www.elsevier.com/locate/geoderma

Geoderma 132 (2

Sediment deposition and pedogenesis in exclosures in the

Tigray highlands, Ethiopia

Katrien Descheemaeker a,*, Jan Nyssen a,b, Joni Rossi a, Jean Poesen c, Mitiku Haile b,

Dirk Raes a, Bart Muys d, Jan Moeyersons e, Seppe Deckers a

aLaboratory for Soil and Water Management, K.U.Leuven, Vital Decosterstraat 102, B-3000 Leuven, BelgiumbDepartment of Land Resource Management and Environmental Protection, Mekelle University, PO Box 231, Mekelle, Ethiopia

cPhysical and Regional Geography Research Group, K.U.Leuven, Redingenstraat 16, B-3000 Leuven, BelgiumdLaboratory for Forest, Nature and Landscape, Vital Decosterstraat 102, B-3000 Leuven, Belgium

eRoyal Museum for Central Africa, B-1960 Tervuren, Belgium

Received 5 July 2004; received in revised form 1 April 2005; accepted 27 April 2005

Available online 6 July 2005

Abstract

In the Tigray highlands of Northern Ethiopia, the establishment of exclosures (i.e., areas closed for grazing and

agriculture) has become an important measure to combat land degradation and restore vegetative cover. Exclosures are

commonly found on steep slopes and downslope from a sediment source area. In this study their sediment trapping

capacity and controlling factors were investigated. Total sediment depth turned out to be related to vegetation cover,

sediment source area and in some cases slope gradient. Thickness of recent, short-term (b20 years) sediment deposits

was strongly related to distance from the top edge of the closed area, slope gradient, vegetation cover and characteristics

of the sediment source area. Mean sediment deposition rates ranged between 26 and 123 Mg ha�1 yr�1. Under

influence of vegetation and sediment deposition dark soils rich in organic matter (Phaeozems) develop. In view of their

high sediment trapping capacity, exclosures are highly valued as efficient soil conservation measures in the Tigray

highlands.

D 2005 Elsevier B.V. All rights reserved.

Keywords: Closed areas; Sediment trapping; Vegetative sediment buffer; Phaeozems; Forest restoration; Eucalyptus; East Africa

0016-7061/$ - see front matter D 2005 Elsevier B.V. All rights reserved.

doi:10.1016/j.geoderma.2005.04.027

* Corresponding author. Tel.: +32 16329765; fax: +32 16329760.

E-mail address: katrien.descheemaeker@agr.kuleuven.ac.be

(K. Descheemaeker).

1. Introduction

Land degradation, comprising degradation of the

natural vegetation cover, soil erosion, loss of soil

fertility and moisture stress is a well-known problem

in the Ethiopian highlands (Herweg and Stillhardt,

1999). Deforestation, overgrazing, expansion of crop-

006) 291–314

K. Descheemaeker et al. / Geoderma 132 (2006) 291–314292

land and unsustainable use of natural resources are the

main causes for the alarming state of land degradation

in the highlands of the northern province Tigray

(Nyssen et al., 2004).

In an attempt to rehabilitate the natural environ-

ment, the Northern Ethiopian society reacted by taking

soil and water conservation measures. Initially, mainly

physical structures were introduced (Chadhokar and

Solomon, 1988), such as the now widespread stone

bunds constructed along contour lines. However, in the

1980s it was realized that vegetation restoration is very

important and often less costly (Fiedler and Belay,

1988). This idea was put into practice by closing the

most degraded areas for grazing and agriculture

through the joint initiative of local communities, Gov-

ernmental and Non-Governmental Organizations.

Later on, this measure was adopted in the policy

framework of the Tigray Forestry Action Plan

(TFAP, 1996). However, the establishment of exclo-

sures should not be seen as an intervention merely

focusing on environmental rehabilitation. Income gen-

eration and community relief through food and cash

for work were also aimed at.

The terms bexclosureQ and bclosed areaQ denom-

inate the areas set aside, where agriculture and

grazing became forbidden (Kebrom et al., 1997;

Wisborg et al., 2000; Tenna et al., 2001; Aerts et

al., 2001, 2003). Exclosures are defined by Le

Houerou (2000) as areas of natural vegetation pro-

tected from the intrusion of major degradation

agents, such as humans, livestock and big game.

In Tigray, exclosures are demarcated on community

land and mostly found on very steep and degraded

slopes. Management and control of the area by a

guard lies within the responsibility of the local

authority. The objectives of exclosure are rehabilita-

tion of degraded land, production of grass for fodder

and thatching, wood for fuel and construction and

non-wood forest products.

Eweg et al. (1998) noted that the recovery process

in exclosures invariably starts with a rapid increase in

richness and soil cover of the herbaceous species.

After 3 to 5 years shrub and tree species gain impor-

tance and suppress the herbaceous species.

According to different studies (Chadhokar, 1992;

Mitiku and Kindeya, 1998; Wisborg et al., 2000;

Aerts et al., 2003, 2004) several positive effects of

exclosures can be observed. Natural vegetation is

regenerating, runoff and sheet and rill erosion are

reduced, the land is stabilised, soil moisture availabil-

ity is increased and a microclimate is reinstalled.

Moreover, new springs originate and during a longer

period more water is flowing in the rivers draining the

catchments. Exclosures thus seem to be a very suc-

cessful measure improving soil and water conserva-

tion. But, at the same time they cause more pressure

on remaining grazing land.

In view of the high land pressure and scarcity of

fuel wood and grazing land, it is important to quantify

the benefits and trade-offs of exclosures and to obtain

a profound knowledge about the effects of the set-

aside policy on different aspects of the natural envi-

ronment. One important process requiring further in-

vestigation is sediment trapping and soil development

in exclosures. From field observations it is clear that

under certain conditions exclosures can trap a consid-

erable amount of sediment. As such, they could con-

tribute to the buildup of fertile soil in formerly

degraded lands and act as buffers reducing sediment

transport in the landscape.

The process of sediment deposition by vegeta-

tion buffers has been studied extensively. Deposi-

tion is the result of a modification of the hydraulic

characteristics of the water flow, such as a reduc-

tion in flow turbulence and velocity (Abu-Zreig,

2001). Through this, infiltration is promoted so

that the sediment load is deposited due to a low-

ered transport capacity of the flow (Lacey, 2000).

In places where all water infiltrates, all sediment is

deposited. Abu-Zreig (2001) gives an overview of

the efforts undertaken to experimentally characterize

the performance of vegetative buffer strips with

different dimensions and characteristics. The most

important factors influencing the sediment trapping

capacity of these buffers seem to be length, inflow

sediment type, slope gradient, soil type and surface

roughness, which is related to vegetation type and

cover. Studies on the effectiveness of forest buffers

in reducing overland flow and its sediment concen-

tration reveal that sediment trapping capacity

depends on vegetation type, surface roughness, to-

pographic variables like slope gradient, length and

shape, soil characteristics like water storage capac-

ity and saturated hydraulic conductivity, and over-

land flow velocity (Croke et al., 1999; Lacey,

2000).

K. Descheemaeker et al. / Geoderma 132 (2006) 291–314 293

In many soil erosion studies however, sediment

trapping is neglected, so that erosion models often

overestimate catchment sediment yields (Rey, 2003).

On a broader landscape scale, the transfer of water and

sediment from disturbed to less disturbed parts of the

landscape is a very important aspect in catchment

impact assessment, but there is still scarcity of data

on the nature of hillslope sediment storage and the

rates of sediment transfer and delivery (Croke et al.,

1999). The importance of redeposition in sediment

yield studies is also endorsed by the phenomenon of

dramatic decrease in sediment delivery ratios when

zooming out from small to larger areas in the land-

scape (Bergkamp, 1998; Braud et al., 2001).

In the literature on sediment deposition only few

studies are available, presenting estimates of

amounts of sediment deposited in field situations.

Notwithstanding Bojo and Cassells (1995) noting

that for the specific case of Ethiopia no empirical

data are available on sediment deposition patterns in

relation to geomorphic parameters, there are some

indications. Hurni (1985) shows, for a 116 ha catch-

ment in Wello, that the rate of sediment redeposition

(17 Mg ha�1 y�1) is more important than the rate

of sediment export through the drainage system (7

Mg ha�1 y�1). In a vegetation-rich catchment in

South-western Ethiopia, redeposition of eroded sed-

iment occurs at a rate of 30 Mg ha�1 y�1, while

sediment export rates through the river are only 1.1

Mg ha�1 y�1. However, also here no details are

provided on specific locations and conditions for

sediment deposition.

The overall objective of this study is to better

understand sediment deposition processes and soil

formation in exclosures with the aim to contribute

to a better land management system in the Tigray

highlands. More specific objectives of this study

are (1) to assess pedogenesis in exclosures, (2) to

characterise different types of deposited sediment

according to their spatial distribution, morphological

and chemical characteristics, age and timescale of

deposition, (3) to develop a field method for deter-

mining thickness of deposited sediment, (4) to

define the role of vegetation cover, topography

and sediment source area characteristics in the de-

position of sediment and its spatial distribution and

(5) to assess the sediment deposition rate within

exclosures.

2. Methodology

To study changes in sediment deposition rates or

conservation of fertile top soil induced by land use

dynamics, one could monitor different land use types

for a long time, and record what happens. This very

costly and time-consuming method can however be

substituted by a more simple method, in which top

soils within different land use types are compared.

The prerequisite for this method is that for the differ-

ent treatments (i.e., land use types) differences in

lithological, topographic and climatic conditions are

negligible, so that space can replace time (Woldeam-

lak and Stroosnijder, 2003). This inferential approach

has been used widely in studies on the effects of land

use changes on soil properties (Adejuwon and Eka-

nade, 1988; Abubakar, 1996, 1997; Hajabbasi et al.,

1997; Lumbanraja et al., 1998; Jaiyeoba, 1998; Pania-

gua et al., 1999; Liu et al., 2002; Woldeamlak and

Stroosnijder, 2003).

In this study, the effect of the exclosure land use

type on pedogenesis and sediment trapping was

assessed by comparing soils in typical exclosure

areas with soils in comparable landscape situations

but under other land use types.

2.1. Study area

The study area (Fig. 1) is located in the Dogu’a

Tembien (Tembien highlands) district in central

Tigray, near the district capital Hagere Selam (13839V N, 398 10V E, altitude 2650 m a.s.l.). This area

is representative for the northern Ethiopian highlands,

because of its elevation and morphology. The lithol-

ogy of the study area is composed of Mesozoic sed-

imentary rocks (Adigrat sandstone, Antalo limestone

and Amba Aradam sandstone), covered by Tertiary

flood basalts (Bosellini et al., 1997). The subhorizon-

tal layering of rock formations of different resistance

to weathering results in a stepped landscape, where

flats alternate with steep escarpments (Nyssen et al.,

2002). Mean annual rainfall is around 700 mm, most-

ly concentrated in the rainy season from June to

September.

According to the ratio of annual precipitation to

potential evapotranspiration, UNEP (1994) qualifies

land degradation in the study area as desertification.

Therefore, soil and water conservation merits high

Fig. 1. Location of the study area in Ethiopia (black circle on country map), and overview of study sites and selected parcels with different land uses. Bold lines in parcels indicate

studied transects.

K.Desch

eemaeker

etal./Geoderm

a132(2006)291–314

294

K. Descheemaeker et al. / Geoderma 132 (2006) 291–314 295

priority attention (Nyssen et al., 2000, 2004). In the

1980s a start was made with closing degraded areas.

According to the local Agricultural Office exclosures

now cover about 12% of the total area in the study

district. They mostly form 50 to 150 m wide and 200

to 2000 m long vegetation strips on steep slopes

parallel to the contour lines.

During a preliminary survey an inventory of 43

exclosure sites was made. These sites were located on

a topographic map and described according to lithol-

ogy, soil type, vegetation type, slope gradient and

aspect, management and year of closure. Based on

the information collected, three representative study

sites (Fig. 1, Table 1) were selected, each of them

located on one of the dominant lithological units of

the study area: i.e., Haragua (sandstone), May Ba’ati

(limestone) and Adewro (basalt). Within the study

sites, several parcels with different land use were

selected in similar landscape positions, to enable com-

parative analysis. The land use types investigated are

exclosures of different ages, degraded rangelands, a

Eucalyptus forest and church forests. The latter are

sacred places where the vegetation has always been

protected. Church forests are often located in small

valleys with more favorable moisture conditions.

However, for this study we selected church forests

located in landscape situations comparable to the ones

of the other parcels. A brief description of all study

sites and parcels is given in Table 1.

Land use dynamics on steep slopes, which are now

under exclosure, went through a number of typical

evolution stages, illustrated in Fig. 2: (1) removal of

the original forest vegetation due to population pres-

sure, (2) severe soil erosion and land degradation

under overgrazing, (3) vegetation regeneration and

sediment trapping after exclosure establishment. In

all study sites, the selected parcels represent the dif-

ferent stages in this land use evolution.

All parcels are located on steep slopes with gradi-

ents of 35% to 70% and widths of 100 to 150 m.

Along their upper edge they are bordered by flatter

(slope gradients generally up to 20%), usually

cropped land, where runoff and soil erosion take

place. The upslope land is the sediment source area.

It produces important volumes of sediment carried by

runoff water to the study areas. This is illustrated in

Fig. 3, representing the May Ba’ati site with the

studied parcels and their sediment source areas.

2.2. Soil study

Soil profiles were described in 15 representative

profile pits in the different parcels, using the Guide-

lines for Soil Description (FAO, 1990). Classification

was carried out following the World Reference Base

for Soil Resources (WRB) (FAO et al., 1998). About

300 soil augerings, for which a selection of variables

(depth, texture, colour, structure, consistence, pres-

ence of CaCO3, stoniness) was recorded, complemen-

ted the detailed profile study.

A distinction was made between breference soil

horizonQ and bsediment layerQ. The latter one consist

of soil material deposited on top of the original soil

profile and conserved under the influence of vegeta-

tion. The reference soil horizon refers to the original

soil profile. Based on profile and augering descrip-

tions and chemical analyses, typical characteristics of

both types of layers were determined for the three

study sites. Depths of deposited sediment were then

determined in the field by measuring thickness of soil

material showing characteristics of sediment layers

until a reference horizon was reached.

2.3. Transect study

A transect study was conducted to record depths of

deposited sediment in all land use types. Therefore,

three transects were laid out along the slope (bold

lines in Fig. 1) in each parcel of the study sites. The

following variables were recorded every 20 m along

these transects starting 10 m from the top edge:

2.3.1. Depth of sediment deposition (SED)

At each location, 5 augerings at one meter distance

from each other were made to determine the average

sediment depth. Soil colour, texture, stoniness and

indications on structure and consistence were

recorded for each stratum.

2.3.2. Slope gradient (SLO)

At each location, slope gradient was measured

using a clinometer.

2.3.3. Vegetation cover (VEG)

Soil cover by trees, shrubs, herbs and grasses was

recorded in the second half of the rainy season of the

year 2002 for each segment of the transect (0–20 m,

Table 1

Description of study sites and the selected parcels with different land uses

Site name Location Approximate

elevation (m a.s.l.)

Lithology Parcel

code

Land use

type

Parcel

area (ha)

Years of

closure

Slope

gradient (%)

Vegetation

cover (%)

Main shrub and

tree species

Haragua 138 40V 54W N,

398 09V 29W E

2100 Adigrat

sandstone

HFO Church forest 1.3 55 75 Calpurnia aurea(Ait.)

Benth. subsp. aurea

Rhus natalensis Krauss

Adathoda schimperana

Nees (Wilson)

Celtis africana Burm. f.

HXY Young exclosure 2.2 5 55 55 Dodonea angustifolia L.f.

Euclea schimperi (A.DC.) Dandy

Lantana viburnoides

(Forssk.) Vahl.

Rhus natalensis Krauss

HRA Rangeland 1.5 50 25 Aloe macrocarpa Tod.

May BaVati 138 38V 52W N,

398 13V 15W E

2300 Antalo

limestone

MFO Church forest 1.8 35 75 Acokanthera schimperi

(A. DC.) Benth.

Euclea schimperi (A.DC.) Dandy

Carissa edulis (Forssk) Vahl

Rhus glutinosa

Hochst. ex A. Rich

MXOa Old

exclosure A

3.0 20 55 50 Acacia etbaica Schweinf.

Dodonea angustifolia L.f.

Carissa edulis (Forssk) Vahl

Euclea schimperi (A.DC.) Dandy

MXOb Old exclosure B 6.7 20 45 50 Acacia etbaicaSchweinf.

Dodonea angustifolia L.f.

Carissa edulis (Forssk) Vahl

Euclea schimperi (A.DC.) Dandy

MXY Young exclosure 1.5 3 60 25 Becium grandiflorum

(Lam.) Pichi-Serm.

Acacia etbaicaSchweinf.

Euclea schimperi (A.DC.) Dandy

Adewro 138 37V 27W N,

398 09V 7W E

2700 Tertiary Ashange

basalt

AFO Church forest 3.3 65 60 Calpurnia aurea (Ait.)

Benth. sub. Aurea

Euphorbia candelabrumKotschy

Olea europaeasubsp. cuspidata

(Wall. Ex. DC) Ciffieri

AEU Eucalyptus

plantation

0.9 16 70 50 Eucalyptus globulus Labill.

Aloe macrocarpa Tod.

AXY Young exclosure 1.1 4 70 35 Rumex nervosus Vahl.

Aloe macrocarpa Tod.

ARA Rangeland 0.5 20 5 Aloe macrocarpa Tod.

Rumex nervosus Vahl.

K.Desch

eemaeker

etal./Geoderm

a132(2006)291–314

296

Fig. 2. Illustration of a typical succession of land use types found on steep slope sections in the Tigray highlands. Forest vegetation (a), degraded

rangeland (b), young (c) and older exclosure (d). Detailed view of the vegetation in a young (e) and an old exclosure (f). All pictures are from

the May Ba’ati study site, taken in the dry season.

K. Descheemaeker et al. / Geoderma 132 (2006) 291–314 297

20–40 m, 40–60 m, etc.) following the line transect

method (Jennings et al., 1999). For this, a 50 m long

tape was laid out from the top edge of the parcel to the

bottom, and the fraction of the tape covered by plant

Fig. 3. Aerial overview of the May Ba’ati site with the studied parcels: MFO

and MXY (young exclosure) and their respective sediment source areas (

canopies was recorded. The percentage of cover was

then the fraction of the tape that overlapped with

vegetation cover. Not only the ground level vegeta-

tion, but also the area below the canopy of trees and

(church forest), MXOa (old exclosure A), MXOb (old exclosure B)

FO, A, B1, B2, C).

K. Descheemaeker et al. / Geoderma 132 (2006) 291–314298

shrubs was included in this measurement. These

places are usually covered with leaf litter and there-

fore contribute to sediment trapping.

2.4. Delineation and description of the sediment

source area

The sediment source area for each studied parcel

was mapped based on field surveys and topographical

maps. Next, the susceptibility of the sediment source

area to water erosion was characterized by field as-

sessment of slope gradient, vegetation cover, soil

erodibility, stoniness and presence of soil conservation

structures.

2.5. Soil analyses

Soil samples were taken from all horizons of the

soil pits, as well as from the topsoil in each of the

studied parcels at three distances (10, 50 and 100 m)

from the top edge of the area. The topsoil samples are

composite samples consisting of a mixture of 5 soil

samples taken at a depth of 0–15 cm, 1 m apart from

each other.

After air-drying, all samples were passed through a

2-mm sieve for laboratory analyses. Texture was de-

termined using the sieve–pipette method (Sheldrik

and Wang, 1993); pH by using a pH-meter in 1 :2.5

soil /water and soil /KCl solution ratio (Hendershot et

al., 1993); available P by using the BrayII-extraction

method (Dewis and Freitas, 1970); cation exchange

capacity (CEC) and exchangeable bases by using

silverthiourea as an index cation at the soil pH (Chha-

bra et al., 1975); CaCO3 by reaction with HCl and

titration with NaOH (Allison and Moodie, 1965).

Total C and N content were determined by the dry

combustion method (Tiessen and Moir, 1993) using a

VARIO MAX total element analyzer (Elementar,

Hanau, Germany). Organic C content was then calcu-

lated by subtracting the C content present in CaCO3

from the total C content.

2.6. Data analysis

Correlation analysis and linear regression were

performed to determine and characterize relations be-

tween thickness of deposited sediment and controlling

factors. In some cases Weighted Least Squares (WLS)

regression was used, to correct for unequal error

variances. ANOVA analysis and post hoc tests

(Tukey and Tamhane T2 for equal and unequal error

variances, respectively) were used to test significant

differences between class means for different vari-

ables. All statistical analyses were performed using

SPSS 11.0 (SPSS, 2001).

3. Results

3.1. Soil study

A brief description of soil types and their proper-

ties is given, followed by an interpretation of pedo-

genetic processes and different sediment types in the

three study sites.

3.1.1. May Ba’ati

Calcisols, Calcaric Regosols and Calcaric Cambi-

sols (FAO et al., 1998) are the most common soil

types encountered under degraded rangeland situa-

tions in May Ba’ati (Fig. 1, Table 1). Typical horizons

show light brown colours, a low organic matter con-

tent (less than 1% organic C), a very high CaCO3

content (up to more than 50%) and can be designated

as either old colluvium (with high rock fragment

content) or in situ weathered parent material. In

some places a darker horizon with higher organic

matter content is still preserved on top of the soil

and in these cases soils key out as Calcaric Cambisols.

In the young exclosure (area MXY), which was

only closed since 3 years, most soils are very similar

to the ones under rangeland, although several auger-

ings show a darker upper horizon, leading to the

assumption that soils were less degraded in this loca-

tion before it was closed. In the first meters from the

top of the exclosure, shallow sediment deposits were

found, deposited since the area was closed. The sed-

iment can be distinguished from other soil material as

it shows characteristics of the soil material from the

upslope sediment source area (non-calcareous materi-

al in which basaltic stones and sandstones are present

and absence of limestone fragments) and by its rela-

tively low rock fragment content (maximum 15%).

The upslope sediment source area consists of flatter

land on top of the limestone cliff, bordering the upper

side of the parcel. Soils in the sediment source area are

K. Descheemaeker et al. / Geoderma 132 (2006) 291–314 299

mostly Cumulihaplic Regosols (FAO et al., 1998)

formed in colluvium from sandstone and basaltic

parent material.

Besides the presence of Humi-calcaric Cambisols

(FAO et al., 1998), which are distinguished by darker

upper horizons, rich in organic matter (N2.5% organic

C), and deeper-lying light brown calcaric horizons

similar to the ones in the degraded rangeland, the

older exclosures (areas MXOa and MXOb) are char-

acterized by the presence of Phaeozems (FAO et al.,

1998) (Table 2), which have been formed in

allochthonous soil material. Phaeozems are dark

soils rich in organic matter, occurring where enough

precipitation ensures some percolation in most years,

but also where the soil can dry out in some periods

(FAO et al., 1998). The classification as Phaeozems is

justified by the presence of a mollic horizon, a base

saturation of at least 50%, and the absence of second-

ary carbonates and no other diagnostic horizons than

cambic horizons. Phaeozems in the old exclosures of

May Ba’ati are mostly found in the upper sections (0–

Table 2

Soil profile description of representative Phaeozems for the three study si

Haragua May BaVati

Ah 0–20 cm; brownish black colour (7.5 YR 2/2);

sandy clay texture; many sandstone rock

fragments; soft, friable, sticky and plastic

consistence; moderate, coarse and granular

structure; many fine and medium pores; 3.3%

organic carbon; 71% base saturation

0–30 cm; dark

(7.5 YR 3/3);

limestone and

fragments; slig

sticky and ve

moderate, coa

structure; man

4.4% organic

saturation

B1 20–55 cm; dark brown colour (7.5 YR 3/3);

sandy clay texture; abundant sandstone rock

fragments; slightly hard, friable, slightly sticky

and slightly plastic consistence; moderate, coarse

and subangular blocky structure; many fine and

medium pores; 1.0% organic carbon; 63% base

saturation

30–70 cm; bro

(7.5 YR 3/2);

limestone and

fragments; ha

very plastic co

medium and s

structure; com

1.6% organic

saturation

B2 55–120 cm; brownish black colour (7.5 YR 3/2);

sandy clay texture; abundant sandstone rock

fragments; slightly hard, friable, very sticky and

plastic consistence; moderate, very coarse and

subangular blocky structure; many fine and

medium pores; 0.6% organic carbon; 53% base

saturation

70–150 cm;

(10 YR 3/2);

limestone and

fragments; ha

plastic consist

medium and s

structure; com

1.1% organic

saturation

30 m from the top edge) of the slopes or in flatter

positions up to 60 m from the top edge. They consist

of a deep soil profile (up to 1.5 m) of dark, deposited

material originating from the upslope source area,

which is evidenced by (1) the presence of sandstones

and basalt stones, and (2) the sharp transition to a

deeper soil layer showing similarities with the soil

layers found in the degraded situations (high CaCO3

content, light brown colours). In the deposited sedi-

ment layers, a mollic horizon was formed near the

surface under influence of regenerating vegetation. In

the subsoil, a cambic horizon with a blocky structure

is present, the lower part of which showing a CaCO3–

pseudomycelium (FAO et al., 1998) on the ped faces

between 60 and 80 cm depth.

Since it is highly unlikely that all this allochtho-

nous material accumulated since closure 20 years ago,

it must be concluded that the deeper, more strongly

structured horizons, consist of older, conserved sedi-

ment. As we find similar, very deep (up to 2 m)

Phaeozems under the old protected church forest vege-

tes

Adewro

brown colour

clay texture; common

sandstone rock

htly hard, very friable,

ry plastic consistence;

rse and crumby

y very fine pores;

carbon; 65% base

0–50 cm; brownish black colour

(10 YR 2/2); sandy loam texture;

common basalt rock fragments; loose,

very friable, non sticky and non plastic

consistence; moderate, medium and

granular structure; many very fine pores;

2.5% organic carbon; 52% base

saturation

wnish black colour

clay texture; common

sandstone rock

rd, friable, sticky and

nsistence; moderate,

ubangular blocky

mon very fine pores;

carbon; 58% base

50–90 cm; very dark brown colour

(7.5 YR 2/3); loamy sand texture; many

basalt rock fragments; loose, very

friable, non sticky and non plastic

consistence; moderate, medium and

granular structure; many fine pores;

0.6% organic carbon; 74% base

saturation

brownish black colour

clay texture; few

sandstone rock

rd, friable, sticky and

ence; moderate,

ubangular blocky

mon very fine pores;

carbon; 60% base

90–145 cm; very dark brown colour

(7.5 YR 2/3); loamy sand texture;

abundant basalt rock fragments; hard,

friable, slightly sticky and slightly plastic

consistence; moderate, coarse rock

structure; few fine pores; 0% organic

carbon; 81% base saturation

K. Descheemaeker et al. / Geoderma 132 (2006) 291–314300

tation, it is assumed that Phaeozems started being

formed when deforestation and agriculture in the

upslope source areas led to erosion and thus sediment

production that could be trapped by the vegetation

belts bordering the fields at their lower side. When

exactly this process of sediment production, transport

and deposition started is not known, but taking into

account the considerable depths of the Phaeozems, it

must be at least several centuries. Nyssen et al. (2004)

mention estimates of 2000 to 2450 years BP for the

time when deforestation started in Ethiopia. Deposi-

tion carried on for centuries until land pressure became

so high that also the forest vegetation on the slopes was

totally removed and land degradation started there as

well. During the degradation phase, the vegetation

belts, except for the church forests, were destroyed

and the Phaeozems started to erode. However, at the

time of closing, the dark cambic horizons were not

fully eroded in areas like MXOa and MXOb, as they

are still present deeper in the profile. Since closing,

new sediment has been deposited on these conserved,

older layers, and a new mollic horizon with a crumby

structure was formed near the surface.

It is thus clear that sediment deposition has taken

place during two periods, as illustrated in Fig. 4. The

first sediment deposition phase, in which the deeper

Phaeozem layers were formed, is designated as boldQand bmedium-termQ, as it lasted for several centuries.

The second deposition phase started recently, at the

time of closing and it is still a short-term process as

Fig. 4. Sketch to illustrate evolution of land use, sediment deposition and pe

explains processes occurring on the hillslope.

the oldest exclosures found in the study area are only

20 years old.

3.1.2. Haragua

Under the degraded rangeland situation in Haragua

(Fig. 1, Table 1), most common soil types are Rego-

sols (FAO et al., 1998), comprising loamy sand layers

of brown colour and low organic matter content (0.5%

organic C) and lacking any profile development.

Under the protected vegetation of the church forest,

Humi-endoskeletic Phaeozems (FAO et al., 1998)

(Table 2) are found with a mollic horizon on top and

a cambic horizon in the deeper soil layers. Colours of

the soil material range from brownish black to black.

The soil horizon succession in the exclosure pre-

sents a dark brown horizon on top of a brownish black

or black horizon. The fact that a lighter coloured hori-

zon overlies a darker horizon leads to a reconstruction

of soil development similar to the one in May Ba’ati:

Phaeozems have developed under vegetation influence

in trapped sediment originating from upslope areas. In

the degradation phase, the soils of the steep slope,

which is now under exclosure, have been eroded, but

not completely. Since closure, new sediment is being

trapped again by the regenerating vegetation, but as the

vegetation is not yet as rich as before (or as compared to

the church forest vegetation), soil organic matter con-

tent is not that high (1.3% compared to 1.6% organic C

in the Phaeozem under the forest vegetation), which is

also demonstrated by a lighter colour. Most common

dogenesis on steep slope sections now under exclosure. Text in italic

K. Descheemaeker et al. / Geoderma 132 (2006) 291–314 301

soils in the exclosure key out as Thaptomollic Cambi-

sols (FAO et al., 1998), reflecting the fact that con-

served, older sediment layers are buried by recently

deposited sediment.

3.1.3. Adewro

In Adewro (Fig. 1, Table 1) the distinction between

recent, short-term sediment deposited since closure,

and older, conserved sediment could not be estab-

lished. In all studied parcels, Phaeozems (FAO et

al., 1998) are the most common soils (Table 2),

even in the rangeland, where degradation apparently

has not yet led to complete erosion of the mollic

horizon at the surface.

Texture analyses reveal that the sand / silt ratio of

soil material in the rangeland and its source area is up

to four times smaller than that in the other areas and

their respective source areas (0.5 compared to values

around 2). Since texture is so different, it was not

found feasible to use horizon characteristics of the

degraded rangeland area to retrieve reference soil

horizons in the exclosures. However, taking into ac-

count the stone pavement at the soil surface in the

rangeland, there is no doubt that no sediment origi-

nating from the source area was recently deposited

here. Moreover, in view of the fact that the sand

fraction increased remarkably compared to the source

area (from 27% to 36%) at the expense of the silt

fraction (from 53% to 42%), it can be concluded that

no sediment deposition but erosion is taking place

here, removing first finer and leaving behind more

sandy soil material.

To determine the depth of the sediment layer in the

other parcels, the same logic as found for May Ba’ati

and Haragua was continued in Adewro. Although no

Table 3

General characteristics and locations of different sediment types in exclos

Study site Sediment type Characteristics

Haragua Recent, short-term Dark brown co

Old, medium-term Black colour

May BaVati Recent, short-term Not calcareous

(b15%), crum

Old, medium-term,

preserved

Low stoniness

strong structur

presence of ps

Old, medium-term,

remobilized

Calcareous, hi

Adewro No differentiation

differences in lithology exist between sediment source

areas and the parcels (all are on basaltic material),

evidence of deposition under vegetation influence

consisted in (a) upper horizons rich in organic matter

(N3% organic C), with (b) low rock fragment content,

leading to (c) a clear boundary with the underlying,

weathered basaltic parent material.

3.1.4. The concept of reference soil horizons and

different types of deposited sediment layers

Based on the detailed soil analysis in the three

study sites, the concept of reference soil horizons

and deposited sediment layers has been further devel-

oped. In our case, reference soil horizons are part of

generally truncated, in situ developed soils, compris-

ing soil material with low organic carbon content.

They can be found in all land use types for a given

study site. In degraded situations however, the refer-

ence soil horizon is found near or at the soil surface,

while in exclosures or in sites, which have not been

strongly degraded, the same reference soil horizons

are found below deposited sediment layers. As such,

reference soil horizons serve as a reference face: all

soil material found on top of the upper reference soil

horizon and showing (1) evidence of transportation

through the allochthonous nature of the material and

(2) characteristics such as dark colours and high or-

ganic matter content indicating soil formation under

influence of vegetation, is considered as material de-

posited and conserved under influence of vegetation

and called deposited sediment.

While studying the characteristics of deposited

sediment layers and reference soil horizons, a need

for further differentiation was felt (Table 3). A refer-

ence soil horizon can be either in situ weathered

ures of the three study sites

Location

lour Top of profiles

Deeper in profiles

, low stoniness

by structure

First 50–60 m from top of area,

upper soil layer of Phaeozems

(b15%),

e, calcareous,

eudomycelium

First 50–60 m from top of area,

especially in flatter positions,

deeper in profile of Phaeozems

gh stoniness (N15%) From 50–60 m from the top

downwards, top of profiles

Table 4

Characteristics of reference soil horizons and sediment layers

Type of soil layer Colora Rock

fragment

contentb

Bulk

density

(t m�3)

Organic

C (%)

Ntot

(%)

CaCO3

(%)

Pav(ppm)

Base

saturation

(%)

CEC (meq/

100 g soil)

Sand

(%)

Silt

(%)

Clay

(%)

Sand/silt (Sand+silt) /

clay

Haragua S Recent, short-term

(n =7)

Dark brown Few–many 1.47

(0.15)

1.3

(0.9)

0.1

(0.1)

1.1

(1.0)

1.2

(1.3)

55.2

(12.2)

18.2

(8.5)

78.4

(1.3)

8.6

(0.4)

13.0

(1.5)

9.1

(0.5)

6.7

(0.9)

S Old, medium-term

(n =10)

Brownish

black–black

Few–many 1.40

(0.11)

1.6

(1.0)

0.1

(0.1)

1.6

(1.1)

8.5

(8.0)

56.3

(8.7)

17.5

(9.2)

66.2

(7.8)

15.9

(3.7)

17.9

(4.1)

4.4

(1.5)

4.8

(1.3)

R Reference (n =5) Brown Few–many 1.56

(0.13)

0.5

(0.1)

0.1

(0.01)

0.2

(0.1)

0.9

(0.7)

38.1

(2.3)

7.7

(2.3)

86.1

(3.3)

5.6

(1.8)

8.3

(1.6)

16.3

(5.7)

11.3

(2.3)

May

BaVati

S Recent,

short-term (n =12)

Dark brown

to black

Very

few–common

1.01

(0.14)

4.2

(2.4)

0.3

(0.2)

2.4

(1.3)

6.3

(10.6)

88.9

(29.1)

42.9

(15.9)

19.6

(7.9)

43.6

(17.5)

37.1

(18.6)

0.5

(0.3)

2.8

(2.6)

S Old, medium-term,

conserved (n =6)

Dark brown

to black

Very

few–common

1.35

(0.05)

1.5

(0.1)

0.6

(0.8)

14.9

(12.2)

5.4

(3.2)

64.9

(11.6)

50.4

(8.1)

17.3

(5.0)

22.7

(5.0)

60.0

(0.0)

0.8

(0.4)

0.7

(0.0)

S Old, medium-term,

remobilised

(n =10)

Dark brown

to black

Many–

abundant

1.16

(0.117)

2.5

(0.7)

0.2

(0.1)

41.6

(20.4)

2.6

(0.6)

83.0

(7.4)

30.9

(3.5)

24.0

(7.2)

28.3

(4.5)

47.6

(9.3)

0.9

(0.3)

1.2

(0.4)

R Old colluvium

(n =7)

Yellow-orange-bright

or yellowish brown

Common–

abundant

1.40

(0.15)

0.72

(0.4)

0.1

(0.01)

49.8

(23.9)

1.7

(3.6)

470.5

(607.1)

31.7

(44.2)

43.2

(17.2)

31.6

(13.1)

25.3

(10.5)

2.0

(1.9)

3.5

(1.7)

R In situ parent

material (n =6)

Orange-bright or

yellowish brown

Very

few–common

1.59

(0.19)

0.2

(0.3)

0.03

(0.02)

59.0

(37.8)

3.0

(3.3)

177.5

(155.9)

10.6

(5.3)

31.0

(17.1)

37.3

(16.3)

32.0

(7.9)

1.0

(0.8)

2.3

(0.7)

Adewro S Sediment (n =12) Brownish black-

dark brown

Few–common 1.10

(0.24)

3.4

(1.2)

0.3

(0.1)

2.5

(0.4)

6.7

(9.1)

50.0

(3.3)

30.2

(2.8)

52.1

(8.5)

31.5

(5.5)

16.4

(3.9)

1.7

(0.6)

5.5

(1.8)

R Reference (n =9) Dark brown Abundant 1.50

(0.18)

0.2

(0.3)

0.05

(0.04)

7.6

(3.8)

0.0

(0.01)

78.9

(13.4)

34.0

(9.0)

76.2

(1.9)

17.8

(2.4)

6.0

(0.6)

4.4

(0.7)

15.8

(1.8)

Average values with standard deviations between brackets. In the second column, S and R indicate deposited sediment layers and reference soil layers, respectively.a Based on Munsell Soil Colours.b Stoniness: very few: b2%; few: 2–5%; common: 5–15%; many: 15–40%; abundant: 40–80%.

K.Desch

eemaeker

etal./Geoderm

a132(2006)291–314

302

Table 5

Correlation coefficients (Spearman’s rho) for total sediment depth

and other recorded variables for the three study sites

DIS SLO VEGtot

Haragua (n =56) SLO �0.500**VEGtot �0.090 �0.067SEDtot �0.124 �0.008 0.659**

May BaVati (n =92) SLO 0.313**

VEGtot �0.016 �0.447**SEDtot �0.110 �0.447** 0.429**

Adewro (n =77) SLO �0.053VEGtot 0.587** 0.132

SEDtot 0.317** 0.149 0.726**

(SLO: slope gradient (%); VEGtot: total vegetation cover (%);

SEDtot: total sediment depth (cm); DIS: distance from top edge

of the parcel (m)).

**: Correlation is significant at the .01 level (2-tailed).

Table 6

Best fitting regression functions relating total sediment depth (SED-

tot (cm)) with slope gradient (SLO (%)) and or total vegetation

cover (VEGtot (%)) for the three study sites (all given regression

functions are significant at p =0.000)

Site Regression function R2

Haragua (n =56) SEDtot =�19.34+1.604VEGtot�0.00762 VEGtot 2a

0.65b

May BaVati (n =92) SEDtot =126.171�23.275ln(SLO) 0.27

SEDtot =7.712+0.594VEGtot 0.24

SEDtot =89.373+0.567VEGtot�21.826ln(SLO)a

0.67b

Adewro (n =77) SEDtot =2.653+0.426 VEGtota 0.60

a Weighted Least Squares regression was used to correct for

unequal error variances.b The adjusted R2 value is given.

K. Descheemaeker et al. / Geoderma 132 (2006) 291–314 303

parent material or old colluvium. The latter contains

a high content of unweathered rock fragments and is

thought to be deposited through mass transport

(Nyssen et al., 2002), before the old sediment was

deposited. Sediment layers are subdivided in brecent,short-term sedimentQ, deposited since exclosure es-

tablishment, and bold, medium-term sedimentQ. In

May Ba’ati, the latter is further refined by the

specifications bconservedQ and bremobilizedQ. Dur-

ing the degradation phase the remobilized sediment

has been redeposited further downslope, resulting

in a more stony and calcareous nature (see also

Fig. 4).

In Table 4 characteristics of reference soil horizons

and deposited sediment layers are given for the three

study sites. Organic carbon content, total nitrogen

content and available phosphorus are the principal

variables allowing to distinguish between reference

soil horizons and deposited sediment layers for all

study sites, while stoniness, bulk density, texture

and CaCO3 content serve this purpose only in some

sites. Based on the data summarized in Table 4, vari-

ables which can be easily used in the field were

determined for the three study sites. In May Ba’ati,

colour, the presence of CaCO3 and rock fragment

content are the most clearly distinguishing variables.

In Haragua indications on colour and texture lead to

correct designations, while in Adewro, knowledge on

rock fragment content and texture will do so. Using

also the indications given in Table 3, a further refine-

ment between types of deposited sediment layers was

carried out in the field.

As two clearly different sediment deposition

phases were distinguished, it was decided to split up

the analysis of the data collected through the transect

study in an analysis of recent, short-term sediment

deposits on the one hand and total deposited sediment

on the other hand. The latter stands for the total

thickness of recent, short-term and old, medium-

term deposited sediment layers taken together and as

such, it reflects the total thickness of fertile soil pres-

ent in exclosure areas. In the remainder of this paper,

the terms recent, short-term and old, medium-term

will be simply referred to as recent and old.

3.2. Total sediment deposits

Relations between total sediment thickness and

explanatory variables, such as slope gradient, current

vegetation cover and distance from the top edge of the

parcel, have been analyzed for all study sites by

correlation (Table 5) and linear regression analysis

(Table 6 and Fig. 5). One can criticise using current

vegetation cover to represent the medium-term vege-

tation cover that is believed to have influenced total

sediment thickness. However, it can be assumed that

within an exclosure, places, which nowadays have

low vegetation cover were most probably bearing

less vegetation during the degradation phase also.

The same reasoning holds for zones which currently

have high vegetation cover. These zones are found on

more favorable locations, supporting also more vege-

tation both during the forest and the degradation

phase. As such it does seem reasonable to take current

Haragua

SE

Dto

t (cm

)S

ED

tot (

cm)

0

20

40

60

80

100

120

May Ba'ati

0

20

40

60

80

100

May Ba'ati

0

20

40

60

80

100

120

140

Adewro

VEGtot (%)

0 20 40 60 80 100

VEGtot (%)

0 20 40 60 80 100

VEGtot (%)

0 20 40 60 80 100

SLO (%)

0 20 40 60 80 100

SE

Dto

t (cm

)S

ED

tot (

cm)

0

10

20

30

40

50

60

n = 77R2 = 0.60SEDtot = 2.653+0.426VEGtot

n = 92R2 = 0.27SEDtot = 126.171–23.275ln(SLO)

n = 92R2 = 0.24SEDtot = 7.712+0.594VEGtot

n = 56R2 = 0.65SEDtot = –19.34+1.604VEGtot –0.00762VEGtot 2

Fig. 5. Best fitting regression functions relating total sediment depth (SEDtot) with slope gradient (SLO) or vegetation cover (VEGtot) for the

three study sites.

K. Descheemaeker et al. / Geoderma 132 (2006) 291–314304

vegetation cover as a substitute for vegetation cover

over the studied medium-term period.

Total sediment depth in Haragua is significantly

correlated with total vegetation cover only (Table 5),

further explained by a significant regression, using the

quadratic term of total vegetation cover as explanatory

variable (adjusted R2=0.65; n=56; p =0.000) (Table

6 and Fig. 5). Neither slope gradient nor distance is

correlated with total sediment depth. As it is expected

that more sediment is deposited close to the sediment

source area near the top edge of the parcel and where

the slope gradient is relatively low, it should not

surprise that for the exclosure and the church forest

in Haragua, total sediment depths are found to be

relatively higher near the top edge (despite the steep

slope) and at the bottom (thanks to the gentle slope).

ANOVA analysis and post hoc testing for com-

parison of mean values of total sediment thickness

were conducted to establish slope gradient and veg-

etation cover classes with significantly different total

sediment thicknesses. This analysis reveals two

clear groups of total vegetation cover in Haragua

(Table 7). The first group (classes with vegetation

cover below 40%) can be found in the rangeland,

which has an average vegetation cover of 25%

(Table 1), while classes with vegetation cover of

more than 40%, forming the second group, are

encountered in the exclosure (average total vegeta-

tion cover of 50%, see Table 1) and in the forest

(average total vegetation cover of 75%, see Table

1), where total sediment thicknesses are more than 3

times higher (Table 7).

Table 8

Correlation coefficients (Spearman’s rho) for all recorded variables

for the Eucalyptus forest in Adewro (n =19)

DIS SLO VEGtot VEGunder VEGeuc

SLO 0.048

VEGtot �0.371 0.283

VEGunder 0.114 0.357 0.792**

VEGeuc �0.253 �0.246 �0.489* �0.830**SEDtot 0.303 0.239 0.258 0.421 �0.413(SLO: slope gradient (%); VEGtot: total vegetation cover (%);

VEGunder: understorey vegetation cover (%); VEGeuc: Eucalyptus

tree cover (%); SEDtot: total sediment depth (cm); DIS: distance

from top edge of the parcel (m)).

**: Correlation is significant at the .01 level (2-tailed); *: Correla-

tion is significant at the .05 level (2-tailed).

Table 7

Mean values of total sediment depth (in cm) for vegetation cover

classes and slope gradient classes for the three study sites and results

of ANOVA analysis and post hoc tests to compare means

Classes Haragua

(n =56)

May Ba’ati

(n =92)

Adewro

(n =77)

Vegetation cover 0–20% 4.3 a 15.2 a 5.5 a

21–40% 16.6 a 29.0 ab 17.2 b

41–60% 58.2 b 36.2 ab 23.4 bc

61–80% 57.3 b 44.4 b 32.3 c

81–100% 56.8 b 66.8 c 33.2 c

F-value

(sign.)

12.415

(0.000)

7.695

(0.000)

18.040

(0.000)

Slope gradient 0–30% 23.8 a 60.4 a 13.9 a

31–65% 40.8 a 34.5 b 24.2 a

66–100% 35.5 a 26.5 b 24.4 a

F-value

(sign.)

0.868

(0.426)

13.508

(0.000)

3.750

(0.028)

Vegetation cover and slope gradient classes within the same site

have significantly different sediment depths if they have no letters in

common (based on Tamhane’s T2 test for unequal variances for

May Ba’ati and on the Tukey test for Adewro and Haragua).

K. Descheemaeker et al. / Geoderma 132 (2006) 291–314 305

In May Ba’ati, total sediment depth is significant-

ly correlated with slope gradient (negative correla-

tion) and with total vegetation cover (positive

correlation). Distance from top is negatively, but

not significantly correlated with total sediment

depth (Table 5). The results of the linear regression

analysis confirm these findings, bringing forward

(the natural logarithm of) slope gradient and total

vegetation cover as best explanatory variables both

in single (R2=0.27 for slope gradient and R2=0.24

for total vegetation cover; n =92; p =0.000) as in

multiple regression (adjusted R2=0.67; n =92; p =

0.000) (Table 6 and Fig. 5).

After ANOVA analysis and post hoc testing, a

difference in total sediment depth between the first

slope gradient class and the other classes was dem-

onstrated for May Ba’ati (Table 7), meaning that in

areas with slope gradient less than 30% almost

the double amount of sediment is trapped and

conserved compared to areas where slope gradients

are higher. Concerning total vegetation cover, it is

striking that there is still a significant difference in

sediment depth under a vegetation cover of 60%

to 80% compared to a cover of 80% to 100%

(Table 7).

In Adewro, total sediment depth is significantly

and positively correlated with total vegetation cover

and distance from the top of the parcels (Table 5).

The correlation between the latter variable and

sediment depth is probably due to a strong positive

correlation between total vegetation cover and dis-

tance. When only the Eucalyptus forest is consid-

ered, it is noted that total sediment depth is not

significantly correlated with any other variable, and

that even a negative correlation exists between soil

cover by Eucalyptus trees and total sediment depth

(Table 8). An explanation for this different behav-

iour can be that Eucalyptus trees suppress the

understorey vegetation (illustrated by the signifi-

cantly negative correlation between these 2 vari-

ables, Table 8), which plays a role in trapping

sediment (positive correlation between sediment

depth and understorey vegetation). Our findings

concerning this phenomenon are confirmed by

Fiedler and Belay (1988) who stated that Euca-

lyptus forests in Ethiopia in areas with rainfall

less than 750 mm have a weakly developed un-

derstorey and forest floor which may enhance run-

off and erosion and thus counteract sediment trapping

processes.

The relation between total vegetation cover and

total sediment depth in Adewro is further illustrated

through linear regression (R2=0.60; n =77; p =0.000)

(Table 6 and Fig. 5) and the results of the ANOVA

analysis (Table 7).

Slope gradient and total vegetation cover measured

along transects explain the spatial distribution of total

sediment thicknesses within a studied parcel. The

average depth of total sediment in a parcel as a

whole is believed to be influenced first of all by the

K. Descheemaeker et al. / Geoderma 132 (2006) 291–314306

ratio between sediment source area and trapping area

and secondly by the land use of the parcel, variables

which do not vary along the studied transects. Not-

withstanding the fact that the current land use of a

parcel is not representative for the land use during

the entire period of sediment deposition, it influences

how much of the deposited sediment could be pre-

served. As such, a logical scoring from 1 to 4 was

attributed to each land use type according to its

degree of degradation over time, so that land use

type could be treated as a quantitative variable in the

analysis. Land use type as well as the ratio of

sediment source area to trapping area are positively

correlated to average total sediment depth, although

only the former significantly (n=11; p =0.027).

When both variables are combined, a significant

Haragua - exclosure, HXY

distance from top (m)

0 20 40 60 80 100 120

0 20 40 60 80 100 120

rece

nt, s

hort

-ter

m s

edim

ent d

epth

(cm

)

0

5

10

15

20

slop

e gr

adie

nt a

nd v

eget

atio

n co

ver

(%)

0

20

40

60

80

100

0

20

40

60

80

100

May Ba'ati - old exclosure B, MXOb

distance from top (m)

rece

nt s

edim

ent d

epth

(cm

)

0

5

10

15

20

slop

e gr

adie

nt a

nd v

eget

aito

n co

ver

(%)

Fig. 6. Overview of transect study results for recent, short-term sediment d

Vegetation cover (dotted lines) and slope gradient (full lines) are also ind

multiple regression function is obtained for the

three study sites taken together:

SEDtot;av ¼ � 26:4þ 15:4LU þ 16:4 S=T

n ¼ 11;R2 ¼ 0:88; p ¼ 0:000� �

Where SEDtot,av: average total sediment depth for

the whole parcel (cm); LU: land use type scoring: 1:

degraded rangeland, 2: young exclosure, 3: old exclo-

sure, 4: church forest; S /T: ratio of sediment source

area to sediment trapping area.

3.3. Recent sediment deposits

Since no distinction between recent and old sed-

iment could be made in Adewro, recent sediment

0 20 40 60 80 100 120

0 20 40 60 80 100 120

0

20

40

60

80

100

0

20

40

60

80

100

May Ba'ati - old exclosure A, MXOa

distance from top (m)

rece

nt, s

hort

-ter

m s

edim

ent d

epth

(cm

)

0

5

10

15

20

slop

e gr

adie

nt a

nd v

eget

atio

n co

ver

(%)

May Ba'ati - young exclosure, MXY

distance from top (m)

rece

nt, s

hort

-ter

m s

edim

ent d

epth

(cm

)

0

5

10

15

20

slop

e gr

adie

nt a

nd v

eget

atio

n co

ver

(%)

epth (bars) as a function of distance from the top edge of the parcel.

icated.

Table 10

Best fitting regression functions relating recent, short-term sediment

depth (SEDrec (cm)) with slope gradient (SLO (%)), distance from

the top of the parcel (DIS (m)) and vegetation cover (VEGtot (%))

for May Ba’ati (all given regression functions are significant at

p =0.000, except the third and the last one, where p =0.008 and

p =0.026, respectively)

Site Regression function R2

All land use types

together (n =42)

SEDrec =18.268�4.462ln(DIS) (*) 0.67

SEDrec =28.512�6.476ln(SLO) (*) 0.48

SEDrec =�0.467+0.110(VEGtot) (*) 0.16

MXOa (n =14) SEDrec =17.096�4.239ln(DIS) 0.85

MXO b (n =16) SEDrec =28.154�6.843ln(DIS) 0.80

MXY (n =12) SEDrec =7.571�1.966ln(DIS) (*) 0.40

(*) Weighted Least Squares regression was used to correct for

unequal error variances.

(MXOa, MXOb, MXY,: old exclosures A and B and young exclo-

sure in May Ba’ati).

K. Descheemaeker et al. / Geoderma 132 (2006) 291–314 307

thickness was only measured in May Ba’ati and

Haragua.

The different pattern of deposition in these two

study sites is immediately striking from Fig. 6. In

Haragua, recent sediment depth shows a regular distri-

bution over the whole hillslope with an average of

about 7 cm, while in May Ba’ati, recent sediment

depth decreases rapidly with distance from the top to

reach a value of zero at a distance of 40 or 60 m for the

young and old exclosures, respectively. In the young

three-year-old exclosure of May Ba’ati (MXY), on

average 5 cm of recent sediment is found just below

the top edge of the area, while in the old exclosures of

20 years, recent sediment depths range from 1.6 cm at

40 m distance up to 10 cm at 5 m distance from the top

edge in MXOa and from 2.5 cm to 17 cm in MXOb at

the same respective distances from the top edge.

From Fig. 6 it can be assumed that in May Ba’ati

strong correlations exist between recent sediment

depth, slope gradient and vegetation cover, while in

Haragua none of these are likely, as was confirmed by

correlation analysis (Table 9). Regression functions

relating recent sediment depth with slope gradient,

distance and vegetation cover were significant only in

May Ba’ati (Table 10 and Fig. 7). Multiple regression

with a combination of explanatory variables could not

increase the R2 of the simple regression with distance.

When regression functions relating recent sediment

depth with distance are calculated for the different

land use types of May Ba’ati separately, R2 increases

except for the young exclosure MXY (Table 10). This

is not surprising since the variation in recent sediment

Table 9

Correlation coefficients (Spearman’s rho) for recent, short-term

sediment depth and other recorded variables for the exclosures of

two study sites

DIS SLO VEGtot

Haragua

(n =18)

SLO �0.637**VEGtot �0.146 0.218

SEDrec 0.278 �0.085 0.098

May Ba’ati

(n =42)

SLO 0.545**

VEGtot �0.213 �0.495**SEDrec �0.784** �0.588** 0.391*

(SLO: slope gradient (%); VEGtot: total vegetation cover (%);

SEDrec: recent, short-term sediment depth (cm); DIS: distance

from top edge of the parcel (m)).

* Correlation is significant at the .05 level (2-tailed).

** Correlation is significant at the .01 level (2-tailed).

depth caused by the fact that different land use types

are taken together is lost. For the young exclosure

however, variation in recent sediment depth is still

relatively high (reflected by a lower R2), because in

this area vegetation cover is less continuous (Fig. 2),

which induces also a higher spatial variability in

deposited sediment depth.

Using averages of the recorded values for recent

sediment depth and a mean bulk density value of 1.47

Mg m�3 for Haragua and 1.01 Mg m�3 for May Ba’ati

as determined for recent sediment in these areas (Table

4), the mass of deposited sediment per hectare since

closure was calculated. Dividing these results by the

respective years of closure to obtain an estimation of

the annual short-term sediment deposition rate resulted

in values of 25 to 93 Mg ha�1 yr�1 for May Ba’ati and

123 Mg ha�1 yr�1 for Haragua (Table 11). For the

exclosures in May Ba’ati average values for annual

short-term sediment deposition rates range from 0.52

to 1.77 cm yr�1 for locations near the top edge of the

parcels and decrease rapidly to zero with distance

(Table 11). This trend is not observed in the exclosure

in Haragua, where the deposition rate fluctuates be-

tween 0.9 and 1.8 cm yr�1. The weighted average for

the strip in which sediment deposition occurs, amounts

to almost 1.4 cm yr�1 in Haragua, whereas the weight-

ed average values for the May Ba’ati exclosures don’t

exceed 0.8 cm yr�1 (Table 11).

Differences in recent sediment deposition rate be-

tween different land use types can be caused by several

factors, such as the area and the condition of the

DIS (m)

0 10 20 30 40 50 60 70

SE

Dre

c (c

m)

0

5

10

15

20

25

n = 42

R2 = 0.67

n = 42R2 = 0.48

SEDrec = 18.268 – 4.462ln(DIS)

SEDrec = 28.512 – 6.476ln(SLO)

n = 42R2 = 0.16SEDrec = –0.467 + 0.11(VEGtot)

SLO (%)

0 20 40 60 80 100 120

0 20 40 60 80 100 120

SE

Dre

c (c

m)

0

5

10

15

20

25

VEGtot (%)

SE

Dre

c (c

m)

0

5

10

15

20

25

Fig. 7. Best fitting regression functions relating recent, short-term

sediment depth (SEDrec) with distance from the top edge of the

parcel (DIS), slope gradient (SLO) and total vegetation cover

(VEGtot) for May Ba’ati.

Table 11

Recent, short-term sediment deposition for Haragua and May Ba’at

DIS HXY MXOa MXOb MXY

SEDrec (Mg ha�1) (years of closure)

n.a. 616 (5) 512 (20) 928 (20) 279 (3)

SEDrec (Mg ha�1 yr�1)

n.a. 123.2 25.6 46.4 93.0

SEDrec (cm yr�1)

5 0.90 0.52 0.84 1.77

20 1.13 0.19 0.43 0.53

40 1.82 0.08 0.13 0

60 1.45 0 0 0

80 1.48 0 0 0

100 1.59 0 0 0

120 1.24 0 0 0

weighted average 1.37 0.26 0.46 0.77

SEDrec: recent, short-term sediment; DIS: distance from top edge of

the parcel (m)).

(HXY: exclosure in Haragua; MXOa, MXOb, MXY: old exclosures

A and B and young exclosure in May Ba’ati).

K. Descheemaeker et al. / Geoderma 132 (2006) 291–314308

sediment source area. Also the state of the exclosure

itself (vegetation cover, slope gradient) could play a

role. In contrast to total sediment however, recent

sediment deposition rate is inversely related to land

use type, as in the young exclosures deposition rates

are much higher. Furthermore, taking into account the

pattern of deposition in relation to distance from the top

in May Ba’ati, it is concluded that all sediment that

enters into the exclosure is deposited, regardless of the

exclosure’s condition. In Haragua on the contrary, re-

cent sediment depths do not tend to zero with increas-

ing distance from the top edge, so that this assumption

does not hold and trapping efficiencies are lower than

100% here. The explanation for different sediment

deposition rates must therefore be related to the ratio

of sediment source to trapping area and the erodibility

of the sediment source area rather than in the conditions

of the exclosure itself.

Assuming complete sediment deposition within the

exclosures, erosion rates for the sediment source area

were calculated (Table 12). Since 100% sediment trap-

ping cannot be assumed in Haragua, erosion rates there

can be even higher than the value indicated in Table

12. The different recent sediment deposition rates

(25.6 versus 46.4 Mg ha�1 yr�1) in otherwise similar

parcels like MXOa and MXOb can be explained by

their different ratio of erosion area to trapping area (2.7

versus 4.6; Table 12). The area ratios of MXY and

MXOb are very similar, but their deposition rates are

not, which is reflected in different erosion rates for the

i

Table 12

Sediment deposition areas, sediment source areas and their assumed erosion rate for exclosures of Haragua and May Ba’ati

Sediment trapping

area (ha)

Sediment source

area (ha)

Ratio source /

trapping area

Short-term deposition

rate (Mg ha�1 yr�1)

Erosion rate

(Mg ha�1 yr�1)

HXY 8.5 26.3 3.1 123.2 39.8

MXOa 1.6 4.3 2.7 25.6 9.5

MXOb Part 1 6.9

Part 2 5.6

Total 2.7 12.5 4.6 46.4 10.0

MXY 0.5 2.5 5.0 93.0 18.6

(HXY: exclosure in Haragua; MXOa, MXOb, MXY: old exclosures A and B and young exclosure in May Ba’ati).

K. Descheemaeker et al. / Geoderma 132 (2006) 291–314 309

source areas (18.6 compared to 10.0 Mg ha�1 yr�1).

The findings concerning erosion rates in May Ba’ati

correspond well with our field observations: the area

just on top of MXY is steeper than the areas on top of

MXOa and MXOb (20% compared to less than 10%)

and moreover, there are no soil conservation measures

taken in the area just on the top of MXY, while on top

of MXOa and MXOb, closely spaced stone bunds are

present in otherwise very similar areas concerning

erodibility. Erosion rates in the sediment source area

for the exclosure in Haragua are two to four times

higher than in May Ba’ati, which can be explained

by the fact that along its’ total length, this area contains

a very steep and bare cliff of highly erodible Adigrat

sandstone, without any soil conservation measures. At

the sandstone rock surface a large amount of loose

sand is observed, which is transported into the exclo-

sure during rainstorms. Apart from the high erosion

rates of about 40 Mg ha�1 yr�1 in the Haragua sedi-

ment source area, the obtained erosion rates for May

Ba’ati correspond well with previously measured ero-

sion rates of 9.9 (Nyssen, 2001) up to 18.13 Mg ha�1

yr�1 (Desta et al., in press) for cropland with soil

conservation structures in the same study region.

Also specific sediment yield rates for catchments

(from 1 up to 24 km2) in Tigray are of the same

order of magnitude, ranging from 2 to 18 Mg ha�1

yr�1 (Nigussie et al., in press).

4. Discussion

4.1. Pedogenesis

According to FAO et al. (1998) Phaeozems are

typical soils of the wetter part of the steppe regions

and their regional distribution does not include the

highlands of Ethiopia. However, the study area fits

into the description of a typical Phaeozem environ-

ment: cool tropical highland climate, humid enough

for water percolation through the soil in most years

and long dry periods. The formation of Phaeozems

under influence of vegetation in deposited sediment is

also somewhat atypical if compared to the WRB

description (FAO et al., 1998). Hence, our findings

are a valuable contribution to the World Reference

Base for Soil Resources. Vegetation plays a double

role in the buildup of our Phaeozems, as shrubs and

grasses trap the incoming sediment but also add to the

organic carbon content of the original soil material

(Descheemaeker et al., submitted for publication). The

obtained sediment deposition rates can be used to

estimate the time which was needed to build up the

deep Phaeozems encountered in the study sites. As-

suming a depth of 150 cm and a deposition rate of 0.5

cm yr�1 (see Table 11) it is concluded that sediment

trapping started about 300 years ago. This estimate

gives also an indication of the time that agriculture

and grazing started to lead to widespread land degra-

dation and erosion in the upslope sediment source

area.

4.2. Total sediment deposits

Results of statistical analyses indicate that in all

sites, the spatial distribution of total sediment was

significantly related to current vegetation cover,

while in only one site, slope gradient was also an

explaining variable. Current vegetation cover can

only provide an indication of vegetation cover over

the medium-term. However, the recurrent correlation

of current vegetation cover with total sediment

K. Descheemaeker et al. / Geoderma 132 (2006) 291–314310

depth, which was demonstrated for all study sites,

proves that it is a valid substitute. In contrast to

recent sediment depth, no relation with distance

was found for total sediment depth. A possible

explanation is that soil material deposited under

former forest vegetation was remobilized after de-

forestation and redeposited further downslope, so

that the initial relation with distance became

blurred. The strong regression predicting average

total sediment depth of a parcel shows that the

parcels’ land use and its’ ratio of sediment source

area to trapping area represent the general condi-

tions for a hillslope bearing more or less total

sediment. Local effects of vegetation cover and

slope gradient are key determinants of total sedi-

ment distribution within a parcel.

4.3. Recent sediment trapping in exclosures

By studying recent sediment deposits, which is

the sediment deposited under the current land use,

i.e., regenerating vegetation, a better understanding

of the process of sediment trapping in exclosure

areas was obtained. For the study site May Ba’ati,

results indicate a significant relation between recent

sediment thickness and both slope gradient (negative

correlation) and vegetation cover (positive correla-

tion). Many studies on water erosion processes

revealed significant relations between slope gradient

(positive) and vegetation cover (negative) on the one

hand and sediment yields on the other hand (Sorriso-

Valvo et al., 1995; Gutierrez and Hernandez, 1996;

Sole-Benet et al., 1997; Woo et al., 1997; Cerda,

1998; Puigdefabregas et al., 1999; Braud et al.,

2001; Erskine et al., 2002). As expected, these rela-

tions for water erosion are just the opposite of the

relations we found for sediment trapping. For the

study site May Ba’ati, our findings further demon-

strate a strong and exponentially decreasing relation

between thickness of recently deposited sediment

and distance from the upper edge of the exclosure.

Van Dijk et al. (1996) found the same type of

relation between sediment concentration and filter

length and also Abu-Zreig (2001) concluded from

a review on factors affecting sediment trapping in

vegetative filter strips that filter length is the most

significant factor. In Haragua however, this typical

pattern of exponentially decreasing recent sediment

depth with distance from the source area was not

found. This can be explained by the fact that very

high slope gradients of almost 70% are found just

below the top edge of the exclosure. This could

have caused the deposition to move downslope as

was also reported by Jin and Romkens (2001).

Differences between study sites in sediment deposi-

tion patterns and the way factors control these pro-

cesses, could not be clearly explained by differences

in lithology, except for the Adigrat sandstone in

Haragua, which, through its high erodibility, can

explain the very high sediment deposition rates in

the exclosure.

In the literature on sediment deposition, only a few

studies (Beuselinck et al., 2000; Wallbrink et al.,

2002) can be found on actual amounts, locations

and conditions of sediment deposited in field situa-

tions. A problem complicating the comparison be-

tween this and their study is the difference in

characteristics of source areas and vegetation strips.

Wallbrink et al. (2002) found an average deposition

rate of 17 Mg ha�1 yr�1 in filter strips bordering

forest logging areas. Beuselinck et al. (2000) recorded

deposition rates of 4 and 15 Mg ha�1 after 1 big

rainfall event for 2 different catchments in the Belgian

Loam Belt. Vegetation-controlled deposition in the

latter study occurred mainly due to small strips of

vegetation in between fields. The values we measured

for yearly sediment deposition (25.6 to 123.2 Mg

ha�1, Table 11) are clearly higher. Compared to the

results of Beuselinck et al. (2000), this can be

explained by the fact that the latter values refer to

deposition in a total catchment area, and result from

1 rainfall event only. An explanation for the discrep-

ancy with the values of Wallbrink et al. (2002) can

be found in the fact that in the latter study already a

significant amount of eroded sediment from log land-

ings and snig tracks is probably deposited in the

general harvest areas (Croke et al., 1999) before it

reaches the filters. Bojo and Cassells (1995) give an

overview of sediment delivery ratios at catchment

scale ranging from 1% to 30%, found by various

authors for different areas in Ethiopia. Consequently,

70% up to 99% of the eroded sediment was depos-

ited before reaching the catchment outlet. The very

high trapping capacity of exclosure areas, which was

demonstrated in this study, can at least partly explain

this phenomenon.

K. Descheemaeker et al. / Geoderma 132 (2006) 291–314 311

The analogy between exclosures in their landscape

context and banded vegetation, described by Valentin

et al. (1999) can be noted, as exclosures are usually

forming vegetation bands on steep slopes bordering

cropland or more degraded rangelands, where natural

vegetation cover is less. Exclosures, as well as band-

ed vegetation phenomena, are source–sink systems

of soil and water. Similar to the so-called interbands,

there is also a lot of runoff and soil erosion in the

cropland and degraded rangelands of our study sites,

since vegetative protection from rain impact is very

low. In both systems the vegetation bands act as

sinks, showing increased infiltration and sediment

deposition.

As the high sediment trapping capacity of exclo-

sures is proved, it is worth considering them as an

effective way to reduce the sediment input to water

reservoirs, which are threatened by rapid siltation

(Nigussie et al., in press), by applying them on se-

lected areas bordering the reservoir. As concluded

from the study on recent sediment, depths of depos-

ited sediment exponentially decrease with distance

from the top edge of exclosures. Therefore, where

slope gradients are not too steep just below the top

edge (less than 50%), an exclosure width of 50–60 m

should be sufficient to trap nearly all of the incoming

sediment.

Furthermore, since exclosures can act as important

sinks of water and sediment, a supplementary option

might be to divert gullies, which are the main trans-

port ways of sediment out of the catchment, into an

exclosure. As such, water and sediment, which would

otherwise be lost, are captured by the exclosure.

However, this potential conservation measure still

has to be tested for its feasibility in the field.

5. Conclusions

In the Tigray highlands, Phaeozems are developing

under influence of vegetation and sediment deposi-

tion. Their formation started probably at least 300

years ago.

In exclosures located downslope from sediment

source areas, two main types of deposited sediment

are found. They are designated as recent, short-term

(b20 years) sediment deposits and old, medium-term

(20–300 years) sediment deposits. Both types are

characterized by specific chemical and morphological

properties, and typical landscape positions.

The detailed description of these characteristics

makes it possible to distinguish the sediment layers

from reference soil layers. By soil augering, the depth

of the deposited sediment layers can be easily deter-

mined in the field.

Total sediment depth in banded exclosure strips is

mainly vegetation-controlled as in only one case a

significant (negative) relation with slope gradient

was found. Besides land use type, the ratio between

sediment source area and trapping area helps to ex-

plain the average depth of total deposited sediment on

a hillslope.

Recent sediment trapping in exclosures is related to

distance from the top edge, slope gradient and vege-

tation cover, but also to the sediment source area and

its’ erodibility. Even on steep slopes (35–50%), very

high short-term (b20 years) sediment deposition rates

were observed in exclosures located below large sed-

iment source areas.

Because of their high sediment trapping capacity,

exclosures established in 50 to 100 m wide strips

across the slope are a very efficient soil and water

conservation measure. They accelerate fertile soil

buildup and prevent important sediment loads from

leaving the catchment or silting up water reservoirs.

However, for exclosures to continue playing their

environmental conservation role in the future, it will

be necessary to take into account also socio-economic

needs. A sustainable and socially fair harvesting sys-

tem of the wood resources or a rotational grazing

system in which several exclosure areas take turns

being opened, are only two examples of how progress

in this direction could be achieved.

Acknowledgements

The research was conducted within the framework

of the Zala–Daget project, funded by the VLIR (Flem-

ish Interuniversity Council, project n8 VLIR/EI 237).Katrien Descheemaeker is funded through an FWO

Vlaanderen Aspirant scholarship and Joni Rossi was

financially supported by a grant from dInterfacultyCouncil for Development Co-operationT of the Katho-lieke Universiteit Leuven. The support by Mekelle

University and the laboratory facilities and help at

K. Descheemaeker et al. / Geoderma 132 (2006) 291–314312

the Faculty of Agricultural and Applied Biological

Sciences of the K.U. Leuven are 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 Girmay Hailemariam, Romha Assefa

and Haftu Assefa for their field assistance.

References

Abubakar, S.M., 1996. Rehabilitation of degraded lands by means

of fallowing in a semiarid area of northern Nigeria. Land

Degradation & Development 7, 133–144.

Abubakar, S.M., 1997. Monitoring land degradation in the semi-

arid tropics using an inferential approach: the Kabomo basin

case study, Nigeria. Land Degradation & Development 8,

311–323.

Abu-Zreig, M., 2001. Factors affecting sediment trapping in vege-

tated filter strips simulation study using VFSMOD. Hydrolog-

ical Processes 15, 1477–1488.

Adejuwon, J.O., Ekanade, O., 1988. A comparison of soil properties

under different land use types in a part of the Nigerian cocoa

belt. Catena 15, 319–331.

Aerts, R., Mitiku Haile, Muys, B., Deckers, J., Hermy, M., Moeyer-

sons, J., 2001. Forest rehabilitation and water conservation in

the Tigray highlands, Northern Ethiopia. European Tropical

Forestry Research Network News 33, 29–31.

Aerts, R., November, E., Mintesinot Behailu,, Deckers, J., Hermy,

M., Muys, B., 2003. Forest rehabilitation: one approach to water

conservation in Central Tigray. Water 6, 34–37 (Eth. J. Wat.

Sci. and Tech.)

Aerts, R., Wagendorp, T., November, E., Mintesinot Behailu,,

Deckers, J., Muys, B., 2004. Ecosystem thermal buffer capac-

ity as an indicator of the restoration status of protected areas in

the Northern Ethiopian Highlands. Restoration Ecology 12,

586–596.

Allison, L.E., Moodie, C.D., 1965. Carbonate. In: Black, C.A., et

al., (Eds.), Methods of Soil Analysis. Part 2. Chemical and

Microbiological Properties, (1st edition) American Society of

Agronomy, Madison, WI.

Bergkamp, G., 1998. A hierarchical view of the interactions of

runoff and infiltration with vegetation and microtopography in

semiarid shrublands. Catena 33, 201–220.

Beuselinck, L., Steegen, A., Govers, G., Nachtergaele, J., Takken,

I., Poesen, J., 2000. Characteristics of sediment deposits formed

by intense rainfall events in small catchments in the Belgian

Loam Belt. Geomorphology 32, 69–82.

Bojo, J., Cassells, D., 1995. Land degradation and rehabilitation in

Ethiopia: a reassessment. AFTES Working Paper, vol. 17. The

World Bank. (48 pp.)

Bosellini, A., Russo, A., Fantozzi, P.L., Assefa, G., Solomon,

T., 1997. The Mesozoic succession of the Mekele outlier

(Tigre Province, Ethiopia). Memorie di Scienze Geologiche

49, 95–116.

Braud, I., Vich, A.I.J., Zuluaga, J., Fornero, L., Pedrani, A., 2001.

Vegetation influence on runoff and sediment yield in the Andes

region: observation and modelling. Journal of Hydrology 254,

124–144.

Cerda, A., 1998. The influence of geomorphological position and

vegetation cover on the erosional and hydrological processes

on a Mediterranean hillslope. Hydrological Processes 12,

661–671.

Chadhokar, P.A., 1992. Area closure for soil conservation in

Ethiopia: potential and dangers. In: Hurni, H., Kebedo Tato

(Eds.), Erosion, Conservation and Small-Scale Farming, a

Reviewed Selection of Papers Presented at the 6th Internation-

al Soil Conservation Conference Held in Ethiopia and Kenya,

6–18 November 1989, Geographica Bernensia, vol. 2. ISCO,

pp. 329–333 (WASWC).

Chadhokar, P.A., Solomon Abate, 1988. Importance of revegetation

in soil conservation in Ethiopia. In: Richwanich, S. (Ed.), Land

Conservation for Future Generations. Proceedings of the 5th

International Soil Conservation Conference, Bangkok, Thailand,

pp. 1203–1213.

Chhabra, R., Pleysier, J., Cremers, A., 1975. The measurement of

the cation exchange capacity and exchangeable cations in soils:

a new method. In: Bailey, S.W. (Ed.), Proceedings of the

International Clay Conference. Applied Publishing Ltd, Wilm-

ette, Illinois, USA, pp. 439–449.

Croke, J., Hairsine, P., Fogarty, P., 1999. Sediment transport, redis-

tribution and storage on logged forest hillslopes in south-eastern

Australia. Hydrological Processes 13, 2705–2720.

Descheemaeker, K., Naudts, J., Nyssen, J., De Geyndt, K.,

Mitiku Haile, Moeyersons, J., Poesen, J., Deckers, J., submit-

ted for publication. Trapping sediment and organic carbon in

exclosures in an Ethiopian highland. Land Degradation and

Development.

Desta Gebremichael, Nyssen, J., Poesen, J., Deckers, J., Mitiku

Haile, Govers, G., Moeyersons, J., in press. Effectiveness

of stone bunds in controlling soil erosion on cropland in

the Tigray highlands, Northern Ethiopia. Soil Use and

Management.

Dewis, J., Freitas, F., 1970. Physical and chemical methods of soil

and water analysis. Soils Bulletin, vol. 10. FAO, Rome, Italy.

Erskine, W.D., Mahmoudzadeh, A., Myers, C., 2002. Land use

effects on sediment yields and soil loss rates in small basins

of Triassic sandstone near Sydney, NSW, Australia. Catena 49,

271–287.

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 Degradation &

Development 9, 529–542.

FAO, 1990. Guidelines for Soil Description. FAO, Rome, Italy.

FAO, ISRIC, ISSS, 1998. World reference base of soil resources.

World Soil Resources Reports, vol. 84. FAO, Rome, Italy.

Fiedler, H.J., Gebeyehu Belay, 1988. Forests and their importance

for soil conservation in Ethiopia. Archiv fur Naturschutz und

Landschaftsforschung 28, 161–175.

Gutierrez, J., Hernandez, I.I., 1996. Runoff and interrill erosion as

affected by grass cover in a semi-arid rangeland of northern

Mexico. Journal of Arid Environments 34, 287–295.

K. Descheemaeker et al. / Geoderma 132 (2006) 291–314 313

Hajabbasi, M.A., Jalalian, A., Karimzadeh, H.R., 1997. Deforesta-

tion effects on soil physical and chemical properties, Lordegan,

Iran. Plant and Soil 190, 301–308.

Hendershot, W.H., Lalande, H., Duquette, M., 1993. Soil reaction

and exchangeable acidity. In: Carter, M.R. (Ed.), Soil Sampling

and Methods of Analysis. Canadian Society of Soil Science,

Lewis Publishers, Boca Raton, USA, pp. 499–512.

Herweg, K., Stillhardt, B., 1999. The variability of soil erosion in

the highlands of Ethiopia and Eritrea. Average and extreme

erosion patterns. Soil Conservation Research Programme, Re-

search Report, vol. 42. University of Berne, Switzerland.

Hurni, H., 1985. Erosion-productivity-conservation systems in

Ethiopia. Proceedings 4th International Conference on Soil

Conservation, Maracay, Venezuela, pp. 654–674.

Jaiyeoba, I.A., 1998. Changes in soil properties related to conver-

sion of savannah woodland into Pine and Eucalyptus planta-

tions, Northern Nigeria. Land Degradation & Development 9,

207–215.

Jennings, S.B., Brown, N.D., Sheil, D., 1999. Assessing forest

canopies and understorey illumination: canopy closure, canopy

cover and other measures. Forestry 72, 59–73.

Jin, C.-X., Romkens, M.J.M., 2001. Experimental studies of factors

in determining sediment trapping in vegetative filter strips.

Transactions of the ASAE 44, 277–288.

Kebrom Tekle, Backeus, I., Skoglund, J., Zerihun Woldu, 1997.

Vegetation on hillslopes in southern Wello, Ethiopia: degra-

dation and regeneration. Nordic Journal of Botany 17,

483–493.

Lacey, S.T., 2000. Runoff and sediment attenuation by undisturbed

and lightly disturbed forest buffers. Water, Air and Soil Pollu-

tion 122, 121–138.

Le Houerou, H.N., 2000. Restoration and rehabilitation of arid

and semiarid Mediterranean ecosystems in North Africa and

West Asia: a review. Arid Soil Research and Rehabilitation

14, 3–14.

Liu, S.L., Fu, B.J., Lu, Y.H., Chen, L.D., 2002. Effects of refores-

tation and deforestation on soil properties in humid mountainous

areas: a case study in Wolong Nature Reserve, Sichuan prov-

ince, China. Soil Use and Management 18, 376–380.

Lumbanraja, J., Syam, T., Nishide, H., Mahi, A.K., Utomo, M.,

Kimura, S., Kimura, M., 1998. Deterioration of soil fertility by

land use changes in South Sumatra, Indonesia: from 1970 to

1990. Hydrological Processes 12, 2003–2013.

Mitiku Haile, Kindeya Gebrehiwot, 1998. Efforts to rehabilitate

degraded lands: local initiatives for planning resource manage-

ment in Tigray (Northern Ethiopia). Sustainable Development of

Dryland Areas of East Africa, Proceedings of the International

Workshop Addis Ababa, November 9–12 1998, European Com-

mission, pp. 319–328.

Nigussie Haregeweyn, Poesen, J., Nyssen, J., De Wit, J., Mitiku

Haile, Govers, G., Deckers, S., in press. Reservoirs in Tigray:

characteristics and sediment deposition problems. Land Degra-

dation and Development.

Nyssen, J., 2001. Erosion processes and soil conservation in a

tropical mountain catchment under threat of anthropogenic de-

sertification— a case study from Northern Ethiopia. PhD Thesis,

Katholieke Universiteit Leuven, Belgium, 380p.

Nyssen, J., Mitiku Haile, Moeyersons, J., Poesen, J., Deckers, J.,

2000. Soil and water conservation in Tigray (Northern Ethio-

pia): the traditional daget technique and its integration with

introduced techniques. Land Degradation & Development 11,

199–208.

Nyssen, J., Moeyersons, J., Poesen, J., Deckers, J., Mitiku Haile,

2002. The environmental significance of the remobilisation of

ancient mass movements in the Atbara–Tekeze headwaters,

Northern Ethiopia. Geomorphology 49, 303–322.

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

Science Reviews 64, 273–320.

Paniagua, A., Kammerbauer, J., Avedillo, M., Andrews, A.M., 1999.

Relationship of soil characteristics to vegetation successions on a

sequence of degraded and rehabilitated soils in Honduras. Agri-

culture, Ecosystems & Environment 72, 215–225.

Puigdefabregas, J., Sole, A., Gutierrez, L., del Barrio, G., Boer, M.,

1999. Scales and processes of water and sediment redistribution

in drylands: results from the Rambla Honda field site in South-

east Spain. Earth-Science Reviews 48, 39–70.

Rey, F., 2003. Influence of vegetation distribution on sediment yield

in forested marly gullies. Catena 50, 549–562.

Sheldrik, B.H., Wang, C., 1993. Particle size distribution. In: Carter,

M.R. (Ed.). Soil Sampling and Methods of Analysis. Canadian

Society of Soil Science, Lewis Publishers, Boca Raton, USA,

pp. 499–512.

Sole-Benet, A., Calvo, A., Cerda, A., Lazaro, R., Pini, R., Barbero,

J., 1997. Influences of micro-relief patterns and plant cover on

runoff related processes in badlands from Tabernas (SE Spain).

Catena 31, 23–38.

Sorriso-Valvo, M., Bryan, R.B., Yair, A., Iovino, F., Antronico, L.,

1995. Impact of afforestation on hydrological response and

sediment production in a small Calabrian catchment. Catena

25, 89–104.

SPSS 11.0.1, 2001. SPSS for Windows, Release 11.0.1. SPSS Inc.

Tenna Shitarek, Sintayehu Manaye, Berihun Abebe, 2001.

Strengthening user-rights over local resources in Wollo,

Ethiopia. IIED Drylands Programme, Issue Paper, vol. 103.

23 pp.

TFAP, 1996. Tigray Forestry Action Plan, Tigray Bureau of Agri-

culture and Natural Resources Development, Regional Govern-

ment of Tigray, Mekelle.

Tiessen, H., Moir, J.O., 1993. Total and organic carbon. In: Carter,

M.R. (Ed.). Soil Sampling and Methods of Analysis. Canadian

Society of Soil Science, Lewis Publishers, Boca Raton, USA,

pp. 187–200.

UNEP, 1994. United Nations Convention to Combat Desertification,

Berne. (71 pp.)

Valentin, C., D’Herbes, J.M., Poesen, J., 1999. Soil and water

components of banded vegetation patterns. Catena 37, 1–24.

Van Dijk, P.M., Kwaad, F.J.P.M., Klapwijk, M., 1996. Retention of

water and sediment by grass strips. Hydrological Processes 10,

1069–1080.

Wallbrink, P.J., Roddy, B.P., Olley, J.M., 2002. A tracer budget

quantifying soil redistribution on hillslopes after forest harvest-

ing. Catena 47, 179–201.

K. Descheemaeker et al. / Geoderma 132 (2006) 291–314314

Wisborg, P., Shylendra, H.S., Gebrehiwot, K., Shanker, R., Tilahun,

Y., Nagothu, U.S., Tewoldeberhan, S., Bose, P., 2000. Rehabil-

itation of CPRS through re-crafting of village institutions: a

comparative study from Ethiopia and India. Paper presented at

the 8th Biennial Conference of the International Association for

the Study of Common Property (IASCP), Bloomington, Indiana,

May 31 – June 4, 2000, 26p.

Woldeamlak Bewket, Stroosnijder, L., 2003. Effects of agroecolo-

gical land use succession on soil properties in Chemoga Water-

shed, Blue Nile basin, Ethiopia. Geoderma 111, 85–98.

Woo, M., Fang, G., di Cenzo, P.D., 1997. The role of vegetation in

the retardation of rill erosion. Catena 29, 145–159.