Catena 131 (2015) 84–91

Contents lists available at ScienceDirect

Catena

j ourna l homepage: www.e lsev ie r .com/ locate /catena

Soil, land use and landform relationship in the Precambrian lowlands ofnorthern Ethiopia

Kassa Teka a,⁎, Jan Nyssen b, Nurhusen Teha a, Mitiku Haile a, Jozef Deckers c

a Department of Land Resources Management and Environmental Protection, Mekelle University, Ethiopiab Department of Geography, University of Gent, Belgiumc Department of Earth and Environmental Sciences, KU Leuven, Belgium

⁎ Corresponding author.E-mail address: kassateka@yahoo.com (K. Teka).

http://dx.doi.org/10.1016/j.catena.2015.03.0100341-8162/© 2015 Elsevier B.V. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 27 March 2014Received in revised form 12 March 2015Accepted 14 March 2015Available online xxxx

Keywords:DMSVLandformLithologyPrecambrianTigray

Soil, landscape and vegetation pattern at a detailed scale (1: 20,000) is non-existent in the Precambrian dominatedlowlands (500–1500 m above sea level) of northern Ethiopia. Current studies at a detailed scale in the regionhave focused on the basalt and Mesozoic rock (limestone and sandstone)-dominated highland (N2300 mabove sea level) areas. The aim of this research was, therefore, to explain the soil distribution as a function oflithology, land use and landform, and to develop a methodology for up-scaling to similar environments. Thisstudy was conducted at Aqushala Watershed in the Precambrian rock dominated Avergelle Lowlands. Theresearch was done based on a discrete model of spatial variation (DMSV). Soil units identified: (1) in the meta-morphosed black limestone formation are Vertic, Endoleptic Calcisol (Humic) at the upper slope (plateau); bothVertic, Endoleptic Cambisol (Calcaric, Humic) and Vertic Leptosol (Calcaric, Humic) at the middle slope (hill);Hypercalcic Calcisol at the foot slope and Grumic Vertisol (Calcaric, Humic) at the lower slope (valley bottom).Majority of the soil units were under cultivation; (2) in the schist and slate formations are Leptosol (Calcaric,Humic) both at the upper slope and at the foot slope positions; Regosol (Calcaric) over Hypercalcic Calcisol atthe mid slope position and Fluvisol (Calcaric, Humic) at the valley bottom; and (3) in the green-reddish-graymetamorphosed banded marl formation are Leptic Calcisol at the upper slope, Haplic Calcisol at the foot slope,and Fluvisol (Calcaric, Humic) at the valley bottom. The model was tested in the Taget (control) Watershedandwas 73% successful. So, if thismodel is used, it can help a lot in every aspect of agricultural or natural resourcemanagement and planning processes.

© 2015 Elsevier B.V. All rights reserved.

1. Introduction

Soil is a dominant factor of the landmainly through its effect on bio-mass production (Gessler et al., 1995; Brunner, 2012). It covers land as acontinuum having properties that vary enormously and continuouslywith depth and with horizontal distances (Gessler et al., 1995). Butthis variation is not random i.e. at any given location on the landscape;there is a particular soil with a unique set of properties (Iqbal et al.,2005). For the most part, soils are the same wherever all elements ofthe five factors (climate, time, vegetation, topography and parentmate-rial) are the same (Jenny, 1941; Dokuchaev, 1883; McKenzie and Ryan,1999; Chaplot et al., 2001; Phillips, 2010). This regularity permits pre-diction of the locations of many different kinds of soils.

When soils are studied in small areas, the effects of topography andparent material on soil becomes apparent (Fikru, 1995; Finzi et al., 1998;Van Breemen and Finzi, 1998; Delin et al., 2000; Bohlen et al., 2001;Fitzpatrick et al., 2003; Venterea et al., 2003). The soil varies along thelandscape, even within limited areas, giving rise to a succession of soil

types, known as a catena (Milne, 1935; Aweto and Enaruvbe, 2010).Studies (e.g. Nizeyimana and Bichi, 1992; Eash and Sandor, 1995;Dahlgren et al., 1997) showed that soil properties and landscape posi-tion are significantly related, mainly where the movement of soil andwater is considered. Landscape topography affects soil physical andchemical properties by erosion and deposition processes, which greatlyinfluence the characteristics and distribution of the soils (Onstad et al.,1985; Nizeyimana and Bichi, 1992; Delin et al., 2000; Bohlen et al.,2001; Venterea et al., 2003; Dotterweich, 2008; Houben, 2008; Awetoand Enaruvbe, 2010; Augustsson et al., 2013; Świtoniak, 2014) in addi-tion to its effect on water table depth which can again impact soilgenesis significantly. Many studies have shown that organic matterand soil nutrient levels are higher in the lower slope segment of thetopography (Abrams et al., 1997; Kravchenko and Bullock, 2000). Thiscan be due to soils in lower slope receive substantial amount of sedi-ments transported fromupslopewhich helps improve their nutrient sta-tus (Aweto and Enaruvbe, 2010). Moreover, soils in lower topographiclocation hold greater quantity of water than higher slope soils and aresaturated with moisture for a much longer period than upper slopesoils (Lopez et al., 2003; Aweto and Enaruvbe, 2010). Studies in dryareas of Australia (Fitzpatrick et al., 2003), for example, have shown

85K. Teka et al. / Catena 131 (2015) 84–91

that shallow loam soils are on the crest, interspersed with shale andsiltstone outcrops (i.e., 5–50% surface cover) and clay soils on steepupper slopes. Surface soil horizons loss occurs on convex parts ofslopes and colluvial material deposition takes place on concaveareas (Dotterweich, 2008; Świtoniak, 2014).

With the notable exception of the constructing patterns of vege-tation in transition zones, local differences in vegetation are alsoclosely associated with differences in relief and parent material(Fikru, 1995). In some cases, soil properties also affect the vegetationtypes (Finzi et al., 1998; Van Breemen and Finzi, 1998) and vice versa(Dotterweich, 2008). Studies (e.g. Finzi et al., 1998; Van Breemenand Finzi, 1998; Dotterweich, 2008) revealed that under naturalconditions, vegetation cover and soil development largely mitigategeomorphic processes, resulting in a stable equilibrium. The naturalwater and matter fluxes change with deforestation and soil erosion.Furthermore, Dotterweich and Dreibrodt (2011) reported thatreducing forest cover and/or increasing frequency of flood eventsrender the fluvial system rich in sediment. The material is washeddown slope and gullies have incised, leading to the deposition of sedi-ments at the slopes and floodplains (Dotterweich, 2008; Hoffmannet al., 2011). These erosional landforms and depositional structuresare the results of past land use, human impact and climate change atbroad temporal and spatial scales. According to Dotterweich andDreibrodt (2011), sediment fluxes in small catchments are highly sensi-tive to changes in local land use. Studies (e.g. Papendick and Miller,1977; De Alba et al., 2004) showed that land use conversion modifiedsoil morphological properties along the slope due to soil materialredistribution.

Nevertheless, the distribution and quality of soil data are not homo-geneous across the world (Dobos et al., 2000). Lacks of financialresources as well as the use of different survey methodologies, soil tax-onomies, and mapping scales are the main obstacles towards creatingglobally valid soil data (Brunner, 2012). Furthermore, the soil, landscapeand vegetation pattern at a detail scale (1: 20,000) is non-existent in thePrecambrian dominated lowlands (500–1500 m above sea level) ofnorthern Ethiopia. The age of the Precambrian rock extended from theorigin of the earth (believed to have been about 4600 million yearsago) to about 570 million years ago, representing nearly ninety percentof geological time (www.oxforddictionaries.com/definition/english/Precambrian).

In the case of Ethiopia, the existing maps, i.e. paper and digital mapsat the national level, have their origin in the 1980s, when the Food andAgricultural Organization (FAO) created a set of maps on geomorpholo-gy and soils at a scale of 1:1 million to assist land use planning (FAO,1983). Current studies at detail scale in northern Ethiopia have focusedon the basalt (Van deWauw et al., 2008), and limestone and sandstone‘Mesozoic rocks’ (Rabia et al., 2013) which is dominated within thehighland (N2300 m above sea level) areas. These studies did not lookat land use.

Hence, this type of soil, landscape and vegetation pattern is likely tobe useful predictions of where these soil issues are likely to occur in thelandscape, land use and lithology. If so, the ability to predict these pat-terns will be important in managing these areas more effectively. Theobjectives of the research were: i) to investigate soil distribution as afunction of lithology, land use and landscape position; ii) to produce amap of these soils and model the soil landscape relationship, and iii)to test the model if it can be used as a tool for similar environments.

2. Materials and methods

2.1. The study area

The Aqushala small scale dam (outlet) (Fig. 1) is located at481,605 m East and 1,479,939 m North Universal Trans MercatorGlobal Positioning System (UTM GPS) coordinates and at an altitudeof 1300 m above sea level. The watershed area is approximately

9.73 km2. This is a dry, lowland (locally called kolla) agro-ecologicalzone. Data taken from Tekeze hydro-electric dam station, about 10 kmsouth west of the Aqushala small scale dam, shows that the averageannual rainfall of the area is 330 mmwith an average annual minimumandmaximum temperature of 20 °C and 28 °C, respectively. The terrainis undulating, hills alternating with plains and valleys. Vegetation coveris made up of scattered acacia trees, riverine forests and bush scrub(Draft Middle Tekeze Livelihood Zone, 2006).

The production system is mixed farming on low lying plains, valleysand foothills. Themain crops grown are Sorghum bicolor (Sorghum), Zeamays (Maize), Eragrostis tef (teff), Sesamum indicum (sesame) andLinum usitatissimum (flax). The major limitations for agricultural pro-duction are low moisture availability and poor soil fertility (DraftMiddle Tekeze Livelihood Zone, 2006). The bed rock is characterizedby its dip direction (to the East) and having different lithologies. Thedominant rock in this catchment is metamorphosed black limestone.Soils in the area are locally classified as Walka (black soil), Baekel(light colored soil) and Afukala (a mixture of soil properties from bothWalka and Baekel).

2.2. Methodology

2.2.1. Pre-field workThe only available document was a small scale (1:250,000) topo-

graphic map. Considering the limited surface (973 ha) of the studyarea, not enough details were available on the topographical map toaccurately delineate the different landforms. Hence, a base map wasdeveloped from Shuttle Radar Topographic Mission (SRTM_90) imag-ery. Land uses were delineated through aerial photo interpretation(API) using Integrated Land and Water Information Systems (ILWIS)3.3 and mapped in ArcView GIS 3.2 software.

2.2.2. Field workPrior to the systematic soil survey, a preliminary soil reconnaissance

of the Aqushala Watershed was conducted. Transect walks were madethrough the watershed to become familiarized with the landscape,landform, land use, etc., as proposed by the World Bank (2005) and todo preliminary establishment of a representative catenae (FAO, 2006).Transect placement and sampling intervals along transects were deter-mined subjectively to capture the full range of soil variability withinlandforms as described by Young et al. (1992). Characterizing both indi-vidual pedon properties and the soil relationships both above and belowon the landscape is important (Soil Survey Staff, 2009). Thus, themethods used to collect data included both surface and subsurfacemethods of land investigation. The surfacemethod of land investigationwas based on visual observation to identify nature and extent of soilproperties, rock outcrops and transportedmaterials. The sub-surface in-vestigations were carried out using profile pits.

The conceptualmodel used in this studywas a discretemodel of spa-tial variation (DMSV) (Bregt, 1992). The model assumes that the land-scape can be divided into discrete polygons of ‘natural’ soil bodies.Hence, the entire area of the catchment was divided into twelve majormapping units based on surface (soil color, vegetation type and density,land use, slope steepness, erosion, stoniness and rock outcrops) andsubsurface landscape parameters (such as CaCO3 content, soil depth,horizon development and profile stoniness). A mapping unit is repre-sented by a set of pixels characterized by the same value for the diag-nostic attributes (Van Orshoven et al., 2011). The identification ofmajor mapping (land) units helped to determine the number of pitsexcavated.

In each land mapping unit, a soil profile pit (at least 1.5 m × 1.5 mand to the depth of the underlying rock) was dug. These soil profilepits were used to demonstrate lateral and vertical changes in the soiland are important for the full description of the soils and for takingsoil samples for chemical and physical laboratory analysis (FAO,2006). The internal properties of the soil (depth of profile and horizon

Fig. 2. Schematic representation of the catena (e.g. slate and limestone succession).

Fig. 1. Location of the Aqushala Watersheds in Tigray province.

86 K. Teka et al. / Catena 131 (2015) 84–91

properties) were described by using the FAO soil description guidelines(FAO, 2006). Explanatory pedogenetic variables at each site wererecorded and a preliminary soil classification was done following(IUSS Working Group WRB, 2007).

Tomap the soils for the entire area, soil augering description follow-ing grid survey techniques was followed. Given the limited time, goodaccessibility and uniformity of the study area, a grid survey pattern of400 m by 200 m (400 m between the lines and 200 m within thelines) was selected, traversing as many physiographic and geologicunits as possible. The distance of the observation within the lines wasnot strictly followed, the choice of the location beingmainly determinedby changes in landform and lithology. The chosen transectswere as per-pendicular as possible to expected boundaries between mapping units(Soil Survey Staff, 2009).

Along these transects, more than 300 soil augering and profiledescriptions were made. The observations in the grid sampling weredone by Dames and Moore Sampler (Auger). Augering was done to a100 cm soil depth or to a limiting layer (rock and very gravely layer)occurring at less than 100 cm (Soil Survey Staff, 2009). Soil descriptionlocations were recorded using a Global Positioning System (GPS). Inthis way, the landscape–geology–soils relationships that are subse-quently used for interpolation of data to cover the whole survey areawere determined.

2.2.3. Post-field workSoil samples were taken from each horizon of the test pits (up to

1.5 m or rock depth). Upon arrival at the laboratory, the samples werefirst air-dried and subsequently, aggregates were broken up by carefulpounding with pestle and mortar. The samples were then passedthrough a 2 mm sieve to obtain the fine earth fraction. Parameters ana-lyzed in the laboratory were of the following: soil texture, pH (of H2O),cation exchange capacity (CEC in cmol+/kg), Ca2+, Mg2+, K+, Na+,available phosphorous (Pava), total N and Organic Carbon (OC).

Standard soil analyses were performed as follows: Particle size— byhydrometer method as outlined by Bouyoucos (1962); Soil pH — by

glass electrode pH meter in 1:2.5 soil/water ratio (Soil Survey Staff,1996); Organic carbon content (OC)— by theWalkley and Black meth-od (Walkley and Black, 1934). Available phosphorus — by Bray No. Imethod (Bray andKurtz, 1945); Exchangeable bases (Potassium, Sodium,CalciumandMagnesium)—byflamephotometry; Total nitrogen content(Nt) — by a modified Kjeldahl method (Christensen and Fulmer, 1927).

Based on the laboratory data, a final soil reclassification was made.Finally, the soil distribution was mapped based on soil variables whichcan be used as input in up scaling it to the other nearby Watersheds.The final mapwas produced at the scale of 1:20,000. Avery (1987) clas-sifies maps at a scale of 1:2500 to 1:25,000 as large scale maps. The soilunitswere then correlatedwith the landforms, parentmaterial and landuse (vegetation cover). To test the quality of the derived model, a con-trol area (TagetWatershed, 325 ha at about 18 kmair distancenortheast

Table 1Identified land units and their description in the Aqushala Watershed, Ethiopia, Africa.

Catena/Rock formation Landunit

Land units description

Black metamorphosedlimestone

1 Located on the plateau, stony, deep soil withsome cracks, and crop land with many scatteredacacia trees.

2 Located at the upper slope, many blacklimestone outcrops with fleshy Aloe succotrina.

3 Located on moderately steep slope (15–30%),with acacia savannah bush land and fleshy Aloesuccotrina, less rock out crop compared to landunit 2.

4 Located at the foot slope, cropland with moresoil cracks and many stones

5 Valley bottom, very deep soil dominated byblack cracking soil type, crop land and fewstones.

Schist and slate 6 Plateau position, flat pattern, many stones butfew rock outcrops–bush land (dense acaciasavannah).

7 At the moderately steep slope (middle slopeposition), deeper and in-situ soil, locallycolluviums, and dominantly woodland.

8 At the foot slope, very stony, shallow soil,undulating topography, and crop land withscattered acacia trees.

9 Flood plain/flat land, very few stones, crop land,very scattered acacia, alluvial/fluvial deposits,on its west edge presence of pebbles of variousdimensions and lithology (looks like ancientriver terrace).

Green ReddishMetamorphosedBanded Marl

10 At the plateau, soft green reddish rock exposedto the surface, resident area and degraded lands.

11 At the foot slope, soft green reddish rock at ashallow depth, degraded + agriculture

12 Flood plain/flat land, mixed vertic and reddishcultivated soils

87K. Teka et al. / Catena 131 (2015) 84–91

of Aqushala Watershed) was selected. The area was chosen becauseno visits were made before creating the base map. Based on the soil–landscape relationship model that was defined in the Aqushala water-shed, the soils of Taget Watershed were predicted. Sixteen profile pits

Fig. 3. Lithological map of the Aqush

(to a maximum depth of 150 cm) were described, classified andmapped in the area, along 2 transects parallel to the slope. The data ob-tained from each soil profile pit were carefully compared with the pre-dicted soil units. A weighting factor for the error was estimated (‘1’denotes for unfit, 0.5 for partially fit and ‘0’ for fit).

3. Results and discussions

3.1. Landscape units

The field survey has indicated three major catenae from which 12landscape units were delineated (Fig. 2 and Table 1). These catenaewere formulatedmainly on the basis of landform and the rock forma-tions. Major landforms identified in the study area (e.g. under theslate and limestone succession) are: the plateau, upper slope, footslope and the valley bottom.

King et al. (1983) suggested that conceptual models of soil distribu-tion in relation to landscape parameters such as ‘landform’ and ‘slopeclass’ are integral to the identification and delineation of mappingunits during soil survey. Studies (e.g. Ellis, 1957; Fikru, 1995; Finziet al., 1998; Van Breemen and Finzi, 1998; Delin et al., 2000; Bohlenet al., 2001; Fitzpatrick et al., 2003; Venterea et al., 2003) also describedcatenary relationships under different topographic situations (gentlyand strongly rolling, roughly undulating and nearly leve1). They notedthat the relationship of soil morphology and degree of development totopographic situation and topographic class determines the areal extentof the different pedons or polypedons. Delin et al. (2000), Bohlen et al.(2001), Venterea et al. (2003) and Augustsson et al. (2013) noted thattopography affects soil physical and chemical properties.

3.2. Soils, lithology, land use and land form relationships

Within the study watershed, three rock formations were identified(Table 1 and Fig. 3):

(i) Metamorphosed black limestone constitutes 42% of the totalarea. This type of formation is a non-foliated metamorphic rockcomposed of re-crystallized carbonateminerals, most commonly

ala Watershed, Ethiopia, Africa.

Fig. 4. Land use map of the Aqushala watershed, Ethiopia, Africa.

88 K. Teka et al. / Catena 131 (2015) 84–91

calcite— CaCO3 or dolomite — CaMg(CO3)2. The term “marble” ismostly used to refer to metamorphosed limestone (Philip, 2001).

(ii) Schist and slate, constitute 34.5% of the total area. Schist ismediumgrade metamorphic rock, formed by the metamorphosis of mud-stone/shale, or some types of igneous rock (Jackson et al., 2005;The University of Auckland, 2005), often finely interleaved withquartz — SiO2 and feldspar (KAlSi3O8–NaAlSi3O8–CaAl2Si2O8)

Fig. 5. Soil map of the Aqushala

(Bishop et al., 1999). Slate is a fine-grained, a low grademetamor-phic rock that splits into thin pieces, foliated metamorphic rockthat is formed through the metamorphism of mudstone/shale orsometimes basalt (Raymond, 1981; The University of Auckland,2005). Clay minerals in the parent rock metamorphose into micaminerals (biotite — (K(Mg, Fe)3(AlSi3O10)(F,OH)2), chlorite —

(Mg,Fe)3(Si,Al)4O10(OH)2(Mg,Fe)3(OH)6) and muscovite —

Watershed, Ethiopia, Africa.

Table 2Soil units and their area in the Aqushalab Watershed, Ethiopia, Africa.

Soil unitsa Soil legend Area (%)

Leptosol (Calcaric, Humic) Lp (ca, hu) 25.9Vertic, Endoleptic Cambisol (Calcaric, Humic) VrlenCM (ca, hu) 19.1Fluvisol (Calcaric, Humic) FL (ca, hu) 17.3Haplic Calcisol HaCL 14.9Vertic, Endoleptic Calcisol VrlenCL 9.9Grumic Vertisol (Calcaric, Humic) GmVR (hu, ca) 4.2Hypercalcic Calcisol CchCL 3.7Vertic Leptosol (Calcaric, Humic) VrLp (ca, hu) 2.4Regosol (Calcaric) over Hypercalcic Calcisol RGbcchCL 2.3Leptic Calcisol LeCL 0.3

a Classification based on IUSS Working Group WRB (2007).b Outlet GPS location of 481,605mEast, 1,479,939mNorth, and 1300mabove sea level.

89K. Teka et al. / Catena 131 (2015) 84–91

KAl2(Si3Al)O10(OH,F)2) which are aligned along foliation planesperpendicular to the direction of pressure (The University ofAuckland, 2005).

iii) Green-reddishmarl constitutes 23.5% of the total area. Marl, alsocalled marlstone, is a calcium carbonate mudstone which con-tains variable amounts of clay and silt (Steneck, 1986).

At every augering, the land use type was recorded. Followingthe land use description and classification methods proposed byAlemayehu et al. (2009) and Belay et al. (2014), four major land usetypes were identified (see Fig. 4): i) Farmland, 67.6% of the total area;ii) Savanna woodland, 29.2%; iii) Residence (home site) area, 2.1%;and iv) Degraded land, 1.2%. On the black metamorphosed limestone,almost all soil units were cultivated unless they are moderately steep(N15%). Whereas, on the schist and slate, the entire plateau was undersavanna woodland/grazing in which terracettes caused directly by live-stock overuse (Buckhouse and Krueger, 1981) can be observed. Themoderately steep slopes of the schist and slate catenae were undersavanna woodlands as the soil was very shallow to support the growthof cereal crops. But the strongly slopping (up to 15%) areas werecropped with shallow rooted crops such as teff, flax and sesame.Hontoria et al. (1999) also showed that land use and soil managementpractices influence the soil nutrients related soil processes such aserosion, oxidation, mineralization and leaching which consequentlymodify the processes of transport and re-distribution of nutrients. Thevery shallow soil unit at the upper and western catchment of thegreen-reddish marl was used for residences as it is too shallow to sup-port crop growth. This supports the findings of Belay et al. (2014) inthat Agricultural land use is usually shifted to other use types due to in-creasingly poor yields from the various agricultural crops.

Based on the delineated geological formations, landform and landuse types; ten soil units were identified in the entire study watershed(Fig. 5). Leptosol (Calcaric, Humic) is the dominant soil unit occupyingabout 26% of the total area, while Leptic Calcisol was the scarcest soilunit with only 0.3% area converge (Table 2). These soil units wereformed in schist and slate, and green-reddish metamorphosed bandedmarl rock formations (Table 3) respectively. Leptic Calcisol is reportedto yield poor crop growth and production (Benin et al., 2003). A similar

Table 3The soil, lithology, and landform relationship model of the Aqushalab Watershed, Ethiopia, Afr

Land form and slope steepness Soils on metamorphosed bla

Plateau and gentle (0–5%) slope VrlenCM (ca, hu) (LUa-1)Hill and moderately steep (15–30%) slope VrLp (ca,hu) (LU-2) + VrlenFootslope and sloping (5–10%) slope CchCL (LU-4)Valley bottom, gentle slope (0–5%) and black color GmVR (hu, ca)/FL (ca, hu) (L

a LU = land units.b Outlet GPS location of 481,605 m East, 1,479,939 m North, and 1300 m above sea level.c Abbreviations are provided in Table 2.

study (Rabia et al., 2013) but in the Mesozoic formations, the Kilte-Awlaelo district, of the Tigray region revealed that Leptosols coverabout 37% of the area. Rabia and others explained that the existence ofthese soil units was related to the dry and sloping nature of the area.According to these authors, the presence of considerable amount ofCambisols in the study area indicates the characteristics of soils usedfor rainfed agriculture and this area is typically affected by erosionprocesses.

Landform, slope and rock formation (Table 3) assessed along threecatenae in the study watershed provided ameans of evaluating soil dis-tribution in relation to landscape features. Similar with the findings ofKing et al. (1983), Fikru (1995), Phillips et al. (1999) and Fitzpatricket al. (2003), the influence of landform was evident in the black meta-morphosed limestone rock formations. The variability of the soil inthis rock formation indicated a marked association of shallow soilswith up-slope convex areas, and of deep soils with lower-slope areas.Within this general framework, occurrence of particular pedons (identi-fied at the subgroup level) may be related to the slope change rate.However, this was not the case in the other rock formations of thestudy area. In contrast to the study of King et al. (1983) who found adeeper soil at the foot slope; schist and slate, and the green-reddishmetamorphosed banded marl rock formations at the foot slope respec-tively yield Leptosol (Calcaric) and Haplic Calcisol, i.e. very shallow soils(FAO, 2006).

3.3. Soil landscape model verification in Taget watershed

The black limestone layers in the Taget Watershed are dippingnorthwest in contrast to the Aqushala watershed where the layers aredipping to the east. The area under green-reddish-graymetamorphosedbanded marl is also absent in the Taget Watershed. Thus, the modeltesting was done under the black limestone, and schist and slate rockformations.

The model developed in the Aqushala Watershed gave satisfactoryresult. As shown in Table 4, the predictions in both black limestoneand schist areas were quite satisfactory. The predicted model was suc-cessful in 73% of the soil units. In the black limestone, 72% of the predic-tion was successful while in the schist and slate, 75% of the predictionfitted with the actual. The biggest error was due to the absence ofHypercalcic Calcisol at the foot slope in the black limestone rock forma-tion in the Taget Watershed. However, the actual soil unit found at thefoot slope is Fluvisol (Calcaric, Humic). In the schist and slate area, thebiggest error was due to the absence of Regosol (Calcaric) OverHypercalcic Calcisol at the mid slope while the actual soil unit isLeptosol (Calcaric, Humic). Therefore, for areas having similar physiog-raphy, lithology and climate; the model can be used but with someground truthing on the less reliable areas.

4. Conclusions

Landform and rock formations were the basis for delineating themajor landscape units in the Aqushala watershed. The major landformsidentified in the study area are: plateau, upper slope, foot slope and thevalley bottom.While the rock formations are: (i) Metamorphosed black

ica.

ck limestone (c) Soils on schistand slate

Soils on green-reddishmetamorphosed banded marl

Lp (ca, hu) LeCLCM (ca, hu) (LU-3) RGbcchCL –

Lp (ca, hu) HaCLU-5) FL (ca, hu) FL (ca, hu)

Table 4Reference soil groups in Tageta Watershed, Ethiopia, Africa.

Lithology Topographic position Slope gradient (%) Land use Mapped soil (b)

Black lime stone Plateau 2–5% Farm land VrlenCLHill 15–30% Bare land VrLp (ca,hu)Foot slope 5–10% Farm land FL (ca, hu)Valley bottom 2–5% Farm land FL (ca, hu)

Schist Plateau 2–5% Farm land Lp (ca, hu)Hill 15–30% Farm land Lp (ca, hu)Foot slope 5–10% Farm land Lp (ca, hu)Valley bottom 2–5% Farm land FL (ca, hu)

a Outlet GPS location of 481,605 m East, 1,479,939 m North, and 1300 m above sea level.b Abbreviations are provided in Table 2.

90 K. Teka et al. / Catena 131 (2015) 84–91

limestone (constituting 42% of the total area); (ii) Schist and slate(34.5% of the total area); iii) Green-reddish marl (23.5% of the totalarea). Four major land use types were identified: i) Farmland, 67.6%;ii) Savanna woodland, 29.2%; iii) Residence (home site) area, 2.1%; iv)Degraded land, 1.2% of the total area.

Based on the lithology, the following major soil units were identi-fied: i) Vertic Endoleptic Calcisol, Vertic Endoleptic Cambisol (Calcaric,Humic), Vertic Leptosol (Calcaric, Humic) and Grumic Vertisol (Calcaric,Humic) that are found on metamorphosed black limestone and weremostly cultivated; ii) Leptosol (Calcaric, Humic), Fluvisol (Calcaric,Humic) and Regosol (Calcaric) Over Hypercalcic Calcisol that arefound on schist and slate, and were mostly used for grazing (as opensavanna wood land); iii) Haplic Calcisol and Leptic Calcisol are foundon the green-reddish gray metamorphosed banded marl, which weremainly degraded lands and used for resident (home sites).

The soil units were also related to their landform. On the black lime-stone catenae, Vertic Endoleptic Calcisol is found at the gentle slope pla-teau position. This soilwas used for cultivation. Vertic Leptosol (Calcaric,Humic) and Vertic Endoleptic Cambisol (Calcaric, Humic) are located atthe moderately steep hill position and were covered by savanna wood-land. Whereas Grumic Vertisol (Calcaric, Humic), the most fertile agri-cultural land, is found at the gentle valley position of the catenae. Atthe schist and slate catenae, Leptosol (Calcaric, Humic) is found bothat the gentle slope plateau and slopping foot slope. Regosol (Calcaric)Over Hypercalcic Calcisol is found on moderately steep hill positionwhere soil deposition occurs. Whereas at the very gentle slope, valleyposition of the catenae, Fluvisol (Calcaric, Humic) is dominant.

The soil–landscape relationship model of the Aqushala Watershedwas verified in an area with similar bio-physiographic setting whichwas the Taget Watershed. The model was found to be efficient (73%efficiency) and can easily be adapted to other areas having similar bio-physiographic setting.

Acknowledgments

Our particular gratitude goes to the Mekelle University-InstitutionalUniversity Cooperation Project (MU-IUC) for their financial and logisticssupport as well as access to essential documents. We would also like tothank the farmers, development agents and teachers of the study area.

References

Abrams, M.M., Jacobson, P.J., Jacobson, K.M., Seely, M.K., 1997. Survey of soil chemicalproperties across a landscape in the Namib Desert. J. Arid Environ. 35, 29–38.

Alemayehu, F., Taha, N., Nyssen, J., Girma, A., Zenebe, A., Mintesinot, B., Deckers, S., Poesen,J., 2009. The impacts of watershedmanagement on land use and land cover dynamicsin eastern Tigray (Ethiopia). Resour. Conserv. Recycl. 53, 192–198.

Augustsson, A., Gaillard, M., Peltola, P., Mazier, F., Bergbäck, B., Saarinen, T., 2013. Effects ofland use and climate change on erosion intensity and sediment geochemistry at LakeLehmilampi, Finland. The Holocene 23, 1247–1259.

Avery, B.W., 1987. Soil survey methods: a review. Technical Monograph No. 18. Soil Sur-vey & Land Resource Centre, Silsoe, p. 86.

Aweto, A.O., Enaruvbe, G.O., 2010. Catenary variation of soil properties under oil palmplantation in South Western Nigeria. Ethiop. J. Environ. Stud. Manag. 3 (1), 1–7.

Belay, K.T., Van Rompaey, A., Poesen, J., Van Bruyssel, S., Deckers, J., Amare, K., 2014. Spa-tial analysis of land cover changes in Eastern Tigray. Land Degrad. Dev. http://dx.doi.org/10.1002/ldr.2275.

Benin, S., Pender, J., Ehui, S., 2003. Policies for Sustainable land management in the EastAfrican highlands: summary of papers and proceedings of a conference held at theUnited Nations Economic Commission for Africa (UNECA), Addis Ababa, Ethiopia,24–26 April 2002. Socio-economics and Policy Research Working Paper 50. ILRI(International Livestock Research Institute), Nairobi, Kenya (194 pp.).

Bishop, A.C., Woolley, A.R., Hamilton, W.R., 1999. Cambridge guide to minerals, rocks andfossils. Cambridge University Press 9780521778817, p. 153.

Bohlen, P.J., Groffman, P.M., Driscoll, C.T., Fahey, T.J., Siccama, T.G., 2001. Plant–soil micro-bial interactions in a northern hardwood forest. Ecology 82, 965–978.

Bouyoucos, G., 1962. Hydrometer method, improved for making particle size analysis ofsoils. Agron. J. 54, 464–465.

Bray, R., Kurtz, T., 1945. Determination of total organic and available forms of phosphorus.SSSA 59, 39–45.

Bregt, A.K., 1992. Processing of soil survey data. Wageningen Agricultural University,Wageningen, The Netherlands.

Brunner, M., 2012. A national soil model of Ethiopia. A geostatistical approach to create anational soil map of Ethiopia on the basis of an SRTM 90 DEM and SOTWIS soil data.(M.Sc thesis). University of Bern, Switzerland, p. 126.

Buckhouse, J.C., Krueger, W.C., 1981. What caused those terracettes? Rangelands 3 (2),72–73.

Chaplot, V., Bernoux, M., Watler, C., Curmi, P., Herpin, U., 2001. Soil carbon storage predic-tion in temperate hydromorphic soils using amorphologic index and digital elevationmodel. Soil Sci. 166, 48–60.

Christensen, L.M., Fulmer, E.I., 1927. A modified Kjeldahl method for the determination ofthe nitrogen content of yeast. Plant Physiol. 2, 455–460.

Dahlgren, R.A., Boettinger, J.L., Huntington, G.L., Amundson, R.G., 1997. Soil developmentalong an elevational transect in the western Sierra Nevada, California. Geoderma 78,207–236.

De Alba, S., Lindstrom, M., Schumacher, T.E., Malo, D.D., 2004. Soil landscape evolutiondue to soil redistribution by tillage: a new conceptual model of soil catena evolutionin agricultural landscape. Catena 58, 77–100.

Delin, G.N., Healy, R.W., Landon, M.K., Bohlke, J.K., 2000. Effects of topography and soilproperties on recharge at two sites in an agricultural field. J. Am. Water Resour.Assoc. 36, 1401–1415.

Dobos, E., Micheli, E., Baumgardner, M.F., Biehl, L., Helt, T., 2000. Use of combined digitalelevation model and satellite radiometric data for regional soil mapping. Geoderma97 (3–4), 367–391.

Dokuchaev, V.V., 1883. Russian Chernozem. ‘Selected Works of V.V. Dokuchaev’ (publ.1967). Kaner N (Trans.). International Program for Scientific Translations, Jerusalem,pp. 1–419.

Dotterweich, M., 2008. The history of soil erosion and fluvial deposits in small catchmentsof central Europe: deciphering the long-term interaction between humans and theenvironment — a review. Geomorphology 101 (1–2), 192–208.

Dotterweich, M., Dreibrodt, S., 2011. Past land use and soil erosion processes in centralEurope. PAGES News 19, 49–51.

Draft Middle Tekeze Livelihood Zone, 2006. Livelihood profile, Tigray Region, Ethiopia.www.heawebsite.org/download/file/fid/365 (Accessed on March 19, 2014).

Eash, N.S., Sandor, J.A., 1995. Soil chronosequence and geomorphology in a semiaridvalley in the Andes of southern Perù. Geoderma 65, 59–79.

Ellis, J.G., 1957. The relationship of topography to soil development. (M.Sc. thesis).Department of Soil Science, University of Saskatchewan, Saskatoon, Sask.

FAO, 1983. Ethiopia — geomorphology & soils map and legend — assistance to land-useplanning. FAO, Ethiopia.

FAO, 2006. Guidelines for soil description (fourth edition). Food and Agriculture Organi-zation of the United Nations, Rome 92-5-105521-1, p. 109.

Fikru, Y.K., 1995. Soil and landscape information system fromusers' requirement perspec-tive case study in the Alora Region-south of Spain. Wageningen Agricultural Univer-sity, Wageningen (the Netherlands).

Finzi, A.C., Canham, C.D., van Breemen, N., 1998. Canopy tree–soil interactions withintemperate forests: species effects on pH and cations. Ecol. Appl. 8, 447–454.

Fitzpatrick, R.W., Thomas, M., Davies, P.J., William, B.G., 2003. Dry saline land: an investi-gation using 505 ground-based geophysics, soil survey and spatial methods nearJamestown, South Australia. CSIRO Land and Water, Adelaide.

Gessler, P.E., Moore, I.D., McKenzie, N.J., Ryan, P.J., 1995. Soil–landscape modeling andspatial prediction of soil attributes. Int. J. Geogr. Inf. Syst. 9 (4), 421–432.

91K. Teka et al. / Catena 131 (2015) 84–91

Hoffmann, T., Thorndycraft, V.R., Brown, A.G., Coulthard, T.J., Damnati, B., Kale, V.S.,Middelkoop, H., Notebaert, B., Walling, D.E., 2011. Human impact on fluvial regimesand sediment flux during the Holocene: review and future research agenda. Glob.Planet. Chang. 72, 87–98.

Hontoria, C., RodrmHguez-Murillo, J.C., Saa, A., 1999. Relationships between soilorganic carbon and site characteristics in Peninsular Spain. Soil Sci. Soc. Am. J.63, 614–621.

Houben, P., 2008. Scale linkage and contingency effects of field-scale and hillslope-scalecontrols of long-term soil erosion: Anthropogeomorphic sediment flux in agriculturalloess watersheds of Southern Germany. Geomorphology 101 (1–2), 172–191.

Iqbal, J., Read, J.J., Thomasson, A.J., Jenkins, J.N., 2005. Relationships between soil–landscape and dry land cotton lint yield. Soil Sci. Soc. Am. 69, 1–11.

IUSS Working Group WRB, 2007. World reference base for soil resources 2006. Firstupdate 2007. World Soil Resources Reports No. 103. Rome, FAO.

Jackson, J.A., Mehl, J.P., Neuendorf, K.K.E., 2005. Glossary of geology. Springer9780922152766, p. 577.

Jenny, H., 1941. Factors of soil formation — a system of quantitative pedology. McGraw-Hill, New York.

King, C.J., Action, D.F., Arnaud, R.J.S.T., 1983. Soil landscape analysis in relation to soil dis-tribution and mapping at a site within the Weyburn association. Can. J. Soil Sci. 63,657–670.

Kravchenko, A.N., Bullock, D.G., 2000. Correlation of corn and soybean grain yield withtopography and soil properties. Agron. J. 92, 75–83.

Lopez, I.F., Lambert, M.G., Mackay, A.D., Valentine, I., 2003. The influence of topographyand pasture management on soil characteristics and herbage accumulation in a hillpasture in the North Island of New Zealand. Plant Soil 255, 421–434.

McKenzie, N.J., Ryan, P.J., 1999. Spatial prediction of soil properties using environmentalcorrelation. Geoderma 89, 67–94.

Milne, G., 1935. Some suggested units of classification and mapping particularly for EastAfrican soils. Soil Res. 4 (3), 183–198.

Nizeyimana, E., Bichi, T.J., 1992. Soil and soil–landscape relationship in the north centralregion of Rwanda, east-central Africa. Soil Sci. 153 (3), 225–236.

Onstad, C.A., Pierce, F.J., Dowdy, R.H., Larson, W.E., 1985. Erosion and productivity interre-lations on a soil landscape. Trans. ASAE 28 (1885–1), 888.

Papendick, R.I., Miller, D.E., 1977. Conservation tillage in the Pacific Northwest. J. SoilWater Conserv. 32, 49–56.

Philip, K., 2001. Dictionary of geology. Penguin Group, London and New York 978-0-14-051494-0, p. 163.

Phillips, P.D., 2010. The perfect soil. 19thWorld Congress of Soil Science, Soil Solutions fora Changing World 1–6 August 2010, Brisbane, Australia, pp. 59–61.

Phillips, J.D., Slattery, M.C., Gares, P.A., 1999. Truncation and accretion of soil profiles oncoastal plain croplands: implications for sediment redistribution. Geomorphology28 (1–2), 119–140.

Rabia, A., Afifi, R.R., Mulualem, Aweke, Bianchi, S., Figuredo, H., Huong, T.L., Lopez, A.A.,Mandala, S.D., Matta, E., Ronchi, M., Hishe, Solomon, Tine, A.K., Youssef, M.S.,Gutirrez, M.G., Yusuf, M.M., Alessandro, V., 2013. Soil mapping and classification: acase study in the Tigray region, Ethiopia. J. Agric. Environ. Int. Dev. 107 (1), 73–99.

Raymond, R.W., 1981. Slate, a glossary of mining andmetallurgical terms. American Insti-tute of Mining Engineers, p. 78.

Soil Survey Staff, 1996. Soil survey laboratory methods manual. Soil Survey InvestigationsRep.No. 42, Ver. 3.0. USDA-SCS, Washington, DC.

Soil Survey Staff, 2009. Soil survey field and laboratory methods manual. In: Burt, R. (Ed.),Soil Survey Investigations Report No. 51, Version 1.0. U.S. Department of Agriculture,Natural Resources Conservation Service.

Steneck, R.S., 1986. The ecology of coralline algal crusts: convergent patterns and adaptivestrategies. Annu. Rev. Ecol. Syst. 17, 273–303.

Świtoniak, M., 2014. Use of soil profile truncation to estimate influence of accelerated ero-sion on soil cover transformation in young morainic landscapes, North-EasternPoland. Catena 116, 173–184.

The University of Auckland, 2005. Flexiblelearning.auckland.ac.nz/rocks_minerals/rocks/schist. html (accessed on March 20, 2014).

Van Breemen, N., Finzi, A.C., 1998. Plant– soil interactions: ecological aspects and evolu-tionary implications. Biogeochemistry 42, 1–19.

Van de Wauw, J., Baert, G., Moeyersons, J., Nyssen, J., De Geyndt, K., Taha, Nurhussein,Zenebe, Amanuel, Poesen, J., Deckers, J., 2008. Soil–landscape relationships in thebasalt-dominated highlands of Tigray, Ethiopia. Catena 75, 117–127.

Van Orshoven, J., Kint, V., Wijffels, A., Estrella, R., Bencsik, G., Vanegas, P., Muys, B.,Cattrysse, D., Dondeyne, S., 2011. Upgrading geographic information systems to spa-tial decision support systems. Math. Comput. Sci. For. Nat. Resour. Sci. 3 (1), 36–41.

Venterea, R.T., Lovett, G.M., Groffman, P.M., Schwarz, P.A., 2003. Landscape patterns of netnitrification in a northern hard wood conifer forest. Soil Sci. Soc. Am. J. 67, 527–539.

Walkley, A., Black, I.A., 1934. An examination of the Degtjareff method for determiningorganic carbon in soils: effect of variations in digestion conditions and of inorganicsoil constituents. Soil Sci. 63, 251–263.

World Bank, 2005. Poverty and Social Impact Analysis Sourcebook. http://go.worldbank.org/ZGZHJEDBZ0.

Young, F.J., Hammer, R., Matta, J.E., 1992. Confidence intervals for soil properties based ondiffering statistical assumptions. In: Millikin, G.A., Schwenke, J.R. (Eds.), Appliedstatistics in agriculture) 4th Annual Kansas State Univ. Conf., 27–28 Apr. 1992,Manhattan, KS. Kansas State Univ., Manhattan.