are important processes
involved in nutrient cycling in terrestrial ecosystems (Vitousek,
1984; Lavelle et al., 1993; Martı́nez-Yrı́zar et al., 1999; Xuluc-
Tolosa et al., 2003). Litter production and accumulation also
contribute to humus formation, carbon sequestration and
soil fertility buildup. In general however, these processes
are not well documented for tropical semiarid forests
22 K. Descheemaeker et al. / Forest Ecology and Management 233 (2006) 21–35
(Martı́nez-Yrı́zar, 1995; Menaut et al., 1995). Notwithstanding
the importance of secondary forests in tropical countries, the
latter are addressed by few ecosystem studies as compared to
mature or climax forest ecosystems (Brown and Lugo, 1990).
More specifically, no information is available on the litterfall
patterns and litter accumulation of shrub and tree species and
vegetation types typical for restoring forests in exclosures of the
Ethiopian highlands.
The general aim of this study is therefore to gain under-
standing on litter production and organic matter accumulation
in regenerating vegetation in a tropical semiarid highland
environment.
Specific objectives are: (1) to determine annual litter
production and litter accumulation in exclosures of different
types and ages, (2) to characterize temporal variation in litter
production in these forest areas, and (3) to study litter
production, litter composition, litterfall seasonality and litter
accumulation of common shrub and tree species of the northern
Ethiopian highlands.
2. Methods
2.1. Study area
The study area is located in the Dogu’a Tembien (Tembien
highlands) district in Central Tigray, Ethiopia, near the district
capital Hagere Selam (138390 N, 398100 E, altitude 2650 m a.s.l.)
(Fig. 1). The lithology of the study area is composed of
Mesozoic sedimentary rocks (Adigrat sandstone, Antalo
limestone and Amba Aradam sandstone), covered by Tertiary
flood basalts (Bosellini et al., 1997). Mean annual rainfall is ca.
700 mm, mainly concentrated in the rainy season from June to
September (Nyssen et al., 2005). Within this area, two
Fig. 1. Location of the two study sites and the study region (indicated as a triangle) in the Northeast of Ethiopia (inset top right corner).
representative study sites were selected, located on different
lithologies, i.e. May Ba’ati (limestone, partly covered by non-
calcareous sediments) and Kunale (mixed sandstone and
limestone lithology, partly covered by transported montmor-
illonite clay material). The two study sites contain different
vegetation types, which are both typical for the study region.
This entails that the selection of two sites for this study is
justified.
Four hillslope sections (having surface areas ranging
between 1.5 and 8 ha) under different land use types (grazing
land, exclosures of different ages and a church forest) were
selected for comparative analysis in each study site (Fig. 1).
Within each land use type, different landscape units, more or
less homogeneous in soil type, slope gradient and overall
vegetation density were identified and demarcated. For more
variable hillslope sections this resulted in a higher number of
landscape units, causing an unbalanced field design with a total
of 21 landscape units. Each landscape unit may still consist of
different microsites, characterized by an internally homo-
geneous humus type. The fraction of the hillslope section
occupied by each landscape unit and in the same way, the
fraction of the landscape unit occupied by each microsite was
determined following the procedures described in Deschee-
maeker et al. (publication). Characteristics of landscape units
are summarized in Tables 1 and 2.
The grazing lands are degraded areas where overgrazing has
led to the disappearance of most vegetation so that severe soil
erosion is taking place, removing nearly all fertile soil. In the
rainy season, a grassy vegetation still develops, but shrubs and
trees are scarce and small (Table 2).
In the protected exclosure areas, vegetative soil cover,
degree of canopy closure, vegetation height and shrub canopy
area increase with the age of closure (Table 2). Average stem
 K. Descheemaeker et al. / Forest Ecology and Management 233 (2006) 21–35 23

Table 1
Topographic and soil-related characteristics of landscape units in the two study sites
Site Land Landscape Soil Slope Slope Clay Silt Sand WHCc BDd
use type unit typea gradient (%) aspectb (8) (%) (%) (%) (%) (Mg m
3)
May Ba’ati Grazing land MR1 Hypercalcic calcisol 20 130 30 32 38 13 1.39
MR2 Hypercalcic calcisol 40 130 30 32 38 13 1.39
Young exclosure MY1 Thaptocalci-humic calcisol 65 200 37 25 39 12 1.24
MY2 Hypercalcic calcisol 85 200 25 27 48 13 1.39
MY3 Thaptocalci-humic calcisol 35 200 48 40 12 12 1.24
Old exclosure MO1 Humi-endocalcaric phaeozem 30 135 52 32 16 13 1.21
MO2 Humi-calcaric cambisol 50 135 43 31 26 12 0.93
MO3 Humi-endocalcaric phaeozem 40 135 62 21 17 16 1.16
MO4 Humi-endocalcaric phaeozem 30 135 62 21 17 16 1.16
MO5 Humi-endocalcaric phaeozem 15 135 52 32 16 13 1.20
MO6 Humi-calcaric cambisol 50 135 43 31 26 12 0.93
MO7 Thaptocalci-humic calcisol 110 135 41 33 26 12 1.24
MO8 Humi-calcaric cambisol 70 135 43 31 26 12 0.93
Church forest MCF Humi-endocalcaric phaeozem 35 200 62 21 17 16 1.16
Kunale Grazing land KR Verti-humic cambisol 50 350 43 23 34 9 1.06
Young exclosure KY2 Humi-episkeletic phaeozem 55 315 41 21 38 14 1.01
KY3 Thaptocalci-humic cambisol 70 315 35 22 43 14 1.06
Middle-aged exclosure KM2 Humi-episkeletic phaeozem 55 315 41 21 38 14 1.01
KM3 Thaptocalci-humic cambisol 75 315 35 22 43 14 1.06
Old exclosure KO2 Thaptocalci-humic phaeozem 45 315 37 21 42 13 1.15
KO3 Thaptocalci-humic phaeozem 50 315 37 21 42 13 1.15
All soil variables refer to the topsoil (0–15 cm).
a
Classification according to the World Reference Base for Soil Resources (FAO et al., 1998).
b
Slope orientation in pole coordinates.
c
Water holding capacity.
d
Bulk density.

Table 2
Vegetation characteristics of landscape units in the two study sites
Site Landscape Age VEG CPAtot,1 CPAtot,2 H D Dmin Dmax # stems CPAshrub
unit (year) (%) (%) (%) (m) (cm) (cm) (cm) (m2)
May Ba’ati MR1 n.a. 20 1 1 0.3 0.6 (0.2) 0.2 1.2 5.4 (4.3) 0.1 (0.1)
MR2 n.a. 24 13 13 0.6 1.7 (0.9) 1.0 2.9 6.0 (6.2) 0.8 (0.6)
MY1 5 44 26 29 0.5 1.5 (2.2) 0.3 8.3 10.2 (9.6) 0.3 (0.4)
MY2 5 36 28 32 0.6 0.8 (0.6) 0.2 2.5 10.8 (8.0) 0.4 (0.3)
MY3 5 61 23 23 0.6 0.6 (0.3) 0.3 1.3 5.6 (5.4) 0.1 (0.2)
MO1 20 77 82 234 1.5 2.8 (2.9) 0.2 9.3 3.8 (2.6) 5.9 (8.5)
MO2 20 66 64 88 1.2 2.3 (2.5) 0.2 9.5 4.7 (5.3) 1.3 (2.2)
MO3 20 67 98 181 1.5 1.8 (1.8) 0.2 7.8 4.9 (3.3) 2.2 (1.8)
MO4 20 95 97 217 1.6 2.0 (1.3) 0.3 5.4 3.8 (5.0) 1.9 (2.6)
MO5 20 64 61 74 1.2 2.6 (2.3) 0.2 8.3 11.2 (14.3) 3.8 (3.2)
MO6 20 72 83 171 1.4 1.8 (1.3) 0.3 5.5 4.9 (5.7) 1.1 (1.1)
MO7 20 61 49 70 1.1 0.9 (0.7) 0.2 3.2 7.3 (5.5) 0.8 (0.9)
MO8 20 82 83 122 1.6 1.6 (1.4) 0.2 7.6 5.5 (5.0) 1.0 (1.9)
MCF n.a. 98 100 307 7.4 15.7 (9.8) 1.6 31.8 2.8 (1.5) 52.4 (49.3)
Kunale KR n.a. 50 2 2 0.3 1.0 (0.5) 0.5 1.6 12.0 (2.8) 0.2 (0.1)
KY2 5 74 40 49 0.9 1.1 (0.9) 0.3 3.8 4.5 (3.9) 0.5 (0.5)
KY3 5 62 37 47 0.8 1.0 (1.3) 0.2 7.6 9.7 (11.1) 0.3 (0.3)
KM2 14 69 48 60 1.4 1.5 (1.9) 0.4 9.5 5.4 (6.8) 0.6 (0.7)
KM3 14 61 50 60 1.1 0.9 (0.5) 0.2 2.5 6.5 (5.4) 0.4 (0.4)
KO2 21 95 77 151 1.5 2.0 (2.0) 0.2 5.9 5.9 (5.8) 3.5 (5.6)
KO3 21 88 100 132 4.4 28.0 (1.8) 26.7 29.3 2.0 (0.8) 89.2 (23.1)
Age: years since exclosure establishment. For the church forest and the grazing lands, this is not applicable (n.a.); VEG: total vegetative soil cover of shrub and tree
and grass and herb layer expressed as percentage of the total area; CPAtot,1: total canopy projection area of all shrubs and trees, with overlap left out, expressed as
percentage of the total area; CPAtot,2: total canopy projection area of all shrubs and trees, with overlap not left out, expressed as percentage of the total area
(percentages above 100 are caused by overlapping canopies); H: average height of the shrub and tree layer; D: average (standard deviation between brackets) stem
diameter; Dmin: minimal stem diameter; Dmax: maximal stem diameter; # stems: average (standard deviation between brackets) amount of stems per shrub or tree;
CPAshrub: average (standard deviation between brackets) canopy projection area of individual shrubs and trees, expressed as m2.
24 K. Descheemaeker et al. / Forest Ecology and Management 233 (2006) 21–35
diameter is small and multi-stemmed shrubs are common
(Table 2). In most landscape units several indigenous shrub
species are mixed and with age biodiversity of the shrub and
tree layer increases. Besides vegetation development, area
protection through exclosure leads to improvements in soil
fertility as well.
The church forest is a sacred place where the forest
vegetation has been protected since long time. Taller trees with
strongly overlapping canopies and thicker stems are found here
(Table 2). In the studied areas, few planted trees are found,
except for the Acacia saligna trees in Kunale and the occasional
Eucalyptus trees in both sites.
2.2. Experimental layout and field methods
2.2.1. Litter production
2.2.1.1. Litter production in exclosures. In each landscape
unit a representative area of 10 m  10 m was delineated and
equipped with 10 randomly placed litter traps. Besides that, in
each land use type, 10 traps were placed at regular distances
along a transect, which runs perpendicular to the contour,
intersecting the different landscape units. This extra sampling
along transects served to better grasp site variability. No
transect was laid out in the grazing lands since they only served
as a reference. Over the two study sites a total of 290 traps were
placed in July 2003 and emptied monthly until June 2005.
Locally available metal cans (16 cm diameter and 10 cm
height) with perforated bottoms to evacuate rainwater were
used as litter traps. The advantages of small traps are discussed
by Búrquez et al. (1999). Economies in terms of time and
money allowed for increasing the number of sample units,
which is especially valuable when studying heterogeneous
environments. The fact that the traps were inconspicuous and
cheap reduced the risk of vandalism or theft. Theft of the traps
could nevertheless not be excluded, especially in the grazing
lands. Missing traps were replaced at every monthly sampling
event and missing records were substituted by the average of
the previous month, the next month and the same month in the
other year of study at the same location. Four percent of the
total monthly samples had to be treated in that way. If a large
branch fell in the can it was broken off to remain with a length
not exceeding the can diameter. The harvested litter was
separated from animal debris and faeces and sorted according
to species and litter component (leaves, woody debris,
reproductive parts). After drying at 65 8C until constant weight
all samples were weighed to 0.1 mg.
2.2.1.2. Litter production on a species basis. Through sorting
of the collected litter on a species basis, the species contribution
to annual litter production was first determined per landscape
unit and ultimately per land use type. Already after the first year
of observation, data analysis revealed which shrub and tree
species were most contributing to annual litter production in the
different sites and exclosures. Based on this result the two
dominating tree or shrub species for each land use type were
selected and their litter production was studied for one year in a
separate observational study. This was done to obtain a more
detailed understanding of their litter production seasonality and
the contribution of different litter components. For each
selected species, five live individuals were chosen in the land
use type where the species was dominant, spread over different
landscape units. The criteria for their selection were the good
state of the canopy, the absence of indications of limited growth
or plant damage and no intermixing with the canopy of other
species. Per individual, three litter traps were placed midway
between the stem and the edge of the canopy. They were
emptied monthly between September 2004 and August 2005
and treated as described above. The average litter production
per species was calculated, providing an indication on the litter
production that could be attained under a hypothetical
homogenous and closed canopy of the selected species in
the land use types where they are dominant.
2.2.2. Standing crop of litter
2.2.2.1. Standing crop of litter in exclosures. Standing crop
of litter, defined as the total weight of all ectorganic layers, was
determined for landscape units and land use types. To account
for variability within the landscape units, transects perpendi-
cular to the contour were laid out every 25 m in each hillslope
section. Every microsite along the transects was recorded and
grouped with similar ones according to litter cover and species
composition. For microsite groups covering at least 2% of a
landscape unit area, one microsite was randomly selected and
subsequently sampled in the field. A metal frame
(25 cm  25 cm) was used to take a sample of the ectorganic
layers in a total of 135 microsites. All ectorganic layers
enclosed by the frame were carefully removed one by one and
collected. Afterwards, the samples were sorted according to
litter components, dried and weighed in the laboratory. Based
on the area contribution of each microsite, the standing crop of
litter per landscape unit was determined. Further upscaling to
standing crop of litter for the land use types was done based on
the area contributions of landscape units to their respective
hillslope sections. As a result of seasonal differences in litter
production and decomposition the standing crop of litter varies
substantially throughout the year. Given that the sampling was
performed in September 2004, before the biggest peak in litter
production and after a period of high decomposer activity in the
rainy season, the minimum standing crop of litter is reported.
2.2.2.2. Standing crop of litter for selected species. Standing
crop of litter was determined for the selected dominant species
in September 2005. For every species, three random litter layer
samples were taken under the canopy of the five individuals
monitored for litter production using a 25 cm  25 cm frame.
The collected litter samples were sorted, dried and weighed.
2.2.2.3. Leaf litter and soil characteristics. Specific leaf mass
was determined for a broad range of shrub and tree species
(Table 4) on fresh leaf litter samples cut into squares of 0.25, 1
of 4 cm2 according to the leaf size. Macronutrient content was
also determined on fresh litter of these species. The Kjeldahl
method (McGill and Figueiredo, 1993) was used for total N.
 K. Descheemaeker et al. / Forest Ecology and Management 233 (2006) 21–35 25

After extraction, P was determined colorimetrically using the
ascorbic acid–molybdenum blue method (Olson and Sommers,
1982), while for K, atomic absorption spectroscopy was used
(Anderson and Ingram, 1993). Organic carbon content was
determined using loss on ignition (Van Reeuwijk, 2002). To
estimate average leaf litter characteristics for a landscape unit
or land use type as a whole, a weighted average was calculated
for C, N, P and specific leaf mass taking into account the
contributions of the species to total annual litter production of
that area.
Procedures for soil characterization are discussed in
Descheemaeker et al. (publication).
2.3. Data analysis
Mean monthly litter production per landscape unit together
with corresponding standard errors was estimated by averaging
the harvested masses in the 10 litter traps. Annual litter
production was obtained by averaging the annual totals of the
collected masses in the traps. The minimum number of litter
traps per landscape unit that was needed for a representative
sample was determined by plotting the running mean of the
annual litter production against an increasing number of traps.
The minimum amount of traps per landscape unit necessary to
limit the oscillation of the mean to 10% of the final value varied
from three for the church forest to seven to nine for the
exclosures and nine for the grazing land, so that the used
experimental layout seemed sufficient to yield a reliable
estimate of litter production per land use type. After a log-
transformation of the data to obtain a normal distribution,
ANOVA was used to test for the effect of site and land use type
on total annual litter production. Multiple pairwise compar-
isons among the land use types were done with Tukey’s test.
Contributions of different species to annual litter production of
landscape units were determined based on the collected litter
separated per species. Upscaling to the land use type level was

Table 3
Litter production (g m
2) for litter components in different land use types in the two study sites for the period July 2003–June 2004 (y1) and July 2004–June 2005 (y2),
average annual litter production (av) and percentage contribution of component to total annual litterfall (%)
Study site Land use typea Leaves Woody litter Reproductive organs Total
y1 y2 av % y1 y2 av % y1 y2 av % y1 y2 avc
May Ba’ati RA (n = 20)
avb 14.5 28.4 21.4 68.3 7.8 4.7 6.2 19.9 5.5 1.9 3.7 11.8 27.8 34.9 31.3 a
CVb 0.66 0.92 0.65 1.07 0.88 0.75 0.88 1.31 0.74 0.49 0.76 0.48
YE (n = 40)
av 71.8 89.8 80.8 76.8 12.8 14.0 13.4 12.7 11.4 10.7 11.1 10.5 96.0 114.5 105.2 b
CV 0.21 0.27 0.17 0.53 0.52 0.37 0.68 0.60 0.46 0.22 0.25 0.17

OE (n = 100)
av 158.4 162.4 160.4 76.0 39.2 22.4 30.8 14.6 19.8 20.0 19.9 9.4 217.3 204.8 211.0 c
CV 0.10 0.12 0.08 0.34 0.34 0.25 0.54 0.40 0.33 0.11 0.12 0.08

CF (n = 30)
av 406.2 327.1 366.7 86.4 51.1 14.2 32.7 7.7 40.8 9.4 25.1 5.9 498.1 350.8 424.4 d
CV 0.20 0.20 0.14 0.76 1.25 0.65 0.68 0.80 0.57 0.21 0.20 0.15
Kunale RA (n = 10)
av 31.0 18.0 24.5 82.5 6.9 0.1 3.5 11.7 1.4 2.1 1.7 5.9 39.2 20.2 29.7 a
CV 0.72 0.89 0.56 1.08 2.65 1.00 1.04 1.49 0.99 0.67 0.81 0.51

YE (n = 30)
av 96.8 71.5 84.1 83.9 8.5 4.6 6.5 6.5 11.8 7.3 9.6 9.5 117.0 83.4 100.2 ab
CV 0.26 0.28 0.19 0.56 0.88 0.48 0.50 0.54 0.38 0.23 0.25 0.17

ME (n = 30)
av 136.1 123.2 129.7 81.1 24.2 15.1 19.7 12.3 15.2 6.1 10.6 6.6 175.5 144.4 160.0 b
CV 0.21 0.25 0.16 0.89 1.26 0.73 0.53 0.96 0.47 0.22 0.25 0.17

OE (n = 30)
av 274.7 187.8 231.3 73.2 73.2 31.5 52.4 16.6 45.5 19.2 32.4 10.2 393.4 238.6 316.0 c
CV 0.20 0.32 0.18 0.86 0.41 0.61 0.36 0.74 0.33 0.22 0.26 0.17
a
Land use type coding: RA: grazing land; YE: young exclosure; ME: middle-aged exclosure; OE: old exclosure; CF: church forest.
b
av: average; CV: coefficient of variation.
c
Average values of total annual litter production are significantly different according to a post hoc Tukey test at the 0.05 level if they have no letters in common
(tested separately for both study sites).
26 K. Descheemaeker et al. / Forest Ecology and Management 233 (2006) 21–35

done based on the area contributions of the landscape units to

the hillslope section to which they belong. Correlation and
regression analysis were applied to investigate relations
between variables. Statistical analyses were performed using
SPSS software (SPSS, 2001).

3. Results

3.1. Annual litter production in land use types

Important effects of wind on litter dispersal were not

observed as the species of the collected litter per can
corresponded with the surrounding vegetation. Total annual
litter production in the different land use types in May Ba’ati
varies from 30  15 g m
2 in the grazing land to
425  64 g m
2 in the church forest (Table 3). ANOVA
indicated a significant effect of land use type, but no effect of
site. Multiple comparisons with Tukey’s test revealed that all
land use types in May Ba’ati have significantly different total
litter production (Table 3), increasing with regeneration stage
and reaching a maximum in the church forest. The same applies
for the land use types in Kunale, although litter production in
the young exclosure was not found significantly different from
the grazing land and the middle-aged exclosure (Table 3). Since
age of exclosure goes hand in hand with age of vegetation,
phase of succession and soil fertility build up, it is not surprising
that in both study sites litter production strongly increases when
a degraded grazing land is closed and the natural vegetation is
allowed to restore. An exponential regression function
describes the dependence of annual litter production on
vegetation cover (Fig. 2) and Fig. 3 shows that there is also
Fig. 3. Average annual litter production and soil nutrient content for different
an effect of soil fertility on litter production when vegetation is land use types expressed relative to the average value obtained for the grazing
restoring. A paired samples t-test on the annual litter production lands. Standard deviations are indicated with error bars (land use type coding,
of landscape units revealed that litter production in the first year RA: grazing land; YE: young exclosure; ME: middle-aged exclosure; OE: old
was significantly higher than in the second year (n = 21; exclosure; CF: church forest).
p = 0.001). On average, litter production in the second year
amounts to only 87% of the first year’s production.
In all land use types leaf litter comprised 70–85% of total
litterfall (Table 3). Reproductive parts do not produce much
more than 10% of the total in each of the land use types, while
the contribution of woody litter ranges between 6 and 20%
(Table 3).
Coefficients of variation (CV) of mean annual litter
production are rather high, which illustrates the high variability
in vegetation density which is typical for the studied areas.
However, CVs vary greatly according to litter component and
land use type. For woody and reproductive litter the CV is
highest, varying between 0.25 and 1, while for mean total litter
production and leaf litter, the highest CV does not exceed 0.65.
Grazing lands in both study sites yield the highest CVs, whereas
CVs of total annual litter production in exclosures and the
church forest lie below 0.20.

3.2. Temporal variation of litterfall per land use type

Fig. 2. Annual litter production versus total vegetation cover for all landscape Litter production in the regenerating forest areas and church
units with plotted regression line. forest is characterized by a unimodal temporal pattern (Figs. 4
 K. Descheemaeker et al. / Forest Ecology and Management 233 (2006) 21–35
Fig. 4. Monthly rainfall and temporal variation of monthly litterfall
(g m
2 mo
1  1 standard error) in Kunale for the period July 2003 to June
2005 (total litter: *; leaf litter: *; woody litter: !; reproductive litter: 5).
and 5). In the grazing lands high coefficients of variation
indicate high spatial and temporal variability. For the
exclosures of Kunale, leaf litter production starts rising in
September, comes to a peak in December and January and
declines afterwards to minimal amounts in the rainy season
(Fig. 4). Reproductive litter production in Kunale is mostly
highest from October to February, while for woody litter there
is no clear temporal trend discernable. Leaf litter production in
both the young and the old exclosure in May Ba’ati peaks after
the rainy season until February. Just as in Kunale, the rainy
27
Fig. 5. Monthly rainfall and temporal variation of monthly litterfall
(g m
2 mo
1  1 standard error) in May Ba’ati for the period July 2003 to June
2005 (total litter: *; leaf litter: *; woody litter: !; reproductive litter: 5).
season is characterized by low litterfall. The church forest of
May Ba’ati deviates from the general trend of high leaf litter
production after the summer rains, with the highest production
recorded from January to April. In the old exclosure of May
Ba’ati several small peaks in reproductive litter production are
recorded. For the young exclosure however, peaks in
reproductive litter are clearly in September and December,
while for the church forest only in the first year of the
experiment, a peak was recorded from November to February
(Fig. 5).
28 K. Descheemaeker et al. / Forest Ecology and Management 233 (2006) 21–35

3.3. Contributions of shrub and tree species to total annual
litterfall
In the two years of study litter belonging to 98 species was
captured in the traps. Sorting of the collected litter according to
species allowed to determine species contribution to total
annual litterfall in the different land use types (Table 4). Most of
the collected species have a negligible contribution to total litter
production. Species contributing considerably to litter produc-
tion are clearly different between the two study sites (Table 4).
Dombeya torrida, Acacia abyssinica, Salvia schimperi and
planted Acacia saligna are important litter producers in the
exclosures of Kunale, while calciphile species such as Euclea
racemosa, Dodonea angustifolia and A. etbaica dominate
litterfall in the May Ba’ati exclosures (Table 4). These
differences can be explained by the ecological preferences
of the different species in relation to site conditions such as soil
nutrient contents and moisture availability. Species composi-
tion of the church forest differs considerably, as illustrated
through the high contribution to total litter production by
climax species like Acokanthera schimperi, Rhus glutinosa,
Carissa edulis and Ficus thonningii. Vegetation composition of
exclosures evolves with time, which is illustrated by the varying
species contributions to litter production in exclosures of
different ages. The most obvious example is given by grasses,
showing their biggest share of total litter production in the
young exclosures of both study sites. As vegetation is restoring,
grasses are suppressed by a rise in shrub abundance, which

Table 4
Procentual contribution of species to total litter production in different land use types from 1 July 2003 to 30 June 2005 in May Ba’ati and Kunale, with indication (*)
for species, which have been analyzed for leaf litter characteristics
Species Family May Ba’atia Kunale a
RA YE OE CF RA YE ME OE
Acacia abyssinica Hochst. ex Benth. Fabaceae * 0.0 0.0 0.1 0.9 0.0 0.7 9.1 16.3
Acacia etbaica Schweinf. Fabaceae * 28.8 17.2 16.1 0.0 0.0 0.0 0.0 0.0
Acacia saligna (Labill.) Wendl§ Fabaceae * 0.0 0.0 0.7 0.0 0.0 0.0 14.0 23.0
Acokanthera schimperi (A. DC.) Benth. Apocynaceae * 0.0 0.0 0.0 26.8 0.0 0.0 0.0 0.0
Aloe macrocarpa Tod. Asphodelaceae * 1.6 0.5 0.2 0.2 30.8 0.9 1.1 0.7
Becium grandiflorum (Lam.) Pichi-Serm. Lamiaceae * 4.4 30.7 0.0 0.0 0.3 18.0 0.0 0.0
Buddleja polystachya Fresen. Loganiaceae 0.0 0.0 0.0 0.3 0.0 0.0 0.0 6.9
Calpurnia aurea (Ait.) Benth. Fabaceae 0.0 0.0 0.3 1.1 0.0 0.0 0.0 0.0
Carissa edulis (Forssk) Vahl Apocynaceae * 0.0 0.0 4.0 12.5 0.0 0.0 0.0 0.0
Danderb Asteraceae 8.8 7.7 0.0 0.0 0.2 3.9 0.0 0.0
Cayratia gracilis (Guill. & Perr.) Suess. Vitaceae 0.0 0.0 0.6 1.1 0.0 0.0 1.9 3.3
Colutea abyssinica Kunth & Bouche Fabaceae 0.0 0.0 0.0 0.0 0.0 0.0 0.1 1.3
Commelina spp c Commelinaceae 0.0 0.0 0.0 2.7 0.0 2.1 0.0 2.0
Dodonea angustifolia L.f. Sapindaceae * 0.0 12.3 14.6 0.0 0.0 0.0 0.0 0.9
Dombeya torrida (J.F. Gmel.) P. Bamps Sterculiaceae * 0.0 0.0 0.8 0.7 0.0 1.2 14.2 22.9
Eucalyptus camaldulensis Dehn§ Myrtaceae * 0.8 0.0 3.9 0.0 0.0 5.0 0.0 0.0
Euclea racemosa Murr. Ebenaceae * 36.7 9.5 29.4 7.2 0.0 0.0 0.0 0.0
Ficus thonningii Blume Moraceae * 0.0 0.0 0.0 10.8 0.0 0.0 0.0 0.0
Grassesd Poaceae * 6.9 10.0 4.2 0.0 13.6 20.0 10.2 2.5
Grewia ferruginea A. Rich. Tiliaceae * 0.0 0.0 0.1 0.7 0.0 0.0 1.7 2.1
Jasminum abyssinicum Hochst. ex. DC. Oleaceae 0.0 0.0 4.8 0.0 0.0 0.0 0.0 0.0
Lantana viburnoides (Forsk.) Vahl Verbenaceae 0.0 1.1 1.7 0.0 0.0 1.7 3.9 0.7
Maytenus arbutifolia (Hochst. ex A. Rich.) R. Wilczek Celastraceae * 0.0 0.0 2.0 0.5 12.7 1.6 10.0 0.1
Olea europaea subsp. cuspidata (Wall. Ex. DC) Ciffieri Oleaceae * 0.0 0.0 0.0 3.2 0.0 0.0 0.0 0.0
Osyris quadripatita Salzm. ex Decaisne Santalaceae 0.0 0.0 1.0 0.0 0.0 6.3 0.5 0.1
Peristrophe paniculata (Forssk) Brummitt Acanthaceae 4.4 1.8 0.3 0.0 3.2 0.1 0.2 0.4
Psydrax schimperiana (A. Rich.) Bridson Rubiaceae 0.0 0.0 0.0 4.8 0.0 0.0 0.0 0.0
Rhus glutinosa Hochst. ex A. Rich. Anacardiaceae * 0.0 0.0 0.6 18.3 0.0 0.0 0.0 1.3
Rhus natalensis Krauss Anacardiaceae 0.0 0.0 1.3 0.0 0.0 0.0 0.0 0.0
Rumex nervosus Vahl. Polygonaceae * 0.0 0.2 0.0 0.0 26.7 10.9 3.6 1.8
Salvia schimperi Benth. Lamiaceae * 0.0 4.5 1.8 0.0 0.0 20.9 5.8 0.1
Senna singueana (Del.) Lock Fabaceae 0.8 1.4 2.7 0.0 0.0 0.0 0.0 0.0
Vernonia spp. e Asteraceae * 0.0 0.5 4.7 0.0 0.0 1.2 17.7 7.1
Unknown species 0.2 1.2 1.3 0.8 0.7 2.1 3.0 3.3
Variaf 6.7 1.4 3.4 8.3 11.8 3.4 2.9 3.0
a
Land use type coding: RA: grazing land; YE: young exclosure; ME: middle-aged exclosure; OE: old exclosure; CF: church forest.
b
Dander is the local name for a series of thistles including Carthamus lanatus L., Carduus nyassanus R.E. Fries, Echinops hispidus Fresen.
c
Commelina spp. include C. imberbis Ehrenb. ex Hassk., C. benghalensis L., C. africana L., C. latifolia Hochst. ex A. Rich.
d
Grasses include Chloris gayana Kunth., Cymbopogon caesius (Nees) Stapf., Cynodon dactylon (L.) Pers., Pennisetum petiolare (Hochst.) Chiov., Themeda
triandra Forssk.
e
Vernonia spp. include V. bipontini Vatke and V. amygdalina Del.
f
Varia is a collection of less important species contributing each less than 1% in every land use type.
§
Planted species.
 K. Descheemaeker et al. / Forest Ecology and Management 233 (2006) 21–35 29

results in lower grass contributions to total litterfall in older
exclosures. The church forest illustrates the ultimate result of
this evolution, as grass litter production is almost nil there. Also
Rumex nervosus, a typical pioneer species exemplifies the trend
of evolving species composition with time. This species is
responsible for one quarter of the total litter production in the
grazing land of Kunale. In the young exclosure, the contribution
is already reduced to 10%, while in the older exclosures of
Kunale the shrub is oppressed by others and its contribution is
driven back to a puny 2 and 3%. The opposite evolution is
illustrated by a climax species such as C. edulis, which is absent
in the grazing land and young exclosure of May Ba’ati, but
appearing in the old exclosure with a contribution of 4% of total
litter production. In the church forest its share amounts to 12%.
3.4. Litterfall for selected species
Of the nine selected species for further investigation of their
litter production, four are dominant in May Ba’ati, another four
are dominant in Kunale and one (Becium grandiflorum) is
important in both sites. Highest total annual litter production was
recorded under the canopy of A. etbaica, closely followed by E.
racemosa and D. torrida (Table 5). Lowest annual production
was recorded for B. grandiflorum, and intermediate productions
for the five other species. In combination with specific leaf mass,
annual leaf litter production can provide an idea on leaf area
index (LAI). The results in Table 5 indicate that D. torrida is
characterized by the highest LAI (3.74) of the nine selected
species. Distinctly lower LAIs around 0.5 are obtained for A.
abyssinica, A. saligna and B. grandiflorum (Table 5). Absolute
values for woody litter production are clearly highest for the
native Acacia species, while reproductive litter is abundant under
the canopies of all three Acacia species (Table 5).
Fig. 6 presents the pattern of litter production for the nine
selected species in a period of one year. The evergreen shrubs E.
racemosa and C. edulis, found in the exclosures of May Ba’ati
show a similar litter production pattern. They have a continuous
turnover of leaves, but a marked increase of leaf-fall before the
first rains (Fig. 6), when leaf renewal starts. Their woody litter
production is more or less constant over the year with some
occasional increases and the contribution of reproductive litter
to total annual litter mass is negligible (Fig. 6). A. etbaica is a
semi-evergreen shrub, and is characterized by a marked pulse of
leaf and woody litter production in March, clearly contrasting
to lower production in other months (Fig. 6). Flowers are
produced at the start of the dry season from October to
December and fruiting takes place form February to June
(Fig. 6). D. angustifolia renews its leaves near the start of the
rainy season and leaf litter production quickly rises from
August to a peak in October and November. Fruits are shed in
March and April (Fig. 6). The deciduous shrubs B. grand-
iflorum and S. schimperi, which are common in both Kunale
and May Ba’ati are characterized by a rise in leaf litter
production from the end of the rainy season to a peak in
November. For S. schimperi, high leaf litter production extends
until March (Fig. 6). After this peak, these shrubs stay more or
less bare until the first rains when their canopy is renewed. S.
schimperi shows a small peak in reproductive litter production
in November, while B. grandiflorum produces fruits and flowers
more vigorously from September to November (Fig. 6). The
introduced tree A. saligna and the indigenous A. abyssinica,
both important species in the Kunale exclosures, show a similar
pattern in reproductive litter production, with flowering starting
in November, followed by a peak in fruit production in January-
February (Fig. 6). Leaf litter production of the evergreen A.
saligna starts increasing after the rains have stopped until
January and is lowest in the wet months of July and August
(Fig. 6). Woody litter production is very low the whole year
around, which cannot be said for A. abyssinica, which has the
highest contribution of woody components to total litter of all
studied species (Fig. 6). A. abyssinica is semi-evergreen, with
new leaves appearing starting from the first rains. The tree
sheds its leaves from November onwards, peaking in March
(Fig. 6), just as the other indigenous A. etbaica in May Ba’ati.
All studied species except the Acacia species, have leaf litter
contributions to total litter production of at least 85% and
varying contributions of woody and reproductive litter (Fig. 6).
The introduced species A. saligna has the highest contribution
of reproductive litter, while the indigenous A. etbaica and A.
abyssinica have markedly higher woody litter shares (Fig. 6).

Table 5
Annual litter production for selected species (g m
2) (average values (av) with coefficient of variation (CV)), split up per litter component and an estimate for leaf area
index (LAI)
Species Leaves Woody Reproductive Total LAI
a a a a
av CV av CV av CV av CV
Acacia abyssinica 47.0a 0.49 141.9a 0.52 46.0ab 2.01 235.0ab 0.51 0.57
Acacia etbaica 172.4ab 0.32 118.9ab 0.28 51.7a 0.90 343.0a 0.15 1.47
Acacia saligna 126.1a 0.45 4.6cd 0.87 77.4a 0.93 208.1ab 0.50 0.48
Becium grandiflorum 86.5a 0.73 4.4cd 0.51 13.3ab 1.17 104.3b 0.74 0.51
Carissa edulils 161.4ab 0.41 19.7cd 1.35 0.2b 2.24 181.3ab 0.45 1.03
Dodonea angustifolia 144.0ab 0.19 5.0cd 1.39 3.3ab 0.55 152.2ab 0.20 1.18
Dombeya torrida 297.0b 0.17 5.2cd 1.71 1.1b 1.67 303.3a 0.17 3.74
Euclea racemosa 300.3b 0.51 20.5bc 0.41 0.2b 1.61 321.0a 0.48 1.63
Salvia schimperi 163.5ab 0.69 1.9d 1.42 3.7ab 1.87 169.1ab 0.66 0.93
a
Average values are significantly different according to a post hoc Tukey test at the 0.05 level if they have no letters in common.
 30
K. Descheemaeker et al. / Forest Ecology and Management 233 (2006) 21–35
Fig. 6. Temporal variation of monthly litterfall (g m
2 mo
1  1 standard error) for selected species for the period September 2004 to August 2005 and relative contribution (%) of each litter component to total annual
litter production (total litter: *; leaf litter: * and light grey bar color; woody litter: ! and black bar color; reproductive litter: 5 and dark grey bar color).
 K. Descheemaeker et al. / Forest Ecology and Management 233 (2006) 21–35 31

Table 6 Table 7
Standing crop of litter for litter components in the different land use types in Standing crop of litter for different litter components for a selection of shrub and
Kunale and May Ba’ati and percentage contribution of component to total tree species and percentage contribution of component to total standing crop of
standing crop of litter litter
Land use typea Leaves Woody Reproductive Total Species Leaves Woody Reproductive Total

2
2
2
2
2
2
2
gm % gm % gm % gm gm % gm % gm % g m
2
May Ba’ati Acacia abyssinica 1 0 208 95 10 4 218
RA 17.6 86.7 2.1 10.3 0.6 3 20.3 Acacia etbaica 72 32 109 49 41 19 222
YE 83 80.1 19.6 18.9 1.1 1 103.7 Acacia saligna 213 66 67 21 41 13 322
OE 497.5 85.2 81.3 13.9 5.4 0.9 584.3 Becium grandiflorum 14 36 23 61 1 3 38
CF 701.9 53.6 608.5 46.4 0 0 1310.4 Carissa edulils 193 95 10 5 0 0 204
Dodonea angustifolia 233 93 14 6 2 1 249
Kunale
Dombeya torrida 139 81 33 19 0 0 172
RA 13.8 93.8 0.7 4.6 0.2 1.6 14.7
Euclea racemosa 709 90 77 10 1 0 786
YE 85.1 69.6 35 28.6 2.2 1.8 122.3
Salvia schimperi 52 57 40 43 0 0 92
ME 99 52.4 87.8 46.5 2 1.1 188.9
OE 261.4 61 138.4 32.3 28.8 6.7 428.6
a
Land use type coding, RA: grazing land; YE: young exclosure; ME: middle-
aged exclosure; OE: old exclosure; CF: church forest. has the highest share of woody litter. Total standing crop of
litter is significantly determined by litter input as is revealed by
the regression function in Fig. 7.
3.5. Standing crop of litter in land use types
3.6. Standing crop of litter under selected species
Total standing crop of litter increases with the age of the
exclosure and varies in May Ba’ati from around Total standing crop of litter varies greatly among species from
100  10 g m
2 in the young exclosure to 1300  95 g m
2 38 g m
2 for B. grandiflorum to 786 g m
2 for E. racemosa
in the church forest. In Kunale total standing crop ranges from (Table 7). Also here litter accumulation is significantly correlated
120  37 to 430  109 g m
2, while in the grazing lands of with total litter input (r = 0.48, p = 0.001, n = 45). The share of
both sites, nearly no litter accumulation takes place, as standing the different litter components is related to their relative input, so
crops do not exceed 20  8 g m
2 (Table 6). Contributions of that it is not surprising to find high woody litter contributions for
different litter components to total standing crop vary A. abyssinica and A. etbaica (Table 7), known for their high
considerably between the studied sites. Leaf litter contribution woody litter production (Table 5, Fig. 6). Also the decay rate of
is highest in the degraded grazing lands. In the exclosures leaf the different litter components plays a role. Generally speaking
litter contribution ranges from 52 to 85%, with the May Ba’ati woody litter is more difficult to decompose, which results in a
exclosures having higher leaf litter contributions than those of higher accumulation than would be expected based solely on the
Kunale. The old exclosure of Kunale is an exception with production rate.
respect to the relatively high share of reproductive litter,
amounting to 7%, which is attributed to the abundance of pod- 4. Discussion
producing species like A. abyssinica and A. saligna. Together
with the middle-aged exclosures of Kunale, the church forest Because of high inter-annual variation in litter production
and strong variability in contributions of the different litterfall
fractions, Proctor (1983) stresses the importance of extending
litter production measurements beyond a single year. The
validity of this concern is also evident from our results, as it was
demonstrated that total annual litter production was signifi-
cantly higher in the first year as compared to the second year. In
this case, the explanation for this trend is higher rainfall in the
first as compared to the second year. During the main rainy
season (June to September) of 2003, rainfall in Kunale and May
Ba’ati amounted to 499 and 437 mm respectively, while in the
same months of 2004 only 390 mm fell in both sites. The
difference in total rainfall between both years is smaller in May
Ba’ati, which could be the reason for the smaller differences in
total litter production that were recorded here as compared to
Kunale (Table 3). For a dry forest in Puerto Rico, Cintrón and
Lugo (1990) detected a direct response of litterfall to annual
variations in rainfall and also Songwe et al. (1995) stressed that
Fig. 7. Standing crop of litter versus annual litter production for all landscape the annual rainfall pattern can cause strong interannual
units with plotted regression line. variation in litter production.
32 K. Descheemaeker et al. / Forest Ecology and Management 233 (2006) 21–35
Martı́nez-Yrı́zar (1995) gives an overview of litter produc-
tion in dry tropical forests. For environments with annual
rainfall depths similar to our study region, annual leaf litter
production varies from 250 to 480 g m
2, while total annual
litter production amounts to 290 to 690 g m
2. Only the old
exclosure in Kunale and the church forest in May Ba’ati have
litter productions within this range (Table 3). The relatively low
production can be explained by the fact that the land use types
under study were degraded areas before they became protected
as exclosures. Within a time span of 10–20 years soil fertility
increased, but is still relatively low given the degraded status
with completely depleted soils from which the exclosures
originated. The low production rates indicate that the younger
exclosures of both sites, and also the old exclosure of May
Ba’ati are still in an early successional stage, with a vegetation
density and soil fertility status not yet at the same level of a
mature dry forest. In an experiment monitoring the environ-
mental effects of agricultural abandonment in the Mediterra-
nean region, López-Bermúdez et al. (1998) recorded a litter
production of 250 g m
2 in an abandoned field (of non-
mentioned age), which is close to the 211 g m
2 produced
annually in the old exclosure of May Ba’ati. The exponential
relation between vegetation cover and litter production (Fig. 2)
merits some extra attention. If an increase in vegetation cover
would merely involve horizontal canopy closure, a more linear
rise in annual litter production would be expected. However,
vegetation restoration goes hand in hand with an increase in
vegetation height as well, leading to a more than linear rise in
leaf area index. Besides this, an altered species composition is
also known to explain the exponentially rising litter production
in older successional stages (Sundarapandian and Swamy,
1999). However, the effect of soil fertility cannot be neglected
to explain the exponential rise in litter production. With
exclosure development a degraded area is restoring, implying
also an improvement in soil fertility (Fig. 3), which favors
biomass production and thus litter production as well. The
increase in soil fertility is brought about by improved nutrient
cycling and by trapping of fertile sediments by the restoring
vegetation (Descheemaeker et al., 2006). Lugo (1992) and
Dames et al. (1998) noted as well that litterfall rates increase
with age.
The contribution of different litter components to total litter
production varies according to species composition. However,
for dry tropical forests leaf litter contributes typically around
70% of the total (Martı́nez-Yrı́zar, 1995). Lisanework and
Michelsen (1994) report leaf litter contributions between 57
and 77% of total litter production in forests in the Ethiopian
highlands. In this respect, the share of leaf litter in exclosures
and church forest in the Tigray highlands is somewhat high, as
it ranges from 68 to 86% of the total (Table 3). Notwithstanding
the possible underestimation of woody litter production due to
trap size, our results are conform to what Martı́nez-Yrı́zar
(1995) reports for tropical dry forests: woody litter is second in
importance followed by reproductive litter.
Considerable seasonality in litterfall was demonstrated for
all restoring forest areas under study (Figs. 4 and 5). Total and
leaf litterfall are characterized by low rates in the wet season
and high rates in the dry season. This seasonality is common to
all dry forests because of pronounced seasonal changes in
rainfall (Martı́nez-Yrı́zar, 1995; Xuluc-Tolosa et al., 2003).
Also other authors (e.g. Lisanework and Michelsen, 1994;
Martı́nez-Yrı́zar et al., 1999; Sundarapandian and Swamy,
1999) reported strong seasonal trends related to rainfall and
moisture stress. Leaf abscission takes place as a result of water
stress (Swift and Anderson, 1989) and because of the inability
of the plant to supply water to leaves and to prevent dessication
and damage to plant structures (Holbrook et al., 1995). In an
environment characterized by a clear-cut dry season with severe
water stress conditions, a vegetation dominated by drought-
deciduous species is expected (Borchert, 1994). Evergreen
species are more drought tolerant and higher investments in
long-lived leathery leaves are paid back by the ability of the
plant to respond to coincidental opportunities, like favorable
moisture conditions. In this respect it is logical that in the
church forest of May Ba’ati the peak in litterfall is not observed
immediately following the rainy season like in the other areas,
but from January to April (Fig. 5). The more favorable moisture
condition in this place due to the forest microclimate certainly
plays a role. This explanation is in line with Borchert (1994)
stating that site dependent differences in soil water availability
are principal environmental causes of variation in tree
phenology. Besides that, there is also a difference in life form
of the dominant species. The exclosures are dominated by
shrubs and small trees, while in the church forest, mature trees
are present, whose roots reach deeper lying water resources.
Moreover, the climax species constituting the church forest
vegetation typically have sclerophyllous leaves with thick
cuticula, which are better adapted to withstand water stress
(Turner, 1994). These advantages concerning water supply and
conservation are a probable explanation for the delay in leaf-
fall recorded in the church forest. The natural forest in the
central highlands of Ethiopia, described by Lisanework and
Michelsen (1994) shows a peak in litter production in spring
and lowest production in the wet summer months, which is
comparable to the pattern displayed by the church forest of May
Ba’ati.
Patterns in reproductive litter production depend largely on
flowering and fruiting of the dominant species (Martı́nez-
Yrı́zar et al., 1999). In the old exclosure of May Ba’ati, this is
illustrated by the presence of several small peaks (Fig. 5), for
which different species are responsible. Also the fact that in the
old exclosure of Kunale fruit shedding takes place until March,
while in the young exclosure this already ends in January
(Fig. 4) is related to the presence and absence of certain species.
In this case A. saligna, absent in the young and co-dominant in
the old exclosure, is responsible.
In a similar way the detailed study of nine species (Fig. 6),
which are dominant in the studied areas, serves to better
understand the temporal and spatial patterns of litter production
in the different sites. Four of the studied species (B.
grandiflorum, S. schimperi, D. angustifolia and D. torrida)
are drought-deciduous and show a peak in leaf litter production
after the rainy season (Fig. 6). This trend in leaf litter
production was also recorded by Martı́nez-Yrı́zar et al. (1999)
 K. Descheemaeker et al. / Forest Ecology and Management 233 (2006) 21–35 33

for drought-deciduous species in Sonoran desert communities Table 8

Leaf litter quality descriptors for the studied land use types and selected species
in Mexico and also Borchert (1994) demonstrated that drought
sensitive species shed their leaves abruptly during the dry Land use typea C/N C/P C/K SLMb (g m
2)
season. Real evergreen shrubs like C. edulis and E. racemosa May Ba’ati
have a more continuous leaf turnover pattern with a peak of RA 43 2736 262 146
leaf-fall later on in spring (Fig. 6), before the main rainy season, YE 46 1929 136 139
which is also true for the evergreen shrubs studied by Martı́nez- OE 47 2215 212 149
CF 51 1826 152 172
Yrı́zar et al. (1999). Hernandez et al. (1992) describe the
seasonal pattern of leaf litter production of the evergreen Kunale
Quercus rotundifolia, and state that shedding of leaves RA 43 1225 258 67
YE 42 1215 118 123
coincides with the appearance of new leaves, which applies ME 40 1570 131 138
to C. edulis and E. racemosa as well. The abundance of the OE 45 2027 153 172
evergreen E. racemosa in the May Ba’ati exclosures can be
explained by the high drought tolerance of this species. Species C/N C/P C/K SLMb (g m
2)
Whether this is achieved through a deep rooting system, Acacia abyssinica 32 1227 252 82
morphological characteristics of the leaves or physiological Acacia etbaica 26 2366 280 117
adaptations is not sure. On the other hand, the three Acacia Acacia saligna 33 3198 216 263
Becium grandiflorum 51 2153 32 170
species that were studied show a distinct temporal litterfall
Carissa edulils 52 1840 68 157
pattern (Fig. 6) with a pronounced peak in leaf litter production Dodonea angustifolia 32 2124 340 122
in January for A. saligna and in March for A. etbaica and A. Dombeya torrida 34 7220 26 79
abyssinica. The fact that the indigenous Acacia species shed Euclea racemosa 53 3473 302 184
their leaves later on in the dry season may be attributed to their Salvia schimperi 25 622 35 175
well-developed dehydration tolerance capacity (Kindeya a
Land use type coding, RA: grazing land; YE: young exclosure; ME: middle-
Gebrehiwot et al., 2005), which is typical for evergreen aged exclosure; OE: old exclosure; CF: church forest.
b
species as compared to drought-deciduous species. SLM: specific leaf mass.
Standing crop of litter was determined in September, before
the period of maximal litter production and just after the rainy
season, with maximal decomposition rates. As such, the records composition is altered in the course of time (Table 4). Eearly
are minimum values and should not be interpreted as average successional forest stages are characterized by pioneer species
values over the year. Standing crops of litter for the old with high litter nutrient content, such as B. grandiflorum and S.
exclosures of both study sites and the church forest (Table 6) are schimperi (Table 8). These species show a relatively low litter
within the range (320–1230 g m
2) reported by Martı́nez- production rate (Table 5) and limited litter accumulation
Yrı́zar (1995) for dry tropical forest in similar rainfall regimes, (Table 7). More mature forest areas are richer in climax species,
while for the younger exclosures standing crops of litter are with commonly sclerophyllous leaves, which are known to
lower. The standing crop of litter in the old exclosures is also decompose more slowly, such as Euclea racemosa and Carissa
comparable to the value of 300 g m
2 reported by Swift et al. edulis. These species are characterized by higher production
(1979) for savannah ecosystems. From Fig. 7 it is clear that the rates (Table 5) and more pronounced litter accumulation
driving force behind litter accumulation is increased litter (Table 7). The species composition is reflected in the average
production. Also for plantations and secondary forests in a litter nutrient composition and the average specific leaf mass
subtropical wet climate, Lugo (1992) reported that litter that was calculated per land use type (Table 8). Generally
accumulation increases with age. The high importance of litter speaking, vegetation restoration goes hand in hand with an
input for explaining litter accumulation (Fig. 7) in restoring increase in C/N ratio and specific leaf mass in both sites
forest areas in Tigray, points to an overall slow decomposition. (Table 8). The effect of litter quality on decomposition rates is
In an ecosystem with fast decomposition on the contrary extensively discussed in the literature and C/N and lignin/N
(typical eumull humus form) all litter would decompose ratios are commonly accepted as main explanatory factors (e.g.
quickly, not allowing litter accumulation, no matter how high Melillo et al., 1982; Berg, 2000;). As specific leaf mass is
the input. The slow decomposition revealed in this study is positively related to lignin and negatively to nitrogen content
explained on the one hand by the annual dry period, during (Medina et al., 1990), this variable serves as an indication for
which decomposition is almost impeded and on the other hand litter accumulation as well. Xuluc-Tolosa et al. (2003) also
by the relatively high litter C/N ratios, which are above 40 in all report on the effects of changing vegetation composition on
land use types (Table 8). Leaf litter with C/N ratio lower than 30 decomposition rates in dry secondary forests in Mexico.
is known to decompose easily and yield a mull humus type, Another effect of vegetation restoration is an improved
whereas C/N ratios above 30 already result in N immobilization microclimate and microbial environment (Xuluc-Tolosa et al.,
(Heal et al., 1997) and decomposition retardation (Wittich, 2003). In semi-arid areas such as Tigray, this means improved
1961). moisture availability (Descheemaeker et al., in press), which
Several processes may play a role in determining litter can promote decomposition. Microclimate development thus
accumulation in a regenerating exclosure. First of all, species counters the general effect of changing species composition.
34 K. Descheemaeker et al. / Forest Ecology and Management 233 (2006) 21–35
Generalizing our results, the evolution in litter production
and accumulation in restoring forest areas can be sketched as
follows. When a degraded area is closed for rehabilitation,
vegetation density, total leaf area and soil fertility increase,
which results in an exponentially increasing litter production
(Figs. 2 and 3). Early successional stages are not yet
characterized by an improved microclimate, so that the annual
dry periods impede fast decomposition, with litter accumula-
tion as a result (Fig. 7). Another evolution favouring litter
accumulation is the altering vegetation composition (Table 4)
with low-quality litter species in more mature forest areas. This
trend is somewhat countered by improved moisture availability
and better microclimate, resulting from canopy closure. In the
church forest, representing the near-to-climax vegetation, not
only microclimate improvement but also a better closed
nutrient cycling is noticed, witnessed by decreased C/P (1826)
and C/K (152) ratios as compared to the old exclosure (Table 8).
The result of the exponentially increasing litter input and the
overall slow decomposition rate is the accumulation of soil
organic matter as forest restoration areas get older.
Under species such as E. racemosa and A. saligna,
characterized by high standing crops of litter, typical Oligomull
or Moder humus forms are found (Descheemaeker et al.,
submitted for publication). These are humus forms with slow
nutrient mineralization and strong accumulation of organic
matter. On the other hand, under species with rapid
decomposition, such as B. grandiflorum and D. torrida, typical
Mull humus forms are present, with negligible organic matter
accumulation (Descheemaeker et al., submitted for publica-
tion). Nutrients contained in the plant residues are mineralized
at a very rapid rate here. Besides knowledge on growth rates,
these results provide valuable information when it comes to
species selection for enrichment planting. When microclimate
enhancement and soil protection is aimed at, species with slow
decomposition rates and high organic matter accumulation are
a valuable option. However, when it comes to biomass
production, a favorable plant nutrient availability brought
about by rapid mineralization can be preferable.
5. Conclusions
Annual litter production in exclosures in the Tigray
highlands is relatively low (30–316 g m
2 year
1) compared
to productions reached in dry tropical forests elsewhere,
because these restoring forest areas originated in very degraded
areas with low soil fertility and are still in an early successional
stage. Litter production significantly rises after exclosure
establishment, which has important beneficial effects with
respect to soil moisture conservation and nutrient return to the
soil. Litterfall in the studied areas is characterized by strong
seasonality, which is typical for forests in dry climates and
attributed to rainfall patterns and species composition.
Litter accumulation in restoring forest areas is mainly
explained by litter input. However, in the successional stages of
forest restoration, litter quality, mediated through changing
species composition, and microclimate development also
determine organic matter accumulation.
Litter production for a set of nine dominant species reveals
three types of litter production patterns. Drought-deciduous
species shed their leaves immediately after the rainy season,
while evergreen species have a continuous turnover with a peak
before the rainy season. Acacia species on the other hand have a
pronounced peak of leaf-fall in the middle of the dry season.
These phenological types can be related to species-specific
drought stress adaptation and tolerance and site-specific
moisture availability. Litter accumulation under the selected
species is quite diverse and can be partly explained by litter
quality.
Acknowledgements
The research was conducted within the framework of two
VLIR (Flemish Interuniversity Council) funded projects (Zala
Daget project and Forest Rehabilitation project) at Mekelle
University. Katrien Descheemaeker is supported through an
FWO-Vlaanderen Aspirant scholarship. The support from
Mekelle University and the Faculty of Bioscience Engineering,
K.U. Leuven, is greatly appreciated. Furthermore, we want to
thank the local Bureau of Agriculture branch, the authorities
and exclosure guards of several villages of the Dogu’a Tembien
district and Haftu Assefa for his field assistance.
References
Aerts, R., Wagendorp, T., November, E., Mintesinot, B., Deckers, J., Muys, B.,
2004. Ecosystem thermal buffer capacity as an indicator of the restoration
status of protected areas in the Northern Ethiopian Highlands. Restor. Ecol.
12, 586–596.
Anderson, J.M., Ingram, J.S., 1993. Tropical Soil Biology and Fertility. A
Handbook of Methods. Commonwealth Agricultural Bureau, Oxon, UK.
Berg, B., 2000. Litter decomposition and organic matter turnover in northern
forest soils. Forest Ecol. Manage. 133, 13–22.
Borchert, R., 1994. Soil and stem water storage determine phenology and
distribution of tropical dry forest trees. Ecology 75, 1437–1449.
Bosellini, A., Russo, A., Fantozzi, P.L., Assefa, G., Solomon, T., 1997. The
Mesozoic succession of the Mekele outlier (Tigre Province, Ethiopia).
Mem. Sci. Geol. 49, 95–116.
Brown, S., Lugo, A.E., 1990. Tropical secondary forests. J. Trop. Ecol. 6, 1–32.
Búrquez, A., Martı́nez-Yrı́zar, A., Núñez, S., 1999. Sonoran desert productivity
and the effect of trap size on litterfall estimates in dryland vegetation. J. Arid
Environ. 43, 459–465.
Cintrón, B.B., Lugo, A.E., 1990. Litterfall in a subtropical dry forest: Guánica,
Puerto Rico. Acta Cientı́fica (San Juan) 4, 37–49.
Dames, J.F., Scholes, M.C., Straker, C.J., 1998. Litter production and accu-
mulation in Pinus patula plantations of the Mpumalanga Province, South
Africa. Plant Soil 203, 183–190.
Descheemaeker, K., Nyssen, J., Rossi, J., Poesen, J., Mitiku, H., Raes, D., Muys,
B., Moeyersons, J., Deckers, J., 2006. Sediment deposition and pedogenesis
in exclosures in the Tigray Highlands, Ethiopia. Geoderma 132, 291–314.
Descheemaeker, K., Nyssen, J., Poesen, J., Mitiku, H., Muys, B., Raes, D.,
Moeyersons, J., Deckers, J. Soil and water conservation through forest
restoration in exclosures of the Tigray highlands. J. Drylands, in press.
Descheemaeker, K., Muys, B., Nyssen, J., Sauwens, W., Mitiku H., Poesen, J.,
Raes, D., Deckers, S. Humus forms of exclosures in the highlands of Tigray,
Northern Ethiopia, Geoderma, submitted for publication.
Eweg, H.P.A., Van Lammeren, R., Deurloo, H., Woldu, Z., 1998. Analysing
degradation and rehabilitation for sustainable land management in the
highlands of Ethiopia. Land Degrad. Dev. 9, 529–542.
FAO, ISRIC, ISSS, 1998. World Reference Base of Soil Resources. World Soil
Resources Reports 84. FAO, Rome, Italy.
 K. Descheemaeker et al. / Forest Ecology and Management 233 (2006) 21–35
Gao, Y., Qiu, G.Y., Shimizu, H., Tobe, T., Sun, B., Wang, J., 2002. A 10-year
study on techniques for vegetation restoration in a desertified salt lake area.
J. Arid Environ. 52, 483–497.
Heal, O.W., Anderson, J.M., Swift, M.J., 1997. Plant litter quality and decom-
position: an historical overview. In: Cadish, G., Giller, K.E. (Eds.), Driven
by Nature. Plant Litter Quality and Decomposition. CAB International,
Wallingford, pp. 3–30.
Hernandez, I.M., Gallardo, J.F., Santa Regina, I., 1992. Dynamic of organic
matter in forests subject to a Mediterranean semi-arid climate in the Duero
basin (Spain): litter production. Acta Oecol. 13, 55–65.
Herweg, K., Stillhardt, B., 1999. Implications of land use and land cover
dynamics for mountain resource degradation in the Northwestern Ethiopian
Highlands. The variability of soil erosion in the highlands of Ethiopia and
Eritrea. Average and extreme erosion patterns. University of Berne, Switzer-
land: Soil Conservation Research Programme, Research Report 42.
Holbrook, N.M., Whitbeck, J.L., Mooney, H.A., 1995. Drought responses of
neotropical dry forest trees. In: Bullock, S.H., Mooney, H.A., Medina, E.
(Eds.), Seasonally Dry Tropical Forests. Cambridge University Press,
Cambridge, pp. 243–276.
Hongo, A., Matsumoto, S., Takahashi, H., Zou, H.Y., Cheng, J.M., Jia, H.Y.,
Zhao, H.Y., 1995. Effect of exclosure and topography on rehabilitation of
overgrazed shrub-steppe in the loess plateau of Northwest China. Restor.
Ecol. 3, 18–25.
Hurni, H., 1990. Degradation and conservation of the resources in the Ethiopian
highlands. Mountain Res. Dev. 8, 123–130.
Kebrom, T., Backeus, I., Skoglund, J., Zerihun, W., 1997. Vegetation on
hillslopes in southern Wello, Ethiopia: degradation and regeneration. Nord.
J. Bot. 17, 483–493.
Kindeya Gebrehiwot, Muys, B., Mitiku Haile, Mitloehner, R., 2005. The use of
plant water relations to characterize tree species and sites in the drylands of
northern Ethiopia. J. Arid Environ. 60, 581–592.
Lavelle, P., Blanchart, E., Martin, A., Martin, S., 1993. A hierarchical model for
decomposition in terrestrial ecosystems: application to soils of the humid
tropics. Biotropica 25, 130–150.
Le Houérou, H.N., 2000. Restoration and rehabilitation of arid and semiarid
Mediterranean ecosystems in North Africa and West Asia: a review. Arid
Soil Res. Rehabil. 14, 3–14.
Lisanework, N., Michelsen, A., 1994. Litterfall and nutrient release by decom-
position in three plantations compared with a natural forest in the Ethiopian
highland. Forest Ecol. Manage. 65, 149–164.
López-Bermúdez, F., Romero-Dı́az, A., Martı́nez-Fernandez, J., Martı́nez-Fer-
nandez, J., 1998. Vegetation and soil erosion under a semi-arid Mediterranean
climate: a case study from Murcia (Spain). Geomorphology 24, 51–58.
Lugo, A.E., 1992. Comparison of tropical tree plantations with secondary
forests of similar age. Ecol. Monogr. 62, 1–41.
Martı́nez-Yrı́zar, A., 1995. Biomass distribution and primary productivity of
tropical dry forests. In: Bullock, S.H., Mooney, H.A., Medina, E. (Eds.),
Seasonally Dry Tropical Forests. Cambridge University Press, Cambridge,
pp. 326–345.
Martı́nez-Yrı́zar, A., Núñez, S., Miranda, H., Búrquez, A., 1999. Temporal and
spatial variation of litter production in Sonoran Desert communities. Plant
Ecol. 145, 37–48.
McGill, W.B., Figueiredo, C.Y., 1993. Total Nitrogen. In: Carter, M.R. (Ed.),
Soil Sampling and Methods of Analysis. Canadian Society of Soil Science,
Lewis Publishers, Boca Raton, USA, pp. 201–211.
Medina, E., Garcia, V., Cuevas, E., 1990. Sclerophylly and oligotrophic
environments: relationships between leaf structure, mineral nutrient con-
35
tent, and drought resistance in tropical rain forests of the upper Rio Negro
region. Biotropica 22, 51–64.
Melillo, J.M., Aber, J.D., Muratore, J.F., 1982. Nitrogen and lignin control of
hardwood leaf litter decomposition dynamics. Ecology 63, 621–626.
Menaut, J.C., Lepage, M., Abbadie, L., 1995. Savannas, woodlands and dry
forests in Africa. In: Bullock, S.H., Mooney, H.A., Medina, E. (Eds.),
Seasonally Dry Tropical Forests. Cambridge University Press,
Cambridge, pp. 64–92.
Nyssen, J., Poesen, J., Moeyersons, J., Deckers, J., Mitiku Haile, Lang, A.,
2004. Human impact on the environment in the Ethiopian and Eritrean
highlands—a state of the art. Earth-Sci. Rev. 64, 273–320.
Nyssen, J., Vandenreyken, H., Poesen, J., Moeyersons, J., Deckers, J., Mitiku
Haile, Salles, C., Govers, G., 2005. Rainfall erosivity and variability in the
Northern Ethiopian Highlands. J. Hydrol. 311, 172–187.
Olson, S.R., Sommers, L.E., 1982. Phosphorus. In: Page, A.L., Miller, R.H.,
Keeney, D.R. (Eds.), Methods of Soil Analysis. Part 2, second ed. American
Society of Agronomy, Madison, WI (Agronomy No. 9).
Proctor, J., 1983. Tropical forest litterfall I. Problems of data comparison. In:
Sutton, S.L., Chadwick, T.C., Whitmore, T.C. (Eds.), Tropical Rain Forest:
Ecology and Management. Blackwell Scientific Publications, Oxford, pp.
267–273.
Songwe, N.C., Okali, D.U.U., Fasehun, F.E., 1995. Litter decomposition and
nutrient release in a tropical rainforest, Southern Bakundu Forest Reserve,
Cameroon. J. Trop. Ecol. 1, 333–350.
SPSS 11.0.1, 2001. SPSS for windows, release 11.0.1. SPSS Inc.
Sundarapandian, S.M., Swamy, P.S., 1999. Litter production and leaf-litter
decomposition of selected tree species in tropical forests at Kodayar in the
Western Ghats, India. Forest Ecol. Manage. 123, 231–244.
Swift, M.J., Heal, O.W., Anderson, J.M., 1979. Decomposition in Terrestrial
Ecosystems. Studies in Ecology, vol. 5. Blackwell Scientific Publications,
372 pp.
Swift, M.J., Anderson, J.M., 1989. Decomposition. In: Lietsh, H., Werger,
M.S.A. (Eds.), Tropical Rain Forest Ecosystems. Ecosystems of the World
14B. Elsevier, Amsterdam, pp. 547–569.
Tenna S., Sintayehu M., Berihun A., 2001. Stengthening user-rights over local
resources in Wollo, Ethiopia. IIED Drylands Programme, Issue paper no.
103, 23 pp.
Turner, I.M., 1994. Sclerophylly: primary protective? Funct. Ecol. 8, 185–206.
Vitousek, P.M., 1984. Litterfall, nutrient cycling, and nutrient limitation in
tropical forests. Ecology 65, 285–298.
Van Reeuwijk, L.P., 2002. Procedures for soil analysis, sixth ed. Technical
Paper 9, ISRIC, Wageningen, The Netherlands.
Wisborg, P., Shylendra, H.S., Gebrehiwot, K., Shanker, R., Tilahun, Y.,
Nagothu, U.S., Tewoldeberhan, S., Bose, P., 2000. Rehabilitation of CPRS
through re-crafting of village institutions: a comparative study from Ethio-
pia and India. In: Paper Presented at the Eighth Biennial Conference of the
International Association for the Study of Common Property (IASCP).
Bloomington, Indiana, May 31–June 4 26 pp.
Wittich, W., 1961. Einfluss der Baumart auf den Bodenzustand. Allgemeine
Forstzeitschrift 16, 41–45.
Xuluc-Tolosa, F.J., Vester, H.F.M., Ramirez-Marcial, N., Castellanos-Albores,
J., Lawrence, D., 2003. Leaf litter decomposition of tree species in three
successional phases of tropical dry secondary forest in Campeche, Mexico.
Forest Ecol. Manage. 174, 401–412.
Zhang, B., Yang, Y., Zepp, H., 2004. Effect of vegetation restoration on soil and
water erosion and nutrient losses of a severely eroded clayey Plinthudult in
southeastern China. Catena 57, 77–90.