Effects of Land-Cover Change on Soil Erosion in the Saxon …sewolf/PDF/Wolf-Walz-Kleber.2009... · 2009-11-30 · • Land-cover patterns – Erosive slope length – Soil erosion

• Land-cover patterns – Erosive slope length – Soil erosion modelling – RUSLE

DIE ERDE 140 2009 (2) Miscellaneous Issue pp. 155-174

Effects of Land-Cover Change on Soil Erosionin the Saxon Switzerland National Park Region

Sebastian Wolf, Ulrich Walz and Arno Kleber

Land-cover patterns changed significantly during the last century due to anthropogenic impact.This also affected soil erosion rates. Changes in land-cover patterns influence slope lengths and,thus, erosional displacement. However, this effect is not yet well quantified. There are variousestablished approaches for the empirical modelling of soil erosion, but none seems appropriate tosufficiently consider changes in land-cover patterns. Our main objective is to quantify theeffects of land-cover changes on soil erosion. Based on historical maps and physical factors,mainly soil properties and precipitation, our modelling techniques utilise Geographic Informa-tion Systems (GIS) to assess these effects, using the Saxon Switzerland (“Sächsische Schweiz”)National Park Region during the past century as a test area. Our results indicate that changes inland-cover patterns affected soil erosion to a greater extent than estimated so far, and thus needto be considered more precisely in erosion modelling.

With 2 Photos, 6 Figures and 4 Tables

Auswirkungen von Landnutzungsänderungen auf die Bodenerosionin der Nationalparkregion Sächsische Schweiz

1. Introduction

Land use and land cover changed considera-bly during the last centuries (Diemann andArndt 2001; Ramankutty et al. 2006; Turner etal. 2007), with industrialisation and extensionof agricultural land being the major driving fac-tors (IPCC 2000; Foley et al. 2005). Such changes

affect environmental processes such as surfacerunoff (Foley et al. 2005) or regional weatherpatterns (Pielke et al. 2007; Schneider and Eug-ster 2007). These, on their parts, impact soil ero-sion, an anthropogenic process that is causedprimarily by land use (Bork et al. 1998). Assess-ing the impacts of these changes helps quanti-fy soil erosion and mitigate hazardous land-use

156 Sebastian Wolf, Ulrich Walz and Arno Kleber DIE ERDE

and land-cover changes in the future (Lambinand Geist 2006).

Following IPCC (2000) and Lambin and Geist(2006), we define land cover (e.g., farmland) asthe biophysical attributes of surface (e.g.,albedo), and land use as the human use andmanaging of the land cover (e.g., tillage, grazing)(Lambin and Geist 2006). Changes in land useare often, but not necessarily always, associat-ed with changes in land cover (IPCC 2000).

Soil erosion is largely triggered by surface runoffgenerated by intense convective precipitationevents (Schwertmann et al. 1987; Renard et al.1997; Richter 1998). The impacts of land use andland cover on soil erosion have been widely inves-tigated, subjected to modelling techniques (Renard

et al. 1997; Schäuble 1999; Wilson and Lorang 1999;Schmidt 2001; Foster et al. 2003; Evans and Bra-zier 2005; Nearing et al. 2005; Szilassi et al. 2006;Bonilla et al. 2007; Flanagan et al. 2007). Model-ling soil erosion is an alternative to measurementsof sediment yields (Deumlich et al. 2006), to quan-tify the displacement of soil through erosive proc-esses, which may be either wind or water. Herewe focus on water-induced soil erosion as thedominating erosional process in Central Europe(Bork et al. 1998; Schmidt 2001; Boardman 2006)and Germany (Deumlich et al. 2006).

There are two different types of soil erosion mod-els, physical and empirical. Physical, process-based models such as EROSION 2D/3D, the Wa-ter Erosion Prediction Model (WEPP) or the Pre-cision Agricultural-Landscape Modelling System

Fig. 1 Location of the Saxon Switzerland National Park RegionLage der Nationalparkregion Sächsische Schweiz

2009/2 Effects of Land-Cover Change on Soil Erosion 157

(PALMS). They provide accurate erosion esti-mates based on hydrological and physical para-meters (Michael 2000; Schmidt 2001; Michael etal. 2005; Bonilla et al. 2007; Flanagan et al. 2007).Due to their complexity, these models requiremany input parameters with high spatial and tem-poral resolution and are, thus, mainly used forsingle events on a plot scale (Michael 2000;Schmidt 2001; Bonilla et al. 2007).

Empirical models, such as the widely used RevisedUniversal Soil Loss Equation (RUSLE), assesslong-time average soil losses (Renard et al. 1997)and may be used for larger areas up to the regionalscale. Utilising statistically derived functions fromstandardised experimental plots, they depend onfewer input parameters, e.g., relief, rainfall, soilstructure, management (Schmidt 1998), but are lessaccurate compared to the complex models de-scribed above (Schmidt 1998; Michael 2000).However, because they are easily applied andrequire widely accessible parameters, they aremore frequently used to assess soil erosion (Re-nard et al. 1997; Bork et al. 1998; Schäuble 1999;Schmidt 2001), particularly for larger areas andannual time scales (Bonilla et al. 2007).

Changes in land use and land cover over timeare commonly derived from remote sensing(Cebecauer and Hofierka 2008) or historic maps(Schäuble 1999; Schmidt 2001; Jordan et al.2005; Szilassi et al. 2006), assuming other fac-tors being invariant (Schäuble 1999; Schmidt2001; Jordan et al. 2005; Cebecauer and Hofier-ka 2008). However, the effect of land-cover pat-terns, which have an impact on the energeticpotential of surface runoff and consequently onsoil erosion, is not considered in most studies.Common soil erosion models consider the impor-tance of land-cover patterns as small (Schäuble1999; Hickey 2000; Schmidt 2001). Nonetheless,transport potential of surface runoff increaseswith slope length, which is primarily regulatedby the patterns of land cover. The longer theslopes within a given homogeneous land cover

type, the greater is the erosive slope length. Con-sequently, changes in land-cover patterns mayaffect soil erosion considerably.

We therefore hypothesise that soil erosion is nota-bly influenced by changing land-cover patterns.Furthermore, we hypothesise that soil erosion po-tential is underestimated in common empirical mod-els as land-cover patterns are not adequately con-sidered. Thus, we developed a method to includeland-cover patterns in soil erosion modelling andtested it with a case study in the Saxon SwitzerlandNational Park Region (Fig. 1) for the last century.Although this study area is not dominated by arableland, the Saxon Switzerland National Park Region hasexperienced enormous, well-documented changes inland use and land cover over the last centuries.

2. Material and Methods

2.1 Study area

The Saxon Switzerland National Park Region (about51°N, 14°E) is a landscape protection area in thesouthern part of eastern Germany (Fig. 1). Most ofthe 380 km² area belongs to the Elbe SandstoneMountains (“Elbsandsteingebirge”) on both sidesof the Elbe river. Mean annual precipitation rangesfrom 700 to 800 mm with a summer maximum(Kaulfuß and Kramer 2000; Flemming 2001). Astructured terrain (Fig. 2), derived from sandstone,fostered poor, usually forested soils (Tröger 2006).Arable soils are mainly related to periglacial loesson plateaus (21 % of the area, Photo 1). In the past,agriculture and forestry dominated the area, withthe former, in particular, having changed signifi-cantly over the last century (Fichtner 2005).

2.2 Land-cover change

We assessed changes in land cover and land-coverpatterns based on digitised historical maps from1900, 1940 and 1992. Historical maps were available


back to 1780 (see Witschas 2002 and Fichtner 2005for details), however, the use of these maps failedbecause of inaccuracies including different aggre-gation scale and discrepancies in georeferencing.These inaccuracies considerably affected slopelength calculations and caused difficulties in thecomparison to the maps from 1900 to 1992, so pre-1900 data was not used in the final analysis.

The Geographic Information Systems (GIS) Arc-GIS 9.1 (ESRI) and the open source GIS SAGA2.0 were used to assess land-cover changes (cal-culation of area and number of agricultural units)and for modelling soil erosion.

2.3 Soil erosion modelling with RUSLE

The empirical Revised Universal Soil Loss Equa-tion (RUSLE) (Wischmeier and Smith 1978; Renardet al. 1997) estimates the longtime average annualsoil loss in t/ha/a (A) by the equation

(1)

where R denotes the Rainfall-runoff erosivity fac-tor (R-Factor), K the Soil erodibility factor (K-Fac-tor), S the Slope steepness factor (S-Factor), L theSlope length factor (L-Factor), C the Cover-man-agement factor (C-Factor) and P the Support prac-

Fig. 2 Relief map of the Saxon Switzerland National Park RegionHöhenschichtenkarte der Nationalparkregion Sächsische Schweiz

A R K S L C P= × × × × ×


tise factor (P-Factor). Schwertmann et al. (1987)adapted RUSLE to Central Europe. BesidesRUSLE, there are simplified versions that regardcertain factors as constant (Marks et al. 1992;Hennings 2000) or even exclude factors (see DIN19708; Hennings 2000). We also used such sim-plification for our case study (C-Factor and P-Fac-tor = 1) and assumed slope length as the onlyvariable factor in time (1900-1992).

Precipitation data from 1961 to 1990 collected at18 stations were corrected (for wind according toRichter 1995) and regionalised with a regressionanalysis related to topography based on a Digit-al Elevation Model (DEM) with a spatial resolu-tion of 20 meter. Because of the higher relevancefor soil erosion, only summer precipitation wasused (see Schwertmann et al. 1987) for the region-alisation. Mean annual precipitation (P, in mm)was regionalised by the equation

where h is elevation (R² = 0.5587); furthermore,orographic Luv-Lee adaption was applied. Basedon this dataset and regional adaptations for Sax-

ony, the R-Factor (R) was calculated accordingto DIN 19708 (2005) using

(3)

with NS being mean annual summer precipitationin mm (R² = 0.972).

K-Factors were derived from digital soil maps(Hennings 2000; DIN 19708 (2005). We calculatedS-factors (S) based on slope angles ( ) derivedfrom a 5-m DEM (selected here to account for theneeded greater spatial resolution, Schwertmannet al. 1987; Hickey 2000; Claessens et al. 2005),using the equation of Nearing (1997)

Unfortunately, the 5 meter DEM only coveredabout 90 % of the study area (see also Fig. 5).

2.4 Modelling erosive slope length

Estimating erosive slope length is often referredto as the most complicated part of erosion model-

24.2378.4100

P h= + ×

0.2755 50.03SR N= × −

2.3 6.1sin

171.51

Se β−= − +

+

Photo 1 Agricultural use of loess plateau near ReinhardtsdorfLandwirtschaftliche Nutzung einer Lössebene bei Reinhardtsdorf

(2)

(4)


ling (Moore and Wilson 1992; Renard et al. 1997;Wilson and Lorang 1999; Hickey 2000). Determi-nation in the field is only feasible for selectedslopes (Hickey 2000). Hence, average slopelengths are usually assumed (Schäuble 1999;Hennings 2000; Hickey 2000), neglecting any het-erogeneity. Accordingly, effects of land-coverpatterns cannot be considered. Previous ap-proaches for slope length calculation (Moore andWilson 1992; Desmet and Govers 1996; Schäuble1999; Wilson and Lorang 1999; Hickey 2000) usu-ally caused difficulties, in particular related to soft-ware (Schäuble 1999). In general, slope length isevaluated from a DEM via flow directions of sur-face runoff (Schäuble 1999; Hickey 2000; VanRemortel et al. 2001). Other approaches are basedon the Unit Stream Power Equation by Mooreand Burch (1986) (see Moore and Wilson 1992;Mitasova 1993; Mitasova et al. 1996; Mitasovaet al. 2001; Schmidt 2001) or use upslope catch-ment area as a substitute for slope length (Des-met and Govers 1996; Wilson and Lorang 1999;

Hickey 2000). In all approaches, surface runoffmodelling is the key to determining proper slopelengths for soil erosion modelling.

2.5 Modelling surface runoff

The major two algorithms for modelling surfacerunoff based on a DEM (Wilson and Lorang 1999)are (1) Single Flow Direction (SFDA) and (2) Mul-tiple Flow Direction (MFDA) algorithms (Fig. 3).Deterministic 8 (D8) by O’Callaghan and Mark(1984) is the established SFDA (Schäuble 1999;Wilson and Lorang 1999). With FD8, Freeman(1991) proposed the most common MFDA so far.Whereas SFDAs can only forward runoff in oneparticular direction, MFDAs facilitate multiple di-rections and are thus more realistic, although theyare less easily implemented. The latter allow com-puting runoff divergences rather than just con-vergences, and produce fewer artificial water di-vides. One disadvantage regarding divergence in

Fig. 3 Types of flow direction algorithms:Single Flow versus Multiple Flow(DEM = Digital Elevation Model,numbers indicate elevation in m a.s.l.;grey box is the origin of runoff, arrowsindicate its forwarding according to thetype of flow algorithm.) / Typen vonFließrichtungsalgorithmen: SingleFlow versus Multiple Flow (DEM=Digitales Geländemodell, die Zahlengeben die Höhe in m über NN an; diegraue Box kennzeichnet den Ursprungdes Abflusses, die Pfeile zeigen dieWeiterleitung entsprechend dem Typvon Fließrichtungsalgorithmen an.)


some MFDAs is the diffusion of surface runoffon flat terrain (like valley floors). This problem canbe solved by using a SFDA only in those partic-ular areas and is referred to as combined flow(Quinn et al. 1995; Schäuble 2003). The lack ofimplementation of MFDAs in standard GIS soft-ware calls for additional tools or GIS extensions,e.g., HydroTools for ArcView 3.3, TauDEM forArcGIS or Open Source GIS like SAGA.

Based on the types of algorithms to model sur-face runoff, two calculation approaches for slopelength can be distinguished: flow length based(FLA) and catchment based (CMA) approaches.Examples of FLAs are the Irregular Slope Conceptby Schäuble (1999) and the methods of Hickey(2000) and Van Remortel et al. (2001). Moore andWilson (1992), Desmet and Govers (1996) and theUnit Contribution Area Concept by Schäuble(1999) are examples of CMAs. While FLAs arebased on SFDAs, CMAs almost entirely rely onMFDAs. However, in general CMAs could alsobe used with SFDAs, but would then incorporatethe associated disadvantages of SFDAs. A moredetailed discussion of algorithms, tools for mod-elling surface runoff with GIS, and calculationapproaches can be found in Wolf (2006).

We conclude from our comparison of differentapproaches to calculate slope length that onlythe CMA of Desmet and Govers (1996) consist-ently implements the concept of the RUSLE andin addition considers irregular slopes. Further-more this approach is applicable in most stand-ard GIS with additional tools by using MFDAs.As an additional modification we propose to usea catchment size weighted slope length expo-nent as recommended by Böhner et al. (2001;see also Böhner and Selige 2006). Desmet andGovers (1996) calculate the local slope lengthfactor Li with the equation

where Ai is the catchment size of slope segmenti in m², D the grid resolution in m and mi the localslope length exponent.

The determination of deposition areas by empiricalmodels is complicated (Schäuble 1999; Wilson andLorang 1999; Schmidt 2001) and few models incor-porate it. RUSLE is not able to predict deposition(Bonilla et al. 2007) without modifications. Thus,this aspect was not considered in our study andaccordingly, we instead estimate a “worst-case”scenario. Furthermore, the relative development ofsoil erosion due to changes in land cover and itspatterns are the main focus of our study and con-sequently net rates of erosion displacement wereless essential. Therefore, it was assumed in our mod-elling approach that erosion rates were in all instanc-es higher than deposition rates.

2.6 Consideration of flow barriers

Although an empirical model is used to quantifysoil erosion, a critical discussion about the proc-ess-related consideration of certain parametersto improve modelling results should not be ne-glected. Land-cover patterns can be consideredin soil erosion modelling via the erosive slopelength. Changes in land use and land cover af-fect surface runoff and thereby the suspendedsoil material. Thus land-cover patterns are a reg-ulating factor of soil erosion, besides the topo-graphic aspects of surface runoff, and it is es-sential to determine the erosive slope length. Theerosive slope length is calculated based on aDEM, for the topographic characteristic, and theland-cover patterns as a characteristic limitingsurface runoff. The adequate combination ofboth inputs is a critical point. In the context ofsoil erosion, flow barriers from land-cover patternsare elements that affect the surface runoff and con-sequently the displacement of soil material. Flowbarriers can be for instance hedges, tree rows,driveways, roads or simply a change of land cover(e.g., from farmland to grassland or forest and vice

( ) 1 12

2 22.13

i i

i i

m mi i

i m m

A D AL

D

+ +

+

+ −=

× (5)


versa). Affecting surface runoff, these flow barri-ers essentially determine the erosive slope length(Photo 2). Flow barriers can slow down surfacerunoff, even stop it (resulting in deposition ofmaterial), or they can transmit and accelerate run-off due to convergence, which might increase theenergetic transport potential and, in turn, increasesoil erosion. For this reason it is crucial to con-sider how flow barriers are incorporated in mod-elling surface runoff and soil erosion. So far, inerosion modelling, flow barriers are considered asareas with unlimited infiltration potential (e.g.,

Schäuble 1999; Schmidt 2001). This results in theproblematic assumption that surface runoff (i.e.mobilised soil material in suspension) is alwaysstopped at flow barriers without any transmission(Fig. 4a). However, in reality surface runoff istopography-dependent and stopped or sloweddown only when either the inclination of a slopeand thus the transport energy is limited (depend-ing on the amount of runoff as well) or a signifi-cant change in surface roughness is limiting run-off (and thus leads to deposition of transportedsoil material in suspension). Hence, erosive slope

Photo 2 Farmland near Pfaffendorf with moderate slope angle and large slope length. Grass, hedges andsmall trees act as flow barriers and separate agricultural units; here they are, however, parallel tothe slope and thus not effective in reducing erosive slope length and soil erosion potential.Ackerland bei Pfaffendorf mit mäßiger Hangneigung und großer Hanglänge. Fließbarrierenin Form von Gras, Sträuchern und kleinen Bäumen trennen die Ackerschläge voneinander.Durch ihren hangparallelen Verlauf reduzieren diese jedoch nicht die erosive Hanglänge unddie Gefahr für Bodenabtrag.


length is limited by any flow barrier independentlyof the topographic situation. Converging effectsalong, or transmission of surface runoff through,flow barriers are thus neglected. Consequently,the surface runoff uphill from flow barriers has noimpact on the downhill areas in this approach ofunlimited infiltration. However, these are partic-ularly vulnerable areas as the cumulated surfacerunoff leads to a greater transport potential andthus larger erosive slope lengths. Therefore wepropose to consider flow barriers in surface run-off for erosion modelling without infiltration (Fig.4b). Soil erosion typically occurs during extremeconvective precipitation events (like thunder-storms), with strong surface runoff providing theconditions to transport soil material. As the soilsurface is then saturated within a short time, theinfiltration potential is relatively low and couldthus be neglected. We think that our approach

represents the process of soil erosion more ade-quately in empirical models.

In GIS modelling, flow barriers are typically consid-ered with runoff masks and modifications of the DEM.Hence, areas which are marked as flow barriers in therunoff mask are set to NoData in the DEM or consid-ered separately with weighting rasters in the model-ling of surface runoff (Schäuble 1999). While the firstapproach results in the explained unlimited infiltra-tion, the application of weighting rasters frequentlyresults in an overflow of flow barriers and thus erro-neous modelling results. Therefore, we applied theproposed approach with an artificial exaggeration offlow barriers to realize a topography-related trans-mission of surface runoff along flow barriers. Toavoid cases of overflow on scarps, we recommendexaggerating flow barriers with the maximum eleva-tion occurring in the investigated study area.

Fig. 4 Illustration of surface runoff in relation to flow barriers. The consequences of the differentconsiderations of flow barriers are shown. With infiltration (a): surface runoff is stoppedcompletely by flow barriers. Without infiltration (b): converging surface runoff along flowbarriers and diverging surface runoff downslope leads to higher energetic transport potential andthus higher soil erosion potential. / Darstellung von Oberflächenabfluss an Fließbarrieren.Dargestellt sind die Auswirkungen einer unterschiedlichen Berücksichtigung von Fließbarrieren.Mit Versickerung (a): Der Oberflächenabfluss wird beim Auftreffen auf Fließbarrieren gestoppt.Ohne Versickerung (b): Konvergierender Oberflächenabfluss entlang von Fließbarrieren unddivergierender Abfluss hangabwärts steigern das energetische Transportpotential und verursa-chen dadurch ein höheres Potential für Bodenerosion.

With infiltration (a) Without infi ltration (b)


3. Results

In order to quantify impacts of changes in landcover and its patterns on soil erosion, we evalu-ated how arable land developed. The area of ar-able land decreased by 25 % from 1900 to 1992,covering about 21 % of the National Park Regionin 1992 (Tab. 1). Few changes in land cover tookplace in the first half of the 20th century, but after

1940 land reform increased the mean size of ag-ricultural units by nearly 400 %, from 2.6 to 9.8hectares. Thus, the number of agricultural unitsdecreased significantly from about 3300 to 700(Tab. 1). These changes in land-cover patternsresulted in an increase in erosive slope lengthsand consequently a rise in soil erosion potential(Tab. 2). In addition to land reform, some arableland was relocated towards siltier soils, areas

1900 1940 1992

Number of agricultural units 3332 3200 663

Max. size of agricultural units [ha] 42.3 55.7 223.9

Mean size of agricultural units [ha] 2.6 2.5 9.8

Total area of arable land [ha] 8697.7 7958.9 6502.9

Tab. 1 Change of number and size of agricultural units in the investigation area, 1900-1992Schlagbezogene Entwicklung der Ackerflächen im Untersuchungsgebiet 1900-1992

Tab. 2 Mean modelling factors and soil erosion displacementMittlere Modellierungsfaktoren und Bodenabtrag

1900 1940 1992

R-Factor 74 74 74

Change [%] - -

K-Factor 0.47 0.48 0.49

Change [%] +2.1 +2.1

L-Factor 2.4 2.3 2.5

Change [%] -4.2 +8.7

S-Factor 1.3 1.2 0.9

Change [%] -7.7 -25.0

LS-Factor 3.7 3.3 2.6

Change [%] -10.8 -21.2

Soil erosion [t/ha/a] 117.5 108.4 91.2

Change [%] -7.7 -15.9


more vulnerable to soil erosion. In contrast, ag-ricultural use of steep slopes >10° was gradual-ly reduced since 1900, which reduced the S-Fac-tor and, thus, soil erosion potential. Obviously,the latter is able to compensate for the changesin the former, as modelled mean soil erosiondecreased by 22 % from 1900 to 1992 with mostof this decrease occurring after 1940 (Tab. 2),exactly in the same period when erosive slopelength increased most significantly due tochanges in land-cover patterns. However, as theslope angle is also included in assessing the L-factor (via slope length exponent m) an increasein slope length may be counterbalanced. Meanerosion values are mainly driven by the with-drawal of agriculture from steep slopes. How-ever, we wanted to reach beyond such net ef-

fects. Therefore, we attempted to separate chang-es in soil erosion due to changes in land-coverpatterns from spatial relocation and the conver-sion to other land-cover types by restricting our-selves to areas which had been continuouslyused as arable land from 1900 to 1992 (determinedby spatial intersection with GIS). Thus, we wereable to separate and analyse independentlyslope-length-related changes of land-cover pat-terns. The results are decisively different fromthose sketched above (Tab. 3). After a slight de-crease from 1900 to 1940, mean soil erosion dis-placement on continuously farmed land in-creased by almost 11 %. In total, soil erosion in-creased from 1900 to 1992 on 41 % of the contin-uously farmed land (Tab. 4, Fig. 5). 9 % showeda strong increase. Only 13 % of the agricultural

1900 1940 1992

Soil erosion [t/ha/a] 79.5 79.4 87.9

Change [%] -0.1 +10.7

Percent of agricultural units

1900-1940 1940-1992 1900-1992

Strong decrease 2.0 1.6 1.9

Decrease 8.9 10.0 10.8

No change 79.4 49.2 46.8

Increase 8.4 30.2 31.6

Strong increase 1.3 9.0 9.0

100 100 100

Tab. 3 Mean soil erosion on continuously-used farmland due to changes in land-cover patternsMittlerer Bodenabtrag durch Veränderungen der Landnutzungsstruktur auf kontinuierlich genutztenAckerflächen

Tab. 4 Slope-length-induced changes in soil erosion on continuously-used agricultural unitsHanglängenbedingte Veränderungen des Bodenabtrags auf kontinuierlich genutzten Ackerflächen


Fig. 5 Soil erosion by agricultural units. Colours indicate soil erosion per agricultural unit; 5 m DEM coversonly about 90 % of the study area. / Schlagbezogener Bodenabtrag. Die Farben kennzeichnen denBodenabtrag je Ackerschlag; 5 m DGM deckt nur etwa 90 % des Untersuchungsgebietes ab.


1900 1940 1992

Lan

d u

se

Lan

d-c

over

patt

ern

black =

flow barriers

white =

arable land

Ero

sio

n d

isp

lacem

en

t

1900-1992

Fig. 6 Development of erosion displacement due to changes in land-cover pattern at a test site nearReinhardtsdorf / Entwicklung des Bodenabtrags durch Veränderungen der Landnutzungsstrukturin einem Testgebiet bei Reinhardtsdorf

black =

flow barriers

white =

arable land


units observed a decrease in soil erosion from1900 to 1992, whereas 47 % remained unchanged.Most of the increases are caused by the substan-tial changes in land-cover patterns after 1940. From1900 to 1940 almost 80 % of the continuous agri-cultural units were unaffected by marked chang-es in soil erosion. Predominantly, areas with en-hanced natural erosion vulnerability (mainly steepslopes at plateau rims) were identified as agricul-tural units with an increase in soil erosion. This isnot surprising, as they are particularly sensitiveto changes in land-cover patterns.

Several test areas were investigated in greaterdetail to assess the spatial differences in thestudy area (e.g., Fig. 6), and to document the de-velopment of land-cover patterns and soil ero-sion over the last century. One test area is locat-ed near the community of Reinhardtsdorf, on aplateau bordering the Elbe valley (Fig. 6). Soilerosion is most intense on foot slopes due togreater slope lengths and slope angles. The ef-fect of convex curvature in combination with higherosive slope length results in high soil erosionrates. In 1900 only driveways perpendicular tothe contour lines existed. Some additional drive-ways along the contour lines had evolved by1940. Consequently, soil erosion decreased be-low these driveways (flow barriers) due to ashortening of erosive slope length. With landconsolidation and elimination of some drivewaysafter 1940, soil erosion increased on more than60 % of the farmland. Newly developed agricul-tural units located in the north-western part ofthe area have only low slope angles. Thus, evenlarge erosive slope lengths cannot be translatedinto high kinetic energy and result in moderatesoil erosion rates only.

4. Discussion

Utilising the widely used and relatively easilyapplied RUSLE brings along the limitations ofan empirical model (cf. e.g., Schmidt 1998;

Michael 2000). Consequently, we regard resultsfrom soil erosion modelling with RUSLE as anapproximation of long-term soil erosion, at best.Yet, to approximate the complex process of soilerosion adequately, empirical models can beimproved by including physically-based ap-proaches. Approaches proposed in this studyare the consideration of flow barriers and theinclusion of the effects of land-cover patterns.Standard GIS and open source software withtools and extensions (e.g. HydroTools, Tau-DEM) provide the functionality to implementthis approach without the need for additionaldata. The basis for this data is critical for thisapproach, particularly when using historicalmaps. Carefully assessing the quality of thesehistorical maps is imperative. The main prob-lem we experienced with historical maps usedin our case study are location discrepanciescompared to the DEM. In boundary areas ofland cover, this resulted occasionally in smallareas of arable land being virtually located onvery steep slopes (>15°; see AG Boden 2005).It is questionable in general whether some ofthe agricultural units are really located on verysteep slopes or whether this is due to classifi-cation errors of mapping.

Due to the interrelation of RUSLE factors (e.g.S- and L-factors), certain effects (e.g. slope an-gle) have a much greater influence on modelledsoil erosion than other factors and, thus, havemore weight within the model (Schwertmann etal. 1987). This fact is crucial to the interpretationof RUSLE-based model results, because soil ero-sion can only take place if combined slope angleand slope length provide a critical amount of ki-netic energy to displace soil material.

Our results indicate a considerable effect ofchanged land-cover patterns on soil erosion.We suggest that this finding is applicable be-yond the boundaries of the Saxon SwitzerlandNational Park Region, because land reform wasintroduced in other regions in Germany after


1940 as well (e.g., Halke 2002; Thormann 2002).Areas with high natural erosion vulnerabilitywere affected most by changes in land-coverpatterns, and areas with similar environmentalconditions may have experienced comparableevolution of soil erosion. Consequently, weassume that changes in land-cover patterns inthe second half of the 20th century have lessseverely affected typical low-angle agricultur-al regions, because effects of increased erosiveslope length were limited there.

Land-cover pattern is an important factor affect-ing soil erosion. Thus, we suggest its study isa new and realistic approach to assess histor-ical changes in soil erosion.

Furthermore, we suggest our approach may helpidentifying future erosion risks under changingenvironmental conditions. Climate change, forinstance, bears the potential for shifts in landuse to adapt to altered climatic conditions(O’Neal et al. 2005). Farming adaptations (e.g.,changed crop types, shifted planting, cultivationand harvest dates) will presumably change thesusceptibility of soil to erosive forces (South-worth et al. 2000; Pfeifer and Habeck 2002;Kundzewicz et al. 2007). For example, O’Neal etal. (2005) suggest that future land-use changeswill increase erosion, largely due to a general shiftfrom wheat and maize to soybean, although otherscenarios of land-use changes lead to differentresults mainly because improved conservationpractices are assumed (Souchere et al. 2005;Kundzewicz et al. 2007).

5. Conclusions

Our study suggests that changes in land-coverpatterns affect soil erosion due to changes inerosive slope length. Thus, assessing soil ero-sion should include the investigation of land-cover patterns to properly calculate erosiveslope length. Using average slope lengths may

lead to an underestimation of soil erosion. Inparticular flow barriers need to be considered.

Our case study in the Saxon Switzerland NationalPark Region revealed that the aggregation of agri-cultural units in the second half of the last centu-ry resulted in an increase of erosive slope lengthsand thus soil erosion. At the same time agricul-ture entirely retreated from steep slopes, reduc-ing average soil erosion. Consequently, modellingaverage soil erosion over larger areas does notnecessarily display the effects of increasing ero-sive slope lengths and may result in misinterpre-tation. Thus, causes of changes in soil erosionshould be assessed in detail in regard to spatialpatterns. Our approach to assess historical chang-es of soil erosion may be of use in simulating de-velopments in soil erosion due to expected chang-es in environmental conditions (e.g., climatechange) or land-use policies. Based on these sim-ulations, optimisation of land-cover patterns ap-pears possible and should be considered in plan-ning decisions which aim at supporting sustaina-ble development and minimising hazardous an-thropogenic soil erosion in the future.

Acknowledgements

This study was conducted within the SISTEMa-PARC project of the Community Initiative Pro-gramme INTERREG III B CADSES that was partlyfunded by the European Union. We thank JohannesFranke, Department of Meteorology, Dresden Uni-versity of Technology, for assistance with theregionalisation of precipitation data and for provi-ding the orographic adaption. We highly apprecia-ted the support by Jürgen Böhner and Olaf Con-rad, Institute of Geography, University of Ham-burg, with discussions and module adaptations ofSAGA. Marco Trommler, Institute of Photogram-metry and Remote Sensing (IPF), Dresden Univer-sity of Technology (TUD), provided the DEM.Furthermore, we like to thank Werner Eugster andNina Buchmann, Institute of Plant Sciences, ETHZurich, for critical comments on the manuscript.


6. References

AG BODEN 2005: Bodenkundliche Kartieranlei-tung. – HannoverBöhner, J., R. Köthe, O. Conrad, J. Gross, A. Ringeler andT. Selige 2002: Soil Regionalisation by Means of TerrainAnalysis and Process Parameterisation. – In: Micheli, E.,F.O. Nachtergaele, R.J.A. Jones and L. Montanarella(eds.): Soil Classification 2001: Contributions to theInternational Symposium “Soil Classification 2001”,8-12 October, Velence, Hungary. – European SoilBureau Research Reports 7. – Luxembourg: 213-222Böhner, J. and T. Selige 2006: Spatial Prediction of SoilAttributes using Terrain Analysis and Climate Regiona-lisation. – In: Böhner, J., K.R. McCloy and J. Strobl (eds.):SAGA – Analyses and Modelling Applications. –Göttinger Geographische Abhandlungen 115: 13-28Bonilla, C.A., J.M. Norman and C.C. Molling 2007:Water Erosion Estimation in Topographically Com-plex Landscapes: Model Description and FirstVerifications. – Soil Science Society of AmericaJournal 71 (5): 1524-1537Boardman, J. 2006: Soil Erosion Science: Reflec-tions on the Limitations of Current Approaches. –Catena 68 (2-3): 73-86Bork, H.-R., H. Bork, C. Dalchow, B. Faust, H.-P.Piorr und T. Schatz 1998: Landschaftsentwicklungin Mitteleuropa. Wirkungen des Menschen auf Land-schaften. – StuttgartCebecauer, T. and J. Hofierka 2008: The Consequencesof Land-Cover Changes on Soil Erosion Distribution inSlovakia. – Geomorphology 98 (3-4): 187-198Claessens, L., G.B.M. Heuvelink, J.M. Schoorl andA. Veldkamp 2005: DEM Resolution Effects onShallow Landslide Hazard and Soil RedistributionModelling. – Earth Surface Processes and Land-forms 30 (4): 461-477Costa-Cabral, M.C. and S.J. Burges 1994: DigitalElevation Model Networks (DEMON) – a Model ofFlow over Hillslopes for Computation of Contri-buting and Dispersal Areas. – Water ResourcesResearch 30 (6): 1681-1692Desmet, P.J.J. and G. Govers 1996: A GIS Procedurefor Automatically Calculating the USLE LS Factor onTopographically Complex Landscape Units. – Jour-nal of Soil and Water Conservation 51 (5): 427-433

Deumlich, D., R. Funk, M. Frielinghaus, W.-A.Schmidt and O. Nitzsche 2006: Basics of EffectiveErosion Control in German Agriculture. – Journal ofPlant Nutrition and Soil Science 169 (3): 370-381Diemann, R. und O. Arndt 2001: Historische Landnut-zung – Chronologische Analysen auf der Querfurt-Merseburger Platte. – scientia halensis 9 (3): 31-32DIN 19708 2005: Bodenbeschaffenheit – Ermitt-lung der Erosionsgefährdung von Böden durch Was-ser mit Hilfe der ABAG. – BerlinEvans, R. and R. Brazier 2005: Evaluation of ModelledSpatially Distributed Predictions of Soil Erosion byWater versus Field-Based Assessments. – Environ-mental Science and Policy 8 (5): 493-501Fichtner, S. 2005: Flächennutzungsmonitoring inder Nationalparkregion Sächsische Schweiz –Methodik und Analyse. – Diploma thesis, Techni-sche Universität Dresden, Institut für Photogram-metrie und FernerkundungFoley, J.A., R. DeFries, G.P. Asner, C. Barford, G. Bonan,S.R. Carpenter, F.S. Chapin, M.T. Coe, G.C. Daily,H.K. Gibbs, J.H. Helkowski, T. Holloway, E.A. Howard,C.J. Kucharik, C. Monfreda, J.A. Patz, I.C. Prentice, N.Ramankutty and P.K. Snyder 2005: Global Consequencesof Land Use. – Science 309 (5734): 570-574Foster, G.R., T.E. Toy and K.G. Renard 2003: Compa-rison of the USLE, RUSLE1.06c, and RUSLE2 forApplication to Highly Disturbed Lands. – In: Renard,K.G., S. McElroy, W. Gburek, E. Canfield and R.L. Scott(eds.): Proceedings of the First Interagency Conferenceon Research in the Watersheds. – Benson: 154-160Flanagan, D.C., J.E. Gilley and T.G. Franti 2007: WaterErosion Prediction Project (WEPP): DevelopmentHistory, Model Capabilities and Future Enhance-ments. – Transactions of the American Society ofAgricultural and Biological Engineers 50 (5):1603-1612Flemming, G. 2001: Angewandte Klimatologie vonSachsen, Basis- und Zustandsklima im Überblick. –Tharandter Klimaprotokolle 4Freeman, T.G. 1991: Calculating Catchment Areawith Divergent Flow Based on a Regular Grid. –Computers and Geosciences 17 (3): 413-422Halke, E. 2002: Historischer Landnutzungswandel unddessen Einfluss auf ausgewählte Landschaftsfunktio-nen – am Beispiel Oelsnitz. – Diploma thesis, Hum-boldt-Universität zu Berlin, Geographisches Institut


Hennings, V. (Hrsg.) 2000: MethodendokumentationBodenkunde. Auswertungsmethoden zur Beurtei-lung der Empfindlichkeit und Belastbarkeit vonBöden. – Geologisches Jahrbuch, Sonderhefte,Reihe G 1. – 2. Auflage. – StuttgartHickey, R.J. 2000: Slope Angle and Slope LengthSolutions for GIS. – Cartography 29 (1): 1-8IPCC 2000: Land Use, Land-Use Change and Fore-stry. – A Special Report of the IntergovernmentalPanel on Climate Change, Summary for Policy-makers. – GenevaJordan, G., A. van Rompaey, P. Szilassi, G. Csillag,C. Mannaerts and T. Woldai 2005: Historical LandUse Changes and their Impact on Sediment Fluxesin the Balaton Basin (Hungary). – Agriculture,Ecosystems and Environment 108 (2): 119-133Kaulfuß, W. und M. Kramer 2000: Naturlandschaftenund Nutzungspotenziale Sachsens. – In: Kowalke, H.(Hrsg.): Sachsen: geographische Strukturen, Entwick-lungen, Probleme. – Gotha: 71-73Kundzewicz, Z.W., L.J. Mata, N.W. Arnell, P. Döll,P. Kabat, B. Jiménez, K. Miller, T. Oki, Z. Sen andI.A. Shiklomanov 2007: Freshwater Resources andTheir Management. – In: Parry, M.L., O.F. Canziani,J.P. Palutikof, P.J. van der Linden and C.E. Hanson(eds.): Climate Change 2007: Impacts, Adaptation andVulnerability. – Contribution of Working Group II tothe Fourth Assessment Report of the Intergovernmen-tal Panel on Climate Change. – Cambridge: 173-210Lambin, E.F. and H.J. Geist (eds.) 2006: Land-Useand Land-Cover Change: Local Processes and GlobalImpacts. – BerlinMarks, R., M.J. Müller, H. Leser und H.-J. Klink (Hrsg.)1992: Anleitung zur Bewertung des Leistungsvermö-gens des Landschaftshaushaltes. – Forschungen zurdeutschen Landeskunde 229. – 2. Auflage. –TrierMichael, A. 2000: Anwendung des physikalischbegründeten Erosionsprognosemodells EROSION2D/3D – Empirische Ansätze zur Ableitung derModellparameter. – Freiberger Forschungshefte 488Michael, A., J. Schmidt, W. Enke, T. Deutschlander andG. Malitz 2005: Impact of Expected Increase in Preci-pitation Intensities on Soil Loss – Results of Compa-rative Model Simulations. – Catena 61 (2-3): 155-164Mitasova, H. 1993: Surfaces and Modelling. – Grass-clippings 7 (1): 18-19

Mitasova, H., J. Hofierka, M. Zlocha and L.R. Iverson1996: Modelling Topographic Potential for Erosionand Deposition using GIS. – International Journal ofGIS 10 (5): 629-641Mitasova, H., L. Mitas and W.M. Brown 2001:Multiscale Simulation of Land Use Impact on SoilErosion and Deposition Patterns. – In: Stott, D.E.,R.H. Mohtar and G.C. Steinhardt (eds.): Sustainingthe Global Farm. – Selected Papers from the 10th

International Soil Conservation Organization Mee-ting, May 24-29, 1999, West Lafayette, IN. – WestLafayette: 1163-1169Moore, I.D. and G.J. Burch 1986: Physical Basis ofthe Length-Slope Factor in the Universal Soil LossEquation. – Soil Science Society of America Journal50: 1294-1298Moore, I.D. and J.P. Wilson 1992: Length-SlopeFactors for the Revised Universal Soil Loss Equation:Simplified Method of Estimation. – Journal of Soiland Water Conservation 47 (5): 423-428Nearing, M.A. 1997: A Single, Continuous Functionfor Slope Steepness Influence on Soil Loss. – SoilScience Society of America Journal 61 (3): 917-919Nearing, M.A., V. Jetten, C. Baffaut, O. Cerdan,A. Couturier, M. Hernandez, Y. Le Bissonnais,M.H. Nichols, J.P. Nunes, C.S. Renschler, V. Souchèreand K. Van Oost 2005: Modeling Response of SoilErosion and Runoff to Changes in Precipitation andCover. – Catena 61 (2): 131-154O’Callaghan, J.F. and D.M. Mark 1984: The Extrac-tion of Drainage Networks from Digital ElevationData. – Computer Vision, Graphics, and ImageProcessing 28 (3): 323-344O’Neal, M.R., M.A. Nearing, R.C. Vining, J. Southworthand R.A. Pfeifer 2005: Climate Change Impacts on SoilErosion in Midwest United States with Changes inCrop Management. – Catena 61 (2-3): 165-184Pfeifer, R.A. and M. Habeck 2002: Farm Level Econo-mic Impacts of Climate Change. – In: Doering, O.C.,J.C. Randolph, J. Southworth and R.A. Pfeifer (eds.):Effects of Climate Change and Variability on Agricul-tural Production Systems. – Boston: 159-178Pielke, R.A., J. Adegoke, A. Beltrán-Przekurat, C.A.Hiemstra, J. Lin, U.S. Nair, D. Niyogi and T.E. Nobis2007: An Overview of Regional Land-Use and Land-Cover Impacts on Rainfall.– Tellus B 59 (3): 587-601


Quinn, P.F., K.J. Beven and R. Lamb 1995: The ln( / tan ) Index: How to Calculate it and how toUse it within the TOPMODEL Framework. –Hydrological Processes 9 (2): 161-182

Ramankutty, N., L. Graumlich, F. Achard, D. Alves,A. Chhabra, R.S. DeFries, J.A. Foley, H. Geist,R.A. Houghton, K. Klein Goldewijk, E.F. Lambin,A. Millington, K. Rasmussen, R.S. Reid and B.L. TurnerII, 2006: Global Land-Cover Change: Recent Pro-gress, Remaining Challenges. – In: Lambin, E.F. andH. Geist (eds): Land-Use and Land-Cover Change:Local Processes and Global Impacts. – Berlin: 9-39

Renard, K.G., G.R. Foster, G.A. Weesies, D.K. McCooland D.C. Yoder 1997: Predicting Soil Erosion by Water:A Guide to Conservation Planning with the RevisedUniversal Soil Loss Equation (RUSLE). – USDA-ARSAgriculture Handbook 703. – Washington D.C.

Richter, D. 1995: Ergebnisse methodischer Unter-suchungen zur Korrektur des systematischen Messfeh-lers des Hellmann-Niederschlagsmessers. – Berichte desDeutschen Wetterdienstes 194. – Offenbach am Main

Richter, G. 1998: Bodenerosion und Kulturland-schaft. – In: Richter, G. (ed.): Bodenerosion –Analyseund Bilanz eines Umweltproblems. – Darmstadt

Schäuble, H. 1999: Erosionsprognosen mit GISund EDV – Ein Vergleich verschiedener Bewertungs-konzepte am Beispiel einer Gäulandschaft. –Diplomarbeit, Eberhard-Karls-Universität Tübin-gen, Institut für Geographie

Schäuble, H. 2003: HydroTools 1.0 für ArcView 3.x –Hydrologische Analysen mit Single-Flow undMultiple-Flow Algorithmen. – Darmstadt

Schmidt, J. 1998: Modellbildung und Prognose zurWassererosion. – In: Richter, G. (ed.): Bodenerosion –Analyse und Bilanz eines Umweltproblems. –Darmstadt: 137–151

Schmidt, M. 2001: Bewertung der Anwendbarkeit zweierGIS-gestützter Modelle zur Vorhersage der Boden-erosion in agrarisch genutzten Einzugsgbieten. –Diplomarbeit, Katholische Universität Eichstätt,Mathematisch-Geographische Fakultät

Schneider, N. and W. Eugster 2007: Climatic Impactsof Historical Wetland Drainage in Switzerland. –Climatic Change 80 (3-4): 301-321

Schwertmann, U., W. Vogl und M. Kranz 1987:Bodenabtrag durch Wasser. Vorhersage des Abtragsund Bewertung von Gegenmaßnahmen. – Stuttgart

Souchere, V., O. Cerdan, N. Dubreuil, Y. Le Bis-sonnais and C. King 2005: Modelling the Impact ofAgri-Environmental Scenarios on Runoff in aCultivated Catchment (Normandy, France). –Catena 61 (2-3): 229-240

Southworth, J., J.C. Randolph, M. Habeck, O.C.Doering, R.A. Pfeifer, D.G. Rao and J.J. Johnston2000: Consequences of Future Climate Change andChanging Climate Variability on Maize Yields in theMid-Western United States. – Agriculture, Eco-systems and Environment 82 (1-3): 139-158

Szilassi, P., G. Jordan, A. van Rompaey and G.Csillag 2006: Impacts of Historical Land UseChanges on Erosion and Agricultural Soil Proper-ties in the Kali Basin at Lake Balaton, Hungary. –Catena 68 (2-3): 96-108

Thormann, D. 2002: Untersuchungen zum Land-nutzungswandel und Auswirkungen auf ausgewähl-te Landschaftsfunktionen. – Diploma thesis, Mar-tin-Luther-Universität Halle-Wittenberg, Institutfür Geographie

Tröger, M. 2006: Analyse der Entwicklung derFlächennutzungsstruktur der NationalparkregionSächsische Schweiz unter Einsatz von Landschafts-strukturmaßen. – Diploma thesis, Technische Uni-versität Dresden, Institut für Geographie

Turner II, B.L., E.F. Lambin and A. Reenberg 2007:The Emergence of Land Change Science for GlobalEnvironmental Change and Sustainability. – Procee-dings of the National Academy of Sciences of theUSA 104 (52): 20666-20671

Van Remortel, R., M. Hamilton and R. Hickey 2001:Estimating the LS Factor for RUSLE Through Itera-tive Slope Length Processing of Digital ElevationData. – Cartography 30 (1): 27-35

Wilson, J.P. and M.S. Lorang 1999: Spatial Modelsof Soil Erosion and GIS. – In: Fotheringham, A.S.and M. Wegener (eds.): Spatial Models and GIS.New Potential and New Models. – London

Wischmeier, W.H. and D.D. Smith 1978: PredictingRainfall Erosion Losses. A Guide to ConservationPlanning. – USDA Agriculture Handbook 537. –Washington D.C.

Witschas, S. 2002: Erinnerung an die Zukunft –sächsische historische Kartenwerke zeigen denLandschaftswandel. – Kartographische Nach-richten 52 (3): 111-117


Wolf, S. 2006: Bodenerosion als Funktion veränderterLandnutzungsstruktur. Modellierung der Entwicklungam Beispiel der Nationalparkregion SächsischeSchweiz. – Diploma thesis, Technische UniversitätDresden, Institut für Geographie

Summary: Effects of Land-Cover Change on SoilErosion in the Saxon Switzerland National ParkRegion

Changes in land cover and its patterns affect lateralprocesses like surface runoff and consequently soilerosion. However, the effects of changes in land-cover patterns are usually not included when mod-elling soil erosion. Therefore, the effects of land-useand land-cover changes on soil erosion are under-estimated. Using an empirical modelling approachbased on the Revised Universal Soil Loss Equation(RUSLE), a methodology is proposed to includechanges in land-cover patterns in soil erosion mod-els. An assessment of existing approaches to calcu-late erosive slope length is included. In addition, animproved method for the consideration of flow bar-riers without infiltration is suggested in order tomodel surface runoff more accurately according tothe process of soil erosion. A case study on soilerosion in the Saxon Switzerland (“SächsischeSchweiz”) National Park Region, Germany, from1900 to 1992, confirmed our hypothesis that chang-es in land-cover patterns considerably affect erosiveslope length and accordingly soil erosion. While soilerosion decreased on average over the whole area asagricultural use of steep slopes was reduced, otherchanges in land-cover patterns caused an increase ofsoil erosion. From 1900 to 1992 soil erosion in-creased on 41 % of the agricultural units in the studyarea due to changing land-cover patterns. This in-crease primarily occurred after 1940, in conjunctionwith land reform and the aggregation of agriculturalunits resulting in longer slopes. Areas with a highNatural Erosion Vulnerability emerged as particu-larly sensitive to changes in land-cover patterns.Overall, our results suggest that the effects of land-cover patterns on soil erosion have been underesti-mated so far and need to be considered more pre-cisely in empirical soil erosion modelling.

Zusammenfassung: Auswirkungen von Land-nutzungsänderungen auf die Bodenerosion in derNationalparkregion Sächsische Schweiz

Veränderungen in der Landnutzung und Landnut-zungsstruktur beeinflussen Prozesse wie den Ober-flächenabfluss und folglich die Bodenerosion. Den-noch werden die Auswirkungen von Veränderun-gen der Landnutzungsstruktur im Allgemeinennicht bei der Modellierung der Bodenerosion be-rücksichtigt. Dadurch werden die Folgen von Land-nutzungsänderungen auf die Bodenerosion unter-schätzt. Mit der Anwendung eines empirischenModellierungsansatzes, basierend auf der verbes-serten Allgemeinen Bodenabtragsgleichung(RUSLE), wird ein Ansatz vorgeschlagen, welcherVeränderungen in der Landnutzungsstruktur mitberücksichtigt. Zudem wird ein Vergleich von be-stehenden Verfahren zur Berechnung der erosivenHanglänge durchgeführt. Ferner wird eine verbes-serte Berücksichtigung von Fließbarrieren ohneVersickerung vorgeschlagen, um Oberflächen-abfluss genauer entsprechend dem Prozess derBodenerosion zu modellieren. Die Anwendung derMethode am Beispiel einer Fallstudie für dieNationalparkregion Sächsische Schweiz (Deutsch-land) von 1900 bis 1992 bestätigt die Hypothese,dass Veränderungen der Landnutzungsstruktur dieerosive Hanglänge und somit die Bodenerosion ent-scheidend beeinflussen. Während die Bodenerosiondurch den Rückzug von Ackerflächen aus steilenHangpositionen im Durchschnitt für das gesamte Un-tersuchungsgebiet abgenommen hat, haben Verän-derungen in der Landnutzungsstruktur ein Ansteigender Bodenerosion verursacht. Von 1900 bis 1992 hatdie Bodenerosion durch Veränderungen in der Land-nutzungsstruktur auf 41 % der Ackerflächen zuge-nommen. Insbesondere nach 1940 ist die Boden-erosion durch flurbereinigende Maßnahmen und denZusammenschluss der Bauern in landwirtschaftlichenProduktionsgenossenschaften angestiegen. Flächenmit einer hohen natürlichen Erosionsgefährdung ha-ben dabei besonders empfindlich auf Veränderungenin der Landnutzungsstruktur reagiert. Insgesamt zei-gen unsere Ergebnisse, dass die Auswirkungen vonVeränderungen der Landnutzungsstruktur auf dieBodenerosion bisher unterschätzt wurden und in derempirischen Modellierung der Bodenerosion genauerberücksichtigt werden müssen.


Résumé: L’influence de l’utilisation des terres surl’érosion du sol dans la région du parc national dela « Suisse saxonne »

Les changements dans l’utilisation et la distributionspatiale des terres influencent directement le proces-sus du drainage superficiel et donc l’érosion du sol.Néanmoins, les effets de ces changements ne sont gé-néralement pas pris en compte dans la modélisationnumérique de l’érosion du sol. Ainsi, l’importance desprocessus liés à l’utilisation des terres est sous-estiméedans ces modèles. Dans ce papier nous proposons uneméthode empirique basée sur l’algorithme RUSLE(équation générale pour l’érosion du sol) qui tientcompte des changements d’utilisation des terres pourla modélisation de l’érosion des sols. De plus, cetteméthode inclut une évaluation des approches existan-tes pour calculer la longueur érosive d’une pente. Cetteméthode prend également en compte le rôle des barriè-res de flux qui limitent l’infiltration pour mieux antici-per les processus d’érosion des sols. Une étude de casréalisée dans la région du parc national de la « Suissesaxonne » (Allemagne) pour la période 1900 à 1992confirme notre hypothèse selon laquelle les changementsd’utilisation des sols modifient considérablement lalongueur des pentes sujettes à l’érosion et donc l’éro-sion des sols. Alors que l’érosion diminue lorsque l’uti-lisation agricole des terres diminue sur les pentes rai-des, d’autres changements dans la couverture terrestreont pu augmenté l’érosion. L’érosion à augmenté de41 % en moyenne sur les terres arables entre 1900 et1992 à cause des changements de l’utilisation des ter-

res. En particulier, l’érosion à augmenté après 1940 àcause de la réorganisation des paysans en coopérati-ves agricoles de production et le besoin d’intensifierla production sur de plus grandes unités et donc plusgrandes pentes. Les zones de Grande Vulnérabilité àl’Erosion sont d’autant plus sensibles aux change-ments d’utilisation des terres. En général, nos résul-tats indiquent que les effets des changements de l’uti-lisation des terres sur l’érosion du sol ont été sous-estimés jusqu’à présent et que ces phénomènes doi-vent être intégrés de façon plus précise dans la modé-lisation empirique de l’érosion du sol.

Dipl.-Geogr. Sebastian Wolf, ETH Zurich, Departmentof Agricultural and Food Sciences, Institute of PlantSciences, LFW A54.1, Universitaetsstr.2, 8092 Zurich,Switzerland, sebastian.wolf @ipw.agrl.ethz.ch

Dr. rer. nat. Ulrich Walz, Leibniz Institute of Ecologicaland Regional Development, Weberplatz 1, 01217 Dres-den, Germany, [email protected]

Prof. Dr. Arno Kleber, Dresden University of Tech-nology, Department of Geography, 01062 Dresden,Germany, [email protected]

Manuscript submitted: 26/01/2009Accepted for publication: 09/06/2009

Documents

Effects of Land-Cover Change on Soil Erosion in the Saxon …sewolf/PDF/Wolf-Walz-Kleber.2009... · 2009-11-30 · • Land-cover patterns – Erosive slope length – Soil erosion