the impact of land use change and check-dams on...

HYDROLOGICAL PROCESSESHydrol. Process. 22, 4922–4935 (2008)Published online 16 September 2008 in Wiley InterScience(www.interscience.wiley.com) DOI: 10.1002/hyp.7115

The impact of land use change and check-dams on catchmentsediment yield

Carolina Boix-Fayos,1* Joris de Vente,2,3 Marıa Martınez-Mena,1 Gonzalo G. Barbera1

and Vıctor Castillo1

1 Soil and Water Conservation Department, CEBAS, CSIC (Spanish Research Council), Campus Universitario de Espinardo, PO Box 164, 30100Murcia, Spain

2 Physical and Regional Geography Research Group, Katholieke Universiteit Leuven GEO-INSTITUTE, Celestijnenlaan 200 E, 3001, Heverlee,Belgium

3 Desertification and Geoecology Department, EEZA, CSIC (Spanish Research Council) General Segura 1, 04001 Almeria, Spain

Abstract:

Extensive land use changes have occurred in many areas of SE Spain as a result of reforestation and the abandonment ofagricultural activities. Parallel to this the Spanish Administration spends large funds on hydrological control works to reduceerosion and sediment transport. However, it remains untested how these large land use changes affect the erosion processesat the catchment scale and if the hydrological control works efficiently reduce sediment export. A combination of field work,mapping and modelling was used to test the influence of land use scenarios with and without sediment control structures(check-dams) on sediment yield at the catchment scale. The study catchment is located in SE Spain and suffered importantland use changes, increasing the forest cover 3-fold and decreasing the agricultural land 2Ð5-fold from 1956 to 1997. In addition58 check-dams were constructed in the catchment in the 1970s accompanying reforestation works.

The erosion model WATEM-SEDEM was applied using six land use scenarios: land use in 1956, 1981 and 1997, each withand without check-dams. Calibration of the model provided a model efficiency of 0Ð84 for absolute sediment yield. Modelapplication showed that in a scenario without check dams, the land use changes between 1956 and 1997 caused a progressivedecrease in sediment yield of 54%. In a scenario without land use changes but with check-dams, about 77% of the sedimentyield was retained behind the dams. Check-dams can be efficient sediment control measures, but with a short-lived effect. Theyhave important side-effects, such as inducing channel erosion downstream. While also having side-effects, land use changescan have important long-term effects on sediment yield. The application of either land use changes (i.e. reforestation) orcheck-dams to control sediment yield depends on the objective of the management and the specific environmental conditionsof each area. Copyright 2008 John Wiley & Sons, Ltd.

KEY WORDS land use change; check-dams; sediment yield; geomorphological impact; management; reforestation

Received 8 January 2008; Accepted 23 June 2008

INTRODUCTION

Two widely applied soil conservation strategies inMediterranean environments are reforestation and theconstruction of check-dams in rivers and streams. Itis well known that land use changes (including refor-estation) and hydrological correction works may altersubstantially the sediment delivery and water dischargeof a catchment, influencing the geomorphological pro-cesses within the river bed, and sometimes inducingnon-desirable processes for river management (Kondolfet al., 2002). However, until now relatively little has beenknown about the effects of changing land use patternsand the introduction of hydrological correction workson sediment yield at the catchment scale. Evaluationof widely applied management measures in mountainstreams is necessary in order to increase our understand-ing of sediment dynamics, and to allocate resources in

* Correspondence to: Carolina Boix-Fayos, Soil and Water Conserva-tion Department, CEBAS, CSIC (Spanish Research Council), CampusUniversitario de Espinardo, PO Box 164, 30100 Murcia, Spain. E-mail:[email protected]

a responsible manner. Particular attention must be paidto upstream/downstream connections, hillslope/channelconnections, process domains, physical and ecologicalroles of disturbance, and stream resilience (Wohl, 2006).

Check-dam effects on river channels

In-channel structures such as check-dams, create seg-mented longitudinal profiles, alter sediment dynamics,bed and bank stability, interrupt longitudinal movementof nutrients and aquatic organisms, and alter the pas-sage of flood waves (Wohl, 2006). Check-dams induceimportant morphological and granulometrical effects inthe river bed (Boix-Fayos et al., 2007) which changethe hydraulic behaviour of the flow in extreme events(Conesa Garcıa et al., 2004). Besides the control ofthe sedimentary load accomplished by check-dams asreported for different environments (Simon and Darby,2002; Martın-Rosales et al., 2003; Surian and Rinaldi,2003), there are also indications that check-dams inducelocal erosion processes (Gomez-Villar and Martınez-Castroviejo, 1991; Porto and Gessler, 1999; Martın Ros-ales, 2002; Castillo et al., 2007; Conesa-Garcıa and

Copyright 2008 John Wiley & Sons, Ltd.

IMPACT OF LAND USE CHANGE AND CHECK-DAMS ON CATCHMENT SEDIMENT YIELD 4923

Garcıa-Lorenzo, 2008) affecting the sediment budget atthe catchment scale. Disturbances induced by check-damscan be negative when induced erosion is high, or whenthey create a false stabilization image (Garcıa-Ruiz et al.,1996b; White et al., 1997; Gutierrez et al., 1998; Alcov-erro et al., 1999; Boix-Fayos et al., 2007). On the con-trary, the effects of check-dams can have short- and long-term benefits. For example, Xiang-Zhou et al. (2004)explain how a check-dam system in gullies is one ofthe most effective ways to conserve soil and water in theLoess Plateau of China, where reforestation methods arenot successful due to the arid climate and barren soils.The sediment retained by the dams provides a uniqueopportunity for high-yield croplands or orchards. A pos-itive influence of check-dams on the riparian vegetationis reported by Bombino et al. (2006), who found a longi-tudinal diversification of vegetation types and creation ofnew habitats for biological and ecological communities,attributed to the geometric characteristics of check-dams.

Land use change effects on sediment yield

Mountain streams are particularly vulnerable tochanges in hillslope processes because of close couplingwith adjacent hillslopes. Hillslope processes are easilyaltered by land use changes. An increase of vegetationcover leads in general to a decrease in runoff generationand sediment detachment resulting in the medium termin the stabilization of sedimentary structures at the catch-ment scale (Garcıa-Ruiz et al., 1996a; Beguerıa et al.,2006). Sometimes a very limited change in land usecan have a significant effect on regional soil erosionrates (Van Rompaey et al., 2002). However, not only thechange in the cover of different land uses induces changeson sediment yield, but also the change in landscape struc-ture has important consequences for the sediment yield.For example, changes in temporal patterns of cultiva-tion practices (Vezina et al., 2006), changes in the spatialconnectivity between sediment- producing areas and theriver network (Vanacker et al., 2003; 2005), changes inthe combined effects of land use and field sizes (Vanackeret al., 2005; Van Rompaey et al., 2007) and changes inthe location of field boundaries (Van Oost et al., 2000)have remarkable effects on sediment yield. So althoughit is well known that changes in land use affect erosionand sediment yield, an integrated evaluation of the effec-tiveness and impacts of check-dams compared with refor-estation and other land use changes is lacking. Recentlyan evaluation of the effect of land use changes, smallfarm dams and large reservoirs on sediment yield in anAustralian catchment has been carried out (Verstraetenand Prosser, 2008).

Perspective and objectives

In south-east Spain significant land use changes haveoccurred as a result mainly of the abandonment ofagricultural activities and the introduction of reforestationplans. In addition, the Spanish Administration spendslarge funds on hydrological control works to reduce

flood risk and sedimentation of reservoirs. However, itremains untested how these land use changes and thehydrological control work affect the erosion processesat the catchment scale, and if they are really efficient.A previous study (Boix-Fayos et al., 2007) analysed thecombined influence of land use change and a networkof check-dams on the morphological evolution of a riverchannel in SE Spain. This paper raises the question towhat extent are (i) land use changes and/or (ii) check-dams responsible for the decreased sediment yield atthe catchment scale. The main objective is to determinethe effectiveness of land use changes and check-damsas management options to control sediment yield at thecatchment scale.

STUDY AREA

The Rogativa catchment (Murcia, SE Spain, 38° 080 N,2° 130 W), with a size of 47Ð2 km2 was selected as thestudy area. The study area and the land use changesthat have occurred within it since 1956 were describedextensively in Boix-Fayos et al. (2007). The catchmentdrains the northern face of Revolcadores (2027 m) andthe Cuerda de la Gitana (1829 m) mountain ranges inS–N direction. It belongs to the catchment of the Taibilla,a tributary of the Segura river, located at the Subbetic unitof the Betic Mountains. The dominant lithology consistsof marls, limestones, marly limestones and sandstones ofthe Cretaceous, Oligocene, and Miocene (IGME, 1978).Mountains are mainly limestone while mid- and bottom-valley locations are dominated by marls. The averageannual rainfall for the period 1933–2004 was 583 mmand the average annual temperature 13Ð3 °C.

The landscape represents a mix of dryland farming,mainly barley, plantations of walnuts (Junglans regia),forests and shrublands. The catchment has been affectedby important land use changes since the second half ofthe twentieth century. These changes consist mainly ofa progressive abandonment of dryland farming activitiesand an increase in forest cover (Boix-Fayos et al., 2007).The surface area of the land use classes in 1997 is listedin Table I. Table II shows changes in surface area of themost important land uses between different periods.

Forest is dominated by Pinus nigra salzmanii, althoughalso some Pinus pinaster and Pinus halepensis occur inthe lower basin. Quercus rotundifolia was reduced dueto intense wood and charcoal production and clearingfor agriculture in the past. Shrublands located higherwithin the catchment are dominated by Erinacea anthylliswhile at lower altitudes Cytisus reverchonii, Rosmarinusofficinalis, Thymus vulgaris and Genista scorpius appear.

In the catchment area 58 check-dams were constructedin 1976 and 1977 (Figures 1 and 2). Of those, 72%are silted and 81% present erosion features directlydownstream of the dam (Boix-Fayos et al., 2007). Inthe main stream 11 check-dams are located, all of themcompletely silted. Further characteristics of check-damsand their drainage areas are given in Table I.

Copyright 2008 John Wiley & Sons, Ltd. Hydrol. Process. 22, 4922–4935 (2008)DOI: 10.1002/hyp

4924 C. BOIX-FAYOS ET AL.

Tabl

eI.

Loc

atio

nan

dm

ain

char

acte

rist

ics

ofth

ech

eck-

dam

san

dth

eir

drai

nage

area

sof

the

Rog

ativ

aca

tchm

ent.

The

area

ofth

ela

ndus

ecl

asse

sis

deri

ved

from

digi

tal

orth

opho

tos

of19

97

Che

ck-d

amC

heck

-dam

cond

itio

nL

ocat

ion

(rav

ine

nam

e)Y

ear

Sedi

men

tsbu

lkde

nsit

y(g

cm�3

)a

Dra

inag

ear

ea(h

a)

Ori

gina

lst

orag

eca

paci

ty(m

3)

Tra

pef

ficie

ncy

(%)b

Sedi

men

tyi

eld

(tyr

�1)

Hig

han

dm

ediu

mde

nsity

fore

st(h

a)

Low

dens

ity

fore

st(h

a)

Shru

blan

d(h

a)Pa

stur

e(h

a)D

ryla

ndag

ricu

ltur

e(h

a)

D1

silt

edR

oble

1976

1Ð55

15Ð36

676Ð9

987

Ð5530

Ð261Ð0

87Ð3

5—

—6Ð9

3D

2si

lted

Rob

le19

761Ð5

558

Ð1613

6Ð26

27Ð22

28Ð77

15Ð30

32Ð11

——

10Ð71

D3

silt

edU

mbr

ıade

las

vıbo

ras

1976

1Ð55

125Ð4

162

5Ð48

44Ð32

81Ð11

48Ð92

60Ð31

16Ð18

——

D4

non-

silt

edU

mbr

ıade

las

vıbo

ras

1976

1Ð55

25Ð17

404Ð4

971

Ð9521

Ð541Ð8

122

Ð25—

—1Ð1

9D

5no

n-si

lted

Can

tarr

ales

119

761Ð5

524

6Ð57

278Ð9

815

Ð3082

Ð3680

Ð9116

4Ð00

——

1Ð63

D6

non-

silte

dC

anta

rral

es1

1976

1Ð55

5Ð80

229Ð9

786

Ð3513

Ð995Ð1

4—

——

0Ð67

D7

non-

silte

dC

anta

rral

es1

1976

1Ð55

11Ð67

326Ð5

281

Ð719Ð1

111

Ð66—

——

0Ð01

D8

non-

silt

edU

mbr

ıade

las

vıbo

ras

1976

1Ð55

230Ð5

690

3Ð34

38Ð47

99Ð45

73Ð72

56Ð66

58Ð81

9Ð20

32Ð18

D9

silt

ed(C

5cm

)C

ueva

del

Agu

a19

761Ð5

585

Ð1012

61Ð68

70Ð29

104Ð4

963

Ð441Ð0

113

Ð85—

6Ð81

D10

non-

silt

edR

ogat

iva

1976

1Ð55

26Ð91

573Ð4

077

Ð2826

Ð905Ð8

815

Ð905Ð0

7—

0Ð06

D11

non-

silt

edR

ogat

iva

1977

1Ð55

37Ð29

1968

Ð7489

Ð3995

Ð5624

Ð532Ð7

20Ð7

5—

9Ð30

D12

silte

dB

arra

ca19

771Ð5

536

Ð0341

9Ð77

65Ð03

37Ð10

8Ð75

1Ð22

4Ð86

—21

Ð20D

13no

n-si

lted

Can

tarr

ales

219

771Ð5

535

Ð7325

8Ð48

53Ð59

22Ð92

35Ð67

0Ð07

——

—D

14si

lted

Can

tarr

ales

219

761Ð5

013

Ð2570

9Ð34

89Ð53

43Ð90

11Ð00

——

—2Ð2

5D

15no

n-si

lted

Poci

cos

1976

1Ð50

86Ð98

238Ð2

530

Ð4221

Ð9224

Ð7753

Ð398Ð8

2—

—D

16si

lted

(C20

.3cm

)So

lana

1977

1Ð52

4Ð73

1070

Ð2297

Ð3166

Ð650Ð0

8—

——

4Ð64

D17

silt

edSo

lana

1977

1Ð52

155Ð6

512

45Ð77

56Ð09

125Ð0

212

9Ð59

1Ð67

——

24Ð40

D18

silt

edR

ogat

iva

1977

1Ð55

42Ð31

8246

Ð1096

Ð8948

9Ð14

5Ð19

——

—37

Ð12D

19si

lted

(C10

cm)

Java

nas

1977

1Ð52

20Ð25

746Ð4

885

Ð4750

Ð7012

Ð91—

1Ð79

—5Ð5

5D

20si

lted

Est

ebas

1977

1Ð52

181Ð3

113

2Ð75

10Ð46

71Ð42

51Ð14

62Ð62

4Ð88

46Ð81

15Ð88

D21

non-

silte

dJa

vana

s19

771Ð5

211

Ð4278

9Ð42

91Ð69

42Ð81

3Ð30

—7Ð4

50Ð0

20Ð6

5D

22si

lted

(C20

cm)

Java

nas

1977

1Ð52

66Ð72

92Ð24

18Ð08

32Ð31

39Ð89

9Ð37

7Ð68

9Ð79

—D

23si

lted

Lom

aPa

rril

la19

771Ð5

244

0Ð05

1058

Ð9927

Ð7521

4Ð81

245Ð6

512

0Ð70

8Ð07

50Ð99

14Ð65

D24

non-

silt

edL

oma

Parr

illa

1977

1Ð52

17Ð08

703Ð8

986

Ð8134

Ð4215

Ð37—

1Ð65

——

D25

silt

edR

ogat

iva

1977

1Ð55

85Ð47

80Ð39

13Ð05

35Ð40

31Ð48

——

—54

Ð00D

26si

lted

Rog

ativ

a19

771Ð5

514

7Ð72

3058

Ð1876

Ð7722

8Ð95

45Ð90

—20

Ð633Ð2

577

Ð91D

27si

lted

(C20

cm)

Est

ebas

1977

1Ð50

41Ð31

47Ð94

15Ð63

19Ð12

14Ð17

—5Ð8

7—

21Ð26

D28

silt

ed(C

10cm

)A

lmen

icas

1977

1Ð50

361Ð7

743

2Ð86

16Ð03

154Ð5

826

3Ð91

25Ð67

61Ð12

9Ð17

1Ð90

D29

non-

silt

edE

scri

bano

1977

1Ð50

7Ð92

2350

Ð1097

Ð9312

6Ð33

4Ð61

——

—3Ð3

1



Tabl

eI.

(Con

tinu

ed)

Che

ck-d

amC

heck

-dam

cond

itio

nL

ocat

ion

(rav

ine

nam

e)Y

ear

Sedi

men

tsbu

lkde

nsit

y(g

cm�3

)a

Dra

inag

ear

ea(h

a)

Ori

gina

lst

orag

eca

paci

ty(m

3)

Tra

pef

ficie

ncy

(%)b

Sedi

men

tyi

eld

(tyr

�1)

Hig

han

dm

ediu

mde

nsity

fore

st(h

a)

Low

dens

ity

fore

st(h

a)

Shru

blan

d(h

a)Pa

stur

e(h

a)D

ryla

ndag

ricu

ltur

e(h

a)

D30

silt

edE

scri

bano

1977

1Ð50

231Ð1

354

6Ð00

27Ð38

110Ð5

119

3Ð34

—8Ð3

1—

29Ð48

D31

silt

edE

scri

bano

1977

1Ð50

5Ð49

750Ð7

595

Ð6243

Ð514Ð7

8—

——

0Ð71

D32

silt

ed(C

25cm

)E

scri

bano

1977

1Ð50

7Ð65

1539

Ð9996

Ð9895

Ð335Ð0

7—

——

2Ð58

D33

silt

edC

irue

los

1977

1Ð50

61Ð29

1529

Ð8279

Ð9310

6Ð06

60Ð38

——

—0Ð9

4D

34si

lted

(C40

cm)

Cir

uelo

s19

771Ð5

014

Ð7515

28Ð65

94Ð30

99Ð28

12Ð32

——

—2Ð4

1D

35si

lted

Cir

uelo

s19

771Ð5

019

Ð9413

78Ð72

91Ð69

83Ð33

11Ð77

0Ð34

0Ð57

—7Ð3

8D

36si

lted

(C50

cm)

Cir

uelo

s19

771Ð5

01Ð0

717

07Ð89

99Ð61

110Ð8

50Ð5

10Ð1

6—

0Ð40

D37

silt

edC

irue

los

1999

1Ð50

71Ð13

2506

Ð8384

Ð9116

3Ð61

43Ð35

—22

Ð92—

4Ð80

D38

silt

edR

ogat

iva

1977

1Ð55

238Ð3

610

841Ð1

487

Ð8970

8Ð88

128Ð3

233

Ð2211

Ð241Ð5

764

Ð02D

39si

lted

Cem

ente

rio

1977

1Ð50

9Ð95

3019

Ð7297

Ð9817

0Ð79

8Ð69

——

—1Ð2

6D

40si

lted

Cem

ente

rio

1977

1Ð50

5Ð11

1005

Ð7496

Ð9257

Ð515Ð1

1—

——

—D

41si

lted

(C30

cm)

Cem

ente

rio

1977

1Ð50

156Ð5

292

5Ð49

48Ð55

116Ð2

068

Ð3388

Ð19—

——

D42

silte

dC

emen

teri

o19

771Ð5

07Ð0

244

9Ð36

91Ð09

27Ð34

3Ð62

—0Ð0

2—

3Ð39

D43

silt

edSa

lval

ejo

1977

1Ð50

36Ð44

931Ð8

180

Ð3264

Ð2915

Ð0219

Ð200Ð0

3—

2Ð19

D44

silt

ed(C

20cm

)Sa

lval

ejo

1977

1Ð50

25Ð79

423Ð6

772

Ð3934

Ð8311

Ð2114

Ð58—

——

D45

non-

silt

edSa

lval

ejo

1977

1Ð50

35Ð91

315Ð1

458

Ð3415

Ð5210

Ð9924

Ð92—

——

D46

non-

silt

edSa

lval

ejo

1977

1Ð50

15Ð18

175Ð2

564

Ð825Ð9

99Ð6

05Ð5

8—

——

D47

non-

silt

edSa

lval

ejo

1977

1Ð50

18Ð58

382Ð1

776

Ð6513

Ð4013

Ð345Ð2

4—

——

D48

non-

silt

edSa

lval

ejo

1977

1Ð55

352Ð4

229

5Ð53

11Ð80

85Ð27

208Ð1

157

Ð6111

Ð5275

Ð19—

D49

silt

ed(C

20cm

)C

asa

Nue

va19

771Ð5

312

3Ð67

1038

Ð6857

Ð2710

9Ð41

0Ð00

86Ð36

3Ð04

14Ð51

19Ð75

D50

silt

ed(C

10cm

)R

ogat

iva

1977

1Ð52

308Ð2

595

89Ð21

83Ð24

670Ð8

411

6Ð29

111Ð7

97Ð9

713

Ð2359

Ð05D

51no

n-si

lted

Suer

teE

stre

cha

1977

1Ð37

48Ð70

745Ð6

970

Ð9645

Ð7025

Ð0523

Ð64—

——

D52

silt

edSu

erte

Est

rech

a19

771Ð3

731

Ð8039

2Ð45

66Ð33

29Ð93

20Ð22

7Ð35

4Ð23

——

D53

silt

edE

rmit

a19

771Ð3

79Ð3

812

24Ð13

95Ð42

64Ð90

8Ð26

1Ð07

——

—D

54si

lted

Erm

ita

1977

1Ð37

30Ð55

662Ð8

577

Ð5943

Ð2230

Ð210Ð4

1—

——

D55

silt

ed(C

10cm

)E

rmit

a19

771Ð3

711

Ð1810

25Ð64

93Ð61

57Ð01

10Ð95

0Ð13

0Ð11

——

D56

silt

ed(C

50cm

)R

uico

1977

1Ð37

168Ð5

225

06Ð00

70Ð36

202Ð7

210

6Ð62

34Ð95

—26

Ð95—

D57

silt

ed(C

10cm

)R

uico

1977

1Ð37

6Ð81

1170

Ð2396

Ð4863

Ð406Ð7

70Ð0

4—

——

D58

silt

ed(C

20cm

)R

ogat

iva

1978

1Ð52

146Ð9

615

142Ð3

894

Ð2797

1Ð16

72Ð67

57Ð95

9Ð87

2Ð39

4Ð09

aA

vera

gebu

lkde

nsity

dete

rmin

edfo

rsa

mpl

esta

ken

inth

eal

luvi

alw

edge

behi

ndea

chda

mb

Tra

pef

ficie

ncy

ofth

eor

igin

alca

paci

tyof

the

chec

k-da

ms

calc

ulat

edac

cord

ing

toB

row

n(1

943)



Table II. Area covered by different land uses in 1956, 1981 and 1997 and ratios between years

Areas Ratios

1956 km2 1981 km2 1997 km2 1981/1956 1997/1981 1997/1956

High density forest 1Ð98 6Ð91 9Ð51 3Ð48 1Ð38 4Ð80Medium density forest 7Ð09 10Ð07 16Ð52 1Ð42 1Ð64 2Ð33Low density forest 15Ð34 8Ð36 10Ð10 0Ð55 1Ð21 0Ð66Shrubland 6Ð00 9Ð92 2Ð47 1Ð65 0Ð25 0Ð41Pasture land 4Ð04 2Ð13 2Ð73 0Ð53 1Ð28 0Ð67Dry land agriculture 12Ð07 9Ð15 5Ð21 0Ð76 0Ð57 0Ð43

Figure 1. Location map of the study area, check-dams and sampling points within the Rogativa catchment

METHODS

The methodological approach to this work consisted ofthree main parts: (i) field surveys in order to estimatesediment yields at the subcatchment level and to charac-terize the soils within the catchment; (ii) GIS analysis tocalculate various model input parameters; and (iii) a mod-elling exercise using the existing spatially distributed soil

erosion model WATEM-SEDEM (Van Oost et al., 2000;Van Rompaey et al., 2001; Verstraeten et al., 2002).

(i) A first field survey to locate all the check-dams con-structed within the catchment was carried out. 58check-dams were found and georeferenced with aGPS Trimble GeoXM with differential correction.



Figure 2. (A) General view of the catchment in November 2004; (B) check-dam in a tributary river at the medium catchment in May 2004;(C) check-dam in a tributary river at the upper catchment in June 2004; (D) check-dam at the outlet of the catchment in May 2005

Measurements of the dam structure (height, length),and detailed mapping of the sediment wedges behindeach dam was carried out. The volume of sedimentswas estimated assuming that the alluvial wedge hasthe form of a prismatic channel with a rectangu-lar section (Prosser and Karssies, 2001; Lien, 2003;Castillo et al., 2007), measuring in the field theheight and area of the wedges. Undisturbed samplingof a selection of sediment wedges was carried out toestimate the average bulk densities of retained sedi-ments. Seven sediment wedges were sampled at thefront and the end of the wedge (two sampling areasper wedge) (Figure 1). Bulk samples of 100 cm3

at 7 cm depth intervals were taken to a maximumdepth of 1Ð25 m, and two replicates also with 7 cmdepth intervals to 35 cm depth were taken at eachsampling area. A total of 189 undisturbed sampleswere collected. With these data, the mass of sedi-ment retained by each check-dam was estimated. Ina second field survey the soil texture for differentareas within the catchment was determined. A totalof 70 sample locations distributed from upstream todownstream in the catchment were sampled at 0–5and 5–10 cm depth. Soils were air dried; the organicmatter was eliminated with H2O2, the samples werechemically dispersed with hexametaphosphate andtheir particle size distribution characterized by laserdiffraction using a Coulter LS200. Three laboratoryreplicates for each sample were done. The percent-age of coarse and fine sand, coarse and fine silt and

clay were obtained as well as the geometric meansize of particles. Results were used to obtain a soilerodibility map (K factor) as explained below.

(ii) GIS analysis was used to derive various modelinput parameters. First of all, the catchment areadraining to each check-dam was calculated using adigital elevation model (DEM) of the catchment. TheDEM was extracted from the contour lines (10 minterval) of digital topographic maps obtained fromthe Spanish National Geographic Institute (IGN). Themaps were published between 2000 and 2002 andbased on photogrammetric information of a flight of1988. From these maps a DEM was created at 30 mresolution using the Idrisi software. Trap efficiencyfor each check-dam was calculated following Brown(1943) using the initial volume of each check-dam. This method estimates trap efficiency accordingto the structure capacity and watershed area ratioand it is useful for estimations of mid- and long-term trap efficiencies, especially when no inflowdata are available (Verstraeten and Poesen, 2000).Specific sediment yield data (SSY, t ha�1 yr�1) andabsolute sediment yield (SY, t yr�1) were calculatedbased on the volume of sediments retained by eachcheck-dam, the bulk density of sediments, the trapefficiency of check-dams and the drainage area ofeach check-dam. Similar approaches are reportedby several authors to estimate sediment yield databehind small dams or check-dams (Verstraeten et al.,2007; Romero-Dıaz et al., 2007a).



Furthermore, several GIS layers were prepared as inputfor the erosion model RUSLE (see third paragraph of thissection) in the following way:

1. The land use scenarios of three years (1956, 1981 and1997) derived from interpretation of aerial photographs(spatial resolution 1 m) of those years were adapted insize and resolution to the study area. Details of themethods to derive the land use scenarios can be foundin Boix-Fayos et al. (2007).

2. Crop factor (C factor) map. C factors were assignedto land use classes based on the estimations of theNational Inventory of Soils for the Region of Murcia(DGCONA, 2002).

3. Rainfall factor (R factor) map. This factor was cal-culated based on mean monthly rainfall data ofthe National Meteorological Institute of Spain forthe period 1937–2004. The equation of Renard andFreimund (1994) as explained in de Vente et al.(2007b) was applied to calculate R.

4. Erodibility (K factor) map. A K factor for each soilsampling point was estimated using the algorithmproposed by Romkens et al. (1987). The data wereextrapolated for the whole catchment using a krigingprocedure. For very steep slopes (>35°), the K factorwas set at a value of 20.

5. Rivers map6. Pond map with a specific trap efficiency7. Parcels map

All layers were resampled at 30 m resolution for themodel application.

GIS analysis was also used to characterize the landuse pattern for 1956, 1981 and 1997 by the applicationof the perimeter to area ratio and the fragmentation index(Monmomier, 1974, Eq. (1)):

F D(

m∑iD1

(c � 1

n � 1

)/n

)�1�

where n D number of different classes present in thekernel of 5 by 5 pixels, c D number of cells consideredin the 5 by 5 pixels kernel and m D number of pixels inthe image. A high perimeter to area ratio and a highfragmentation index indicate a more fragmented landuse pattern with more small isolated patches of differentland use.

Since the available land use maps from aerial pho-tographs (1956, 1981, 1997) did not provide an imageof the land use at the moment of the installation of thehydrological control works (i.e. starting 1976), a classifi-cation was performed of a historic Landsat MultispectralScanner (MSS) satellite image of 21 January 1974 in aneffort to characterize land use conditions for this timeperiod. The MSS sensor took images in four bands ofthe electromagnetic spectrum: two in the visible partof the spectrum, and two in the infrared part. A falsecolour image based on the combination of bands 4, 2, 1was used to perform a supervised classification of land

uses. Because of the relatively low level of detail in theLandsat MSS images (spatial resolution 75 m), in thisclassification two of the original land use classes weremerged (i.e. high and medium density forest). Becauseof the lower spatial resolution of the satellite image com-pared with the aerial photographs, the satellite image wasonly used to characterize the land use at the moment ofconstruction of the control works and not as a land usescenario for modelling purposes.

(iii) The spatially distributed model WATEM-SEDEM (VanOost et al., 2000; Van Rompaey et al., 2001; Ver-straeten et al., 2002) was used to estimate the sedimentyield at the outlet of the catchment under different landuse scenarios. WATEM-SEDEM provides long-termmean annual soil erosion rates by water and sedimentyield. The model was extensively described in earlierpublications (e.g. Van Oost et al., 2000, Van Rompaeyet al., 2001, Verstraeten et al., 2002, de Vente et al.,2007b), so here we summarize the most important fea-tures. The model calculates on a pixel basis and hasthree main components: soil erosion assessment, sedi-ment transport capacity and sediment routing. WithinWATEM-SEDEM mean annual soil erosion is pre-dicted with a modified version of the revised universalsoil loss equation (RUSLE; Renard et al., 1997) thatwas proposed by Desmet and Govers (1996a, b) as:

E D R ð K ð LS2D ð C ð P �2�

where E is the mean annual soil erosion (kg m�2 yr�1),R is the rainfall erosivity factor (MJ mm m�2 h�1 yr�1),K is the soil erodibility factor (kg h MJ�1 mm�1),LS2D is the two-dimensional topographic factor, C isthe crop and management factor, and P is the erosioncontrol practice factor. In the original WATEM-SEDEMmodel the sediment transport capacity is calculated asthe product of the rill and interrill erosion potential anda constant factor called the transport capacity coefficient.Here we used an adapted formula to calculate transportcapacity as suggested by Verstraeten et al. (2007):

TC D KTC ð R ð K ð A1Ð4 ð S1Ð4 �3�

where, TC is the transport capacity (kg m�2 yr�1), KTCis the transport capacity coefficient (�), R and K are therainfall intensity and soil erodibility factor of the RUSLE,A is the upslope area (m2), and S the local slope gradient(m m�1). The difference between this equation and theoriginal formulation of sediment transport capacity is thatit allows a high transport capacity throughout zero-orderbasins, which is essential in basins where gully erosionis an important erosion process (de Vente et al., 2007b,Verstraeten et al., 2007).

Once soil erosion and sediment transport capacity areknown for each pixel, sediments are routed through thebasin towards the river along a runoff pattern that iscalculated with a multiple-flow algorithm (Desmet andGovers, 1996a). The location of the check-dams withinthe catchment was introduced in the model as sedimen-tation areas within the channels using the calculated trap



Figure 3. Land use maps 1956, 1981, 1997 (modified from Boix-Fayos et al., 2007) derived from air photographs

Table III. Area covered by different land uses in 1956, 1974, 1981 and 1997 (data from different sources: 1956, 1981 and 1997 fromdigital air photographs, 1974 from Landsat MSS satellite image)

1956 km2 1974 km2 1981 km2 1997 km2

Forest (total cover) 24Ð42 25Ð18 25Ð34 36Ð13Forest (high C medium density) 9Ð08 16Ð20 16Ð98 26Ð03Low density forest 15Ð34 8Ð98 8Ð36 10Ð10Shrubland 6Ð00 7Ð91 9Ð92 2Ð47Pasture land 4Ð04 0Ð90 2Ð13 2Ð73Dry land agriculture 12Ð07 12Ð48 9Ð15 5Ð21

Table IV. Change in landscape metric indicators between the studied periods

Perimeter to Area ratio (km km�2) Fragmentation index

1956 1981 1997 1956 1981 1997

High density forest 25Ð09 11Ð93 4Ð54 0Ð055 0Ð030 0Ð035Medium density forest 15Ð94 7Ð94 3Ð71 0Ð038 0Ð020 0Ð024Low density forest 9Ð16 7Ð05 9Ð17 0Ð024 0Ð018 0Ð023Shubland 14Ð73 11Ð43 18Ð41 0Ð037 0Ð027 0Ð041Pasture land 15Ð06 14Ð05 13Ð03 0Ð037 0Ð036 0Ð032Dry land agriculture 9Ð47 10Ð75 12Ð93 0Ð023 0Ð025 0Ð029Mean values 14Ð91 10Ð53 10Ð30 0Ð036 0Ð026 0Ð031

efficiency for each dam. In this way calculation of sed-iment yields for each river segment upstream of eachcheck-dam was carried out.

RESULTS AND DISCUSSION

Land use scenarios and sediment yield dataLarge differences in land use occurred between 1956

and 1997. The area occupied by high density forestincreased 4Ð8-fold, the area covered by medium densityforest increased 2Ð3-fold and the area dedicated to agri-culture decreased by 57% (Table II, Figure 3). By 1981

many important land use changes had already occurred,the high density forest had already increased 3Ð5-fold, themedium density forest had increased 1Ð42-fold and theagricultural land had been reduced by 24% by that time.Also remarkable is that the area covered by shrublandhad increased in 1981 1Ð65-fold.

The land use map derived from the satellite imageof 1974 offers an image of the situation just before thestart of the hydrological correction works (reforestationand check-dam construction) in 1976. Table III offers acomparison of the land use areas at the four dates afteraggregation of the classes, high- and medium-density



forest. However, the 1974 land use map is difficult tocompare with those of 1956, 1981 and 1997 becauseof the different spatial resolution of the remote sensingsources (see methods section). In the classification of the1974 satellite image the class ‘pasture land’ producedconfusion with the class ‘dryland agriculture’, and as aresult the pasture land class is probably underestimated.Nevertheless, the different land use classes in 1974occupy areas of very similar size to those in 1981,which means that important changes in forest cover hadalready occurred before the initiation of the hydrologicalcorrection works of 1976. The total forest cover in1974 had approximately the same extension as in 1956(24Ð42 km2 and 25Ð18 km2, respectively) but a higherdensity of forest in 1974 can be observed through theincrease of the area with high- and medium- density forest(9Ð08 km2 and 16Ð20 km2 in 1956 and 1974, respectively)and a decrease in the low-density forest. On the otherhand, the area of agricultural land in 1974 had notdecreased with respect to 1956 (Table III).

The spatial pattern of land use also experiencedimportant changes during the whole study period, asindicated by the fragmentation index and the perimeterto area ratio (Table IV). The perimeter to area ratio isthe highest for high density forest and medium densityforest and the lowest for agricultural land in 1956. Thefragmentation index is the highest for high-density forestand the lowest for agricultural land in 1956, while in1997 the fragmentation index was still high for high-density forest but had considerably decreased with respectto 1956. For agricultural land the fragmentation indexhad slightly increased in 1997 with respect to 1956.All these indicators show that agricultural land had lowfragmentation and was well connected in 1956. Howeverthe land use mosaic in 1997 shows a decreased areaof agricultural land with higher fragmentation than in1956. These results are comparable with those obtainedby Vanacker et al. (2005) in a comparison of land usechanges between 1963 and 1995 in the Deleg catchmentin Ecuador, where the barren land was characterizedby a strong increase in fragmentation during the studyperiod.

Absolute sediment yield (Table I) and specific sed-iment yield for the subcatchments (Figure 4a) of all58 check-dams within the study area was estimatedbased on the measured sediment volumes trapped behindthe check-dams. The specific sediment yield valuesvary between 0Ð25 t ha�1 yr�1 and 107Ð33 t ha�1 yr�1

for subcatchments of 352Ð4 ha and 1Ð1 ha, respectively(Figure 4b). Specific sediment yield values are drasticallyreduced at catchment areas larger than 40 ha; above thisthreshold 85% of the subcatchments show specific sedi-ment yield values lower than 2 t ha�1 yr�1 (see Table Ifor catchment areas and Figure 4a for specific sedimentyield, subcatchments with a drainage area larger than40 ha belong to check-dams 2, 3, 5, 8, 9, 15, 17, 18,20, 22, 23, 25, 26, 27, 28, 30, 33, 37, 38, 41, 48, 49, 50,51, 56, 58). A large variability in specific sediment yieldwas observed and in general, sediment yield decreased

with increasing catchment area (Figure 4b). This situationslightly differs from the model proposed by de Vente andPoesen (2005), where an increase in specific sedimentyield values was described in subcatchments until 1 km2,and a decreasing tendency above this threshold.

The check-dams with catchment areas >40 ha and withSSY > 2 t ha�1 yr�1 are marked in Figure 4b. Thesecorrespond to the subcatchments of check-dams locatedin the main channel of the Rogativa catchment. InFigure 4c the specific sediment yield estimated fromcheck-dams located in the main channel of the Rogativaare plotted from upstream to downstream. There is aslight increase in SSY in the downstream direction; itis possible that this can be caused because other erosionprocesses, such as channel erosion and bank erosion froma certain threshold, are introducing large sediment yieldsin the main channel, as has been observed for other cases(de Vente et al., 2007a).

Model calibration

The WATEM-SEDEM model was calibrated for theentire study catchment using the 1981 land use map asthe land use scenario and including within the modelthe position of the check-dams. The land use of 1981was used for calibration because it was consideredthe most representative of the land use pattern at thetime of construction of check-dams (1976–1977), andfurthermore it has a high level of detail (i.e. more thanthe 1974 map). Only the subcatchments with non-siltedcheck-dams were used for the calibration procedure,since sediment yield of subcatchments of silted dams isless accurate because it is not known when the damswere silted and so no accurate sediment yield can becalculated. The limitation of this approach is obviouslythat the non-silted check-dams are probably located inthe less erodible subcatchments, thus an underestimationof sediment yield for the whole catchment is expected inthe modelling exercise. However, given that the objectiveof the paper is to compare the relative impact of land usescenarios and check-dams on the total sediment yield,the suggested approach is considered valid for relativecomparisons.

The WATEM-SEDEM model was run with a widerange of transport capacity coefficients (KTc). The KTcvalues in Equation (3) were assigned for two contrastingland cover categories, reflecting their different sensitivityto overland flow sediment transport. For well-vegetatedsurfaces (i.e. natural vegetation classes) a low KTc valuewas used (KTc Low), and for poorly-vegetated surfaces(i.e. dry land agriculture) a high KTc value was applied(KTc High). For each parameter combination (i.e. KTcLow and KTc High), the absolute sediment yield (SY ; tyr�1) was calculated for each catchment. The optimal KTcvalues were selected according to the model efficiency



Figure 4. (a) Specific sediment yield estimated at the 58 subcatchments of the check-dams. The error bars represent the change in sediment yieldestimated with the loss of trap efficiency of the check-dams since the moment of their construction (1976–1977), until the moment of sedimentvolume estimation (2003). The change in sediment yield due to the loss of trap efficiency is only represented for the non-silted check-dams, thesilted check-dams have lost their trap efficiency completely and are supposed to let pass through all the sediment produced (Ł107Ð33 t ha�1 yr�1).(b) Specific sediment yield versus drainage area at the check-dams subcatchments with indication of the relation with the Taibilla reservoir. (c) Specific

sediment yield in the check-dams located along the main stream of the Rogativa catchment

(ME ) described by Nash and Sutcliffe (1970):

ME D 1 �

n∑iD1

�Oi � Pi�2

n∑iD1

�Oi � Omean�2

�4�

where n is the number of observations, Oi is the observedvalue, Omean is the mean observed value and Pi is thepredicted value. The closer the ME value approaches 1,the more efficient the model is. It was decided to calibrate

on absolute sediment yield (t yr�1) instead of relativesediment yield (t km�2 yr�1) since the objective of thestudy was to assess the effect of land use changes and theconstruction of check dams on absolute sediment outputfrom the catchment.

The optimal KTc values were 10�6 for KTc Low and3 ð 10�5 for KTc High, showing a model efficiencyof 0Ð84 (Figure 6). The model efficiency obtained inthis application falls within the range of model efficien-cies obtained in previous applications of the WATEM-SEDEM. Model efficiencies reported oscillate between0Ð14 and 0Ð89 for applications of WATEM-SEDEM in



Figure 5. Land use classification extracted from the Landsat MSS imageof January 1974

Italy, Belgium, Spain, Australia and Czech Republic (VanRompaey et al., 2005; Verstraeten et al., 2006; de Venteet al., 2007b; Verstraeten et al., 2007 and Van Rompaeyet al., 2007, respectively).

Influence of land use changes and check-damson sediment yield

After calibration, the WATEM-SEDEM model was runwith six scenarios. First, modelling the effect of differentland use patterns (scenarios with land use of 1956, landuse of 1981 and land use of 1997) and check-dams(original scenario with land use of 1956 and check-dams)on sediment yield separately; and the combined effectof land use patterns and check-dams (scenario with landuse of 1981 and check-dams; scenario with land use of1997 and check-dams). Table V shows the main results ofthe modelling exercise. The results of the model appliedto the land use scenarios without check-dams show aprogressive decrease in sediment yield from 1956 to1997, which can probably be explained by the increasingforest cover and decreasing dryland agriculture from 1956to 1997 (Table II). With the 1981 landuse scenario a 44%decrease of sediment yield with respect to 1956 appears.In 1997 a 54% less sediment yield appears with respectto 1956. Notice that between 1981 and 1997 sedimentyield was reduced only 10%, whereas a decrease of 9%in agricultural areas and an increase of 18% in forestcover was observed.

Figure 6. Calibration curves using the land use scenario of 1981 and thenon-silted check-dams

Table V. Results of the modeled sediment yield at the catch-ment scale under different land use scenarios with and without

check-dams

Land usescenarios

1956 1981 1997

Sediment yieldWithout check-dams 1123 t 632 t 519 tWith check-dams 257 t 19 t 137 t

Decreased sediment yieldWithout check-dams — �44 % �54 %With check-dams �77 % �98 % �88 %

Sediment retained behindcheck-dams

866 t 613 t 382 t

Retention of the total producedsediment

77% 97% 74%

With a scenario without land use change (maintainingthe 1956 landuse scenario) but with the construction ofcheck-dams, there is a decrease of 77% sediment yield. Ina scenario of land use change and check-dams in 1981,there is a decrease of 98% of sediment yield. With ascenario of land use changes and check-dams in 1997,the sediment yield was reduced by 88%. The 58 check-dams network added a 54% reduction in sediment yield



to the land use scenario of 1981, and a 34% reduction insediment yield to the land use scenario of 1997.

The greatest control of sediment yield in the long termis made by land use; in 1981 sediment decreased 44%but most of it went to the dams (97%); in 1997 sedimentyield decreased 54%, 10% more, but only 74% went tothe dams (Table V). The most likely explanation for thisis that there were no dams close to the most importantsource areas of sediment in the land use scenario of 1997.This stresses the importance of the landscape pattern andconnectivity for the prediction of sediment yield, as wasearlier mentioned by Van Oost et al. (2000) and Bakkeret al. (2008).

The application of one type or another of sediment con-trol measures (e.g. reforestation or check dam construc-tion), or a combination of them, depends on the objectiveof the management and on the environmental conditionsof each specific area. For example, when the objective isto drastically reduce sediment delivery in badland areaswith high difficulties for the establishment of vegetation,it will probably be more efficient to combine check-dams with adapted revegetation strategies for the area.A recent proposal focuses on re-vegetation strategies thatreduce the connectivity of overland flow, after evaluationof past reforestation experiences that induced landscapedegradation (Recondes team, 2007). It must be taken intoaccount that check-dams induce erosion processes down-stream (Boix-Fayos et al., 2007; Castillo et al., 2007).Romero-Dıaz et al. (2007b) estimated that up to 20%of the sediments retained by the dams could be inducedby the construction of upstream check-dams and accesstracks. In general check-dams are a temporary solutionbecause they often fill up rapidly.

In contrast, in areas without urgent problems relatedto high sediment delivery from rivers and streams,sufficiently deep soils and favourable climatologicalconditions for the establishment of shrubland and forest,revegetation is a suitable and sustainable solution toreduce catchment sediment yield. In the case of theRogativa catchment, the environmental conditions causeda very positive evolution of the forest cover detectedbetween 1956 and 1974, just before the reforestationand hydrological control works. The model applicationshowed an important control of land use on sedimentyield in this area. Sediment yield was reduced by up to54% only by changing land use conditions. Therefore, itis likely that the natural recovery of the forest togetherwith the progressive abandonment of agriculture wouldhave led to an important decrease of sediment yieldbefore the construction of the check-dams. This makesthe relevance of the construction of the check-damsas an additional measure in the Rogativa catchmentquestionable.

CONCLUSIONS

Both land use changes and check-dams are effective mea-sures decreasing sediment yield in catchments, however

they act at very different temporal scales. Check-dams arevery effective in the short term but potentially increaseerosion downstream. In the studied catchment the effectof land use changes (mainly decrease of agriculturalactivities and increase of forest cover) decreased the sed-iment yield by 44% in 25 years and 54% in 40 years.The addition of 58 check-dams to the land use pattern of1997 meant an extra decrease of 34% in sediment yield.The construction of check-dams without land use changeswith respect to the 1956 scenario controlled 77% of thesediment yield in the study area. In the case of the Rog-ativa catchment it seems that the change in the spatialpattern of land use at the moment of the construction ofcheck-dams could have reduced sediment yield alreadyby 40%.

Land use changes are long-term sustained sedimentcontrol measures compared with check-dams, which areshort-term effective sediment control measures. The useof each of them should be conditioned by the environ-mental characteristics of the area and the specific objec-tives of the management project. In high erodibility areaswith problems for vegetation establishment, check-damscan be effective for reducing sediment yield. In areaswith favourable conditions for vegetation establishment,land use changes leading to an increased vegetation coverare sustainable measures to reduce sediment yield, andcheck-dams can be confined to important source areas ofsediment.

ACKNOWLEDGMENTS

This work was financially supported by a researchproject of the Fundacion Instituto Euromediterraneo delAgua (CE) and by the Spanish Ministry of Scienceand Education through the projects CANOA CGL2004-04 949-C02-01; PROBASE CGL2006-11 619 and ERCOCGL2007-62 590/BTE. The first author was financiallysupported by an I3P postdoc contract CSIC-CE and a“Ramon y Cajal” contract from the Spanish Ministry ofScience and Education. Gonzalo G. Barbera also wassupported by I3P postdoc contract. Joris de Vente wassupported by a Seneca (Region of Murcia Government)foreign researchers grant and through a contract in theintegrated project DESIRE of the EU FP6 037 046(GOCE).

David Correa Calvo helped in the field work. TheDireccion General del Medio Natural of the Region ofMurcia (Directorate of Environment) is acknowledged forproviding the digital ortophotos. The Mancomunidad deCanales del Taibilla (Community of channels of Taibillariver) is acknowledged for the interest shown in thisresearch.

REFERENCES

Alcoverro J, Corominas J, Gomez M. 1999. The Barranco de Aras floodof 7 August 1996 Biescas, central Pyrenees, Spain. EngineeringGeology 51: 237–255.



Bakker MM, Govers G, van Doorn A, Quetier F, Chouvardas D,Rounsevell M. 2008. The response of soil erosion and sediment exportto land use change in four areas of Europe: the importance of landscapepattern. Geomorphology 98: 213–216.

Begueria S, Lopez Moreno J, Gomez Villar A, Rubio V, Lana-Renault N, Garcıa-Ruiz J. 2006. Fluvial adjustments to soil erosionand plant cover changes in the Central Spanish Pyrenees. GeografiskaAnnaler, Series A: Physical Geography 88: 177–186.

Boix-Fayos C, Barbera GG, Lopez-Bermudez F, Castillo VM. 2007.Effects of check-dams, reforestation and land-use changes on riverchannel morphology: Case study of the Rogativa catchment (Murcia,Spain). Geomorphology 91: 103–123.

Bombino G, Tamburino V, Zimbone SM. 2006. Assessment of theeffects of check-dams on riparian vegetation in the mediterraneanenvironment: A methodological approach and example application.Ecological Engineering 27: 134–144.

Brown CB. 1943. Discussion of Sedimentation in reservoirs, byJ. Witzig. Proceedings of the American Society of Civil Engineers 69:1493–1500.

Castillo VM, Mosch WM, Conesa Garcıa C, Barbera GG, NavarroCano JA, Lopez-Bermudez F. 2007. Effectiveness and geomorpholog-ical impacts of check dams for soil erosion control for in a semiaridMediterranean catchment: El Carcavo (Murcia, Spain). Catena 70:416–427.

Conesa Garcıa C, Belmonte Serrato F, Garcıa Lorenzo R. 2004. “Efectosde los diques de correccion hidrologico-forestal en la competenciay estabilidad de corrientes efımeras. Aplicacion a la cuenca de larambla de la Torrecilla (Murcia)”. In: Conesa Garcıa C, MartınezGuevara JB Territorio y Medio Ambiente: Metodos Cuantitativosy Tecnicas de Informacion Geografica, Asociacion de GeografosEspanoles y Universidad de Murcia: Murcia, pp. 69–83 (in spanish).

Conesa-Garcıa C, Garcıa-Lorenzo R. 2008. Effectiveness of check-damsin the control of general transitory bed scouring in semiarid catchmentareas (South-East Spain). Water and Environment Journal 1–14.

de Vente J, Poesen J. 2005. Predicting soil erosion and sediment yieldat the basin scale: Scale issues and semi-quantitative models. Earth-Science Reviews 71: 95–125.

de Vente J, Poesen J, Arabkhedri M, Verstraeten G. 2007a. The sedimentdelivery problem revisited. Progress in Physical Geography 31:155–178.

de Vente J, Poesen J, Verstraeten G, Van Rompaey A, Govers G. 2007b.Spatially distributed modelling of soil erosion and sediment yield atregional scales in Spain. Global and Planetary Change 60: 393–415.

Desmet PJJ, Govers G. 1996a. A GIS procedure for automaticallycalculating the USLE LS factor on topographically complex landscapeunits. Journal of Soil and Water Conservation 51: 427–433.

Desmet PJJ, Govers G. 1996b. Comparison of routing algorithms fordigital terrain models and their implications for predicting ephemeralgullies. International Journal of Geographic Information Science 10:311–331.

DGCONA, Ministerio de Medio Ambiente. 2002. Inventario Nacional deSuelos 2002–2012 . Region de Murcia: Murcia.

Garcia-Ruiz JM, Lasanta T, Ruiz-Flano P, Ortigosa L, White S. 1996a.Land-use changes and sustainable development in mountain areas: Acase study in the Spanish Pyrenees. Landscape Ecology 11: 267–277.

Garcıa Ruiz JM, White S, Martı C, Valero Paz M, Gomez-Villar A.1996b. The catastrophe of the Barranco de Aras Biescas, AragonPyrenees and its spatial-temporal context . Instituto Pirenaico deEcologıa. CSIC: Zaragoza (in Spanish).

Gomez-Villar A, Martınez-Castroviejo R. 1991. Channel degradation asa response to erosion control works: a case study. In Sala M, Rubio JL,Garcıa-Ruiz JM (eds). Soil Erosion Studies in Spain. GeoformaEdiciones: Logrono, Spain; 109–122.

Gutierrez F, Gutierrez M, Sancho C. 1998. Geomorphological andsedimentological analysis of a catastrophic flash flood in the Arasdrainage basin central Pyrenees, Spain. Geomorphology 22: 265–283.

Instituto Geologico y Minero de Espana (IGME). 1978. Mapa geologicode Espana 1 : 50000, Nerpio 909: 23–26.

Kondolf GM, Piegay H, Landon N. 2002. Channel response to increasedand decreased bedload supply from land use change: contrasts betweentwo catchments. Geomorphology 45: 35–51.

Lien HP. 2003. Design of slit dams for controlling stony debris flow.International Journal of Sediment Research 18: 74–87.

Martın-Rosales W. 2002. Effects of check dams in the southern border ofthe Gador mountain range (Almerıa). Departamento de GeodinamicaUniversidad de Granada, Grupo de Investigacion Recursos Hıdricos yGeologıa Ambiental; 266 pp (in Spanish).

Martin-Rosales W, Pulido-Bosch A, Gisbert J Vallejos A. 2003. Sedi-ment yield estimation and check dams in a semiarid area (Sierra deGador, southern Spain). IAHS-AISH Publication 279: 51–58.

Monmonier MS. 1974. Measures of pattern complexity for choroplethmaps. The American Cartographer 2: 159–169.

Nash JE, Sutcliffe JV. 1970. River flow forecasting through conceptualmodels part I—A discussion of principles. Journal of Hydrology 10:282–290.

Porto P, Gessler J. 1999. Ultimate bed slope in calabrian streamsupstream of check-dams: field study. Journal of Hydraulic Engineering125: 1231–1242.

Prosser I, Karssies L. 2001. Designing filter strips to trap sedimentand attached nutrient. River and riparian land management technicalguideline No 1, Land & Water Australia, Camberra.

RECONDES Team. 2007. Combating Land Degradation by MinimalIntervention: The Connectivity Reduction Approach. The Universityof Portsmouth.

Renard KG, Freidmun JR. 1994. Using monthly precipitation data toestimate the R-factor in the revised USLE. Journal of Hydrology 157:287–306.

Renard KG, Foster GR, Weesies GA, McCool DK, Yoder DC. 1997.Predicting soil erosion by water: a guide to conservation planningwith the revised universal soil loss equation (RUSLE). AgricultureHandbook, 703. US Department of Agriculture: Washington DC.

Romero-Diaz A, Alonso-Sarria F, Martinez-Lloris M. 2007a. Erosionrates obtained from check-dam sedimentation (SE Spain). A multi-method comparison. Catena 71: 172–178.

Romero Dıaz A, Martınez Lloris M, Alonso Sarria F, BelmonteSerrato F, Marın Sanleandro P, Ortiz Silla R, Rodrıguez Estrella T,Sanchez Toribio MI. 2007b. The hydrological correction check-dams:catchment of the Quıpar river (Southeast Spain). Universidad deMurcia, Servicio de Publicaciones (in Spanish).

Romkens M, Prasad J, Poesen J. 1987. Soil erodibility and properties.Transactions of the XIII Congress of International Society of SoilScience V: 492–504.

Simon A, Darby SE. 2002. Effectiveness of grade-control structures inreducing erosion along incised river channels: The case of HotophiaCreek, Mississippi. Geomorphology 42(3–4): 229–254.

Surian N, Rinaldi M. 2003. Morphological response to river engineeringand management in alluvial channels in Italy. Geomorphology 50:307–326.

Van Oost K, Govers G, Desmet PJJ. 2000. Evaluating the effects ofchanges in landscape structure on soil erosion by water and tillage.Landscape Ecology 15: 577–589.

Van Rompaey AJJ, Verstraeten G, Van Oost K, Govers G, Poesen J.2001. Modelling mean annual sediment yield using a distributedapproach. Earth Surface Processes and Landforms 26: 1221–1236.

Van Rompaey AJJ, Govers G, Puttemans C. 2002. Modelling land usechanges and their impacts on soil erosion and sediment supply to rivers.Earth Surface Processes and Landforms 27: 481–494.

Van Rompaey A, Bazzoffi P, Jones RJA, Montanarella L. 2005.Modelling sediment yields in italian catchments 65: 157–169.

Van Rompaey A, Krasa J, Dostal T. 2007. Modelling the impact of landcover changes in Czeh Republic on sediment delivery. Land Use Policy24: 576–583.

Vanacker V, Govers G, Barros S, Poesen J, Deckers J. 2003. The effectof short-term socio-economic and demographic change on landusedynamics and its corresponding geomorphic response with relationto water erosion in a tropical mountainous catchment, Ecuador.Landscape Ecology 18: 1–15.

Vanacker V, Molina A, Govers G, Poesen J, Dercon G, Deckers S. 2005.River channel response to short-term human-induced change inlandscape connectivity in Andean ecosystems. Geomorphology 72:340–353.

Verstraeten G, Poesen J. 2000. Estimating trap efficiency of smallreservoirs and ponds: methods and implications for the assessmentof sediment yield. Progress in Physical Geography 24: 219–251.

Verstraeten G, Van Oost K, Van Rompaey A, Poesen J, Govers G. 2002.Evaluating an integrated approach to catchment management to reducesoil loss and sediment pollution through modelling. Soil Use andManagement 19: 386–394.

Verstraeten G, Poesen J, Gillijns K, Govers G. 2006. The use of riparianvegetated filter strips to reduce river sediment loads: an overestimatedcontrol measure? Hydrological Processes 20: 4259–4267.

Verstraeten G, Prosser IP, Fogarty P. 2007. Predicting the spatial patternsof hillslope sediment delivery to river channels in the Murrumbidgeecatchment, Australia. Journal of Hydrology 334: 440–454.



Verstraeten G, Prosser IP. 2008. Modelling the impact of land use changeand farm dam construction on hillslope sediment delivery on rivers atthe regional scale. Geomorphology 98: 199–212.

Vezina K, Bonn F, Van CP. 2006. Agricultural land-use patterns and soilerosion vulnerability of watershed units in Vietam’s northern highlands.Landscape Ecology 21: 1311–1325.

White S, Garcıa-Ruiz J, Martı C, Valero B, Errea M P, Gomez-Villar A.1997. The 1996 Biescas campsite disaster in the Central Spanish

Pyrenees, and its temporal and spatial context. Hydrological Processes11: 1797–1812.

Wohl E. 2006. Human impacts to mountain streams. Geomorphology 79:217–248.

Xiang-Zhou X. 2004. Development of check-dam systems in gullieson the Loess Plateau, China. Environmental Science & Policy 7:79–86.


the impact of land use change and check-dams on...

Documents