Changes in Sediment Flux and Storage Within a Fluvial System: Some

HYDROLOGICAL PROCESSES Hydrol. Process. 17, 3321–3334 (2003) Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/hyp.1389

Changes in sediment ﬂux and storage within a ﬂuvial system: some examples from the Rhine catchment

Andreas Lang,1* Hans-Rudolf Bork,2 Rudiger¨ Mackel,¨ 3 Nicholas Preston,1 Jurgen¨ Wunderlich4 and Richard Dikau1 1 Geographisches Institut, Universität Bonn, Meckenheimer Allee 166, D-53115 Bonn, Germany 2 Okologie-Zentrum¨ Kiel, Christian-Albrechts-Universität zu Kiel, Schauenburger Str. 112, D-24118 Kiel, Germany 3 Institut für Physische Geographie, Albert-Ludwigs-Universität, Werderring 4, D-79085 Freiburg, Germany 4 Institut für Physische Geographie, Johann Wolfgang Goethe-Universität, D-60054 Frankfurt, Germany

Abstract: The Rhine river system can look back on a long history of human impact. Whereas significant anthropogenic changes to the river channel started only 200 years ago, the impacts of land use have been felt for more than 7500 years. Here, we review results from several case studies and show how land-use change and climate impacts have transformed the fluvial system. We focus on changes in sediment delivery pathways and slope–channel coupling, and show that these vary in time and depend on the magnitude of a rainfall event. These changes must be accounted for when trying to use sediments as archives of land-use change and climatic impacts on fluvial systems. For example, human impact is recorded in slope deposits only as long as rainfall intensity and runoff generation do not exceed the threshold for gullying. Similarly, climatic impacts are only recorded in alluvium when both the landscape is rendered sensitive by human activities (e.g. deforestation) and rainfall thresholds for gullying are exceeded. Copyright  2003 John Wiley & Sons, Ltd.

KEY WORDS Rhine river; land-use change; climate change; sediment delivery; system evolution

INTRODUCTION The River Rhine drains large parts (189 700 km2 of central Europe. The main river channel stretches 1320 km and drains into the North Sea. Along its course the hydrological regime of the Rhine changes from glacio-nival in the Alps and upland areas to pluvially dominated lowlands in the Netherlands. Here, the mean discharge is 2500 m3 s1, the mean flood discharge is 6000 m3 s1 and the mean discharge at low flow is 1000 m3 s1 (IHP/OHP, 1996). A detailed description of the physiographic setting and the present-day hydrology of the Rhine is presented by Middelkoop and Asselmann (2003). The Rhine river system has probably been studied in greater detail than any other drainage basin in the world. A simple literature search returned more than 800 scientific articles and books, not including publications dealing with the cultural and political importance of the Rhine. Agricultural activities in the Rhine drainage basin date back to the Early Neolithic (¾7500 years ago). The loess landscapes of northern Switzerland, southern Germany, and France were especially favourable for settlement, due to their fertile soils and relatively mild climate. By medieval times the whole Rhine catchment had been settled, with only a few exceptions in remote mountain environments. Today, the Rhine catchment can be characterized as ‘advanced industrial’ according to Wasson (1996). Until the beginning of the 19th century the channel dynamics of the larger rivers within the Rhine drainage basin were dominantly climatically driven. While changes in land use resulted in strongly increased

* Correspondence to: Andreas Lang, Department of Geography, University of Liverpool, Liverpool L69 72T, UK. E-mail: [email protected]

Copyright  2003 John Wiley & Sons, Ltd. 3322 A. LANG ET AL. sedimentation (floodloams), the river form itself remained largely unchanged. The impact of anthropogenic structures, like water-mills and small dams, can be traced back to the Iron Age. Their influence, however, seems to have been only local. During the 19th century more dramatic changes were initiated by engineering work to permit the passage of ships, hydro-electric power plants, and flood protection. The most prominent of these changes was the channelizing of the course of the Rhine river in the Upper Rhine graben, starting with the work of Tulla in the early 19th century. In addition to these major impacts, the combination of increasingly numerous small impacts (e.g. surface sealing due to construction of roads and buildings, draining of wetlands and channelizing of smaller creeks and ditches) began to change the fluvial regime of the Rhine drastically. However, influences on the fluvial regime have not been restricted to engineering works. Many of the high flood events recorded during the Middles Ages and early modern times were related to ice dams and their failure. Owing to the impact of power plants using river water for cooling and releasing warm water back into the river, and the impact of urban waste water, ice has been absent from the larger rivers during recent decades. The anthropogenic impact of large engineering structures is rather obvious compared with the impacts of land use. Over the longer term, however, land use and land-use change also produce significant changes through the summed effect of small but more frequent events. Here, we report results of several case studies from the German part of the Rhine catchment, exemplifying the long-term development of a fluvial system undergoing human impact, and with these results we address some of the LUCIFS research questions.

CASE STUDIES Upper Rhine lowlands The first case study is located between the towns of Neuenburg and Offenburg in the southern part of the Rhine graben (Figures 1 and 2). The Upper Rhine graben system is a dominant tectonic feature in the course of the Rhine. The graben is filled with Tertiary and Quaternary sediments to depths of up to several hundreds of metres. During the Wurm¨ Glacial, the Rhine was a braided river system, fully in accordance with the periglacial climate conditions. At least one channel was located east of the Kaiserstuhl, which is an upland area within the alluvial flood plain of the Rhine river. At the beginning of the Late Glacial period the fluvial system changed and a meandering river developed. This change was associated with a decrease of gravel load and higher and more continuous runoff. The new channels incised into the Wurmian¨ gravels and a distinct erosional scarp was formed (Hochgestade), which today still forms a pronounced boundary between the higher Wurmian¨ terrace (Niederterasse) and the lower Holocene alluvial plain. According to pollen analysis and 14C data, the eastern course of the Rhine existed as a perennial stream until the end of the Late Glacial (Friedmann, 1999). However, the eastern Rhine channels were also partially reactivated during the Holocene in response to exceptional floods. The Late Glacial environment was largely unaffected by human activity, and changes in the fluvial system were caused by climatic changes. During the Atlantic Period and the Holocene climatic optimum (warmer temperatures, especially during summer), the Neolithic Revolution introduced widespread human impacts on the natural environment for the first time. The first settlements and agricultural activities in the southern Upper Rhine Valley are reflected in pollen diagrams, as for example in the Wasenweiler Ried (Figure 2). A period of geomorphic stability during the early Atlantic Period (about 8000 years ago) is indicated by the development of forests (mixed oak and hazel) and the black lowland (or floodplain) soil (Schneider, 2000). About 6000 years ago the situation changed: pollen from cereals and agricultural weeds appears (Friedmann, 1999). The pollen records show declining percentages of woodland pollen, whereas percentages of pollen indicating open land increase. This is due to woodland grazing and the first existence of range land, as can also be seen by the increase in grass pollen. The earliest occurrence of soil erosion in the loess areas took place during this period, and led to the accumulation of colluvial and fluvial sediments (Mackel¨ and

7°E9°E Elevation / m a.s.l Sieg < 220 Vogels- 220-340 3 berg 340-450 Lahn 450-570 Frankfurt Mosel Rhein 570-730 730-950 Main 50°N 950-1280 >1280

Mannheim Neckar Nürnberg

2 Stuttgart

Upper Rhine Graben Donau München

48°N Rhein 1 Schwäbische Alb

Schwarzwald N Bodensee Thur W E Rhein Alps S Rhein

0 20 40 60 80 km 1 Case study ‘Upper Rhine Lowlands’

Town 2 Case study ‘Vaihingen/Enz’

River Rhein 3 Case study ‘Amöneburger Becken’

Figure 1. Map of southern central Europe showing the location of the study sites

Friedmann, 1999; Schneider et al., 2000). A more noticeable human impact on the environment is documented from the Iron Age: the La Tene` culture (4th to 1st centuries BC) and the Roman period (1st to 4th centuries AD). A further decrease in woodland and an increase in open land during this period is apparent in pollen diagrams from the Upper Rhine Lowlands. Intensive agricultural land use can clearly be seen by an increase in pollen of cereals, weeds and pasture grasses. The growing human population resulted in an increase in the number of settlements and an extension of the farming area. In addition, mining activities caused a further reduction of the woodlands. For this period, higher rates of erosion are clearly documented in fluvial sediments and changed morphodynamic responses of the river systems. With the retreat of the Romans, the population density declined and many settlements and agricultural fields were abandoned. A significant impact on the environment of the climatically favoured and fertile loess-covered areas of the Upper Rhine Lowlands occurred once again during the Alamannic land acquisition and consolidation phase (5th to 7th centuries AD), as documented in alluvial sediments from the Elz and Rhine channels (Schneider, 2000). Widespread alluvial sedimentation in the Rhine tributaries occurred as a result of both mining activities and medieval monastic and

Figure 2. Map of Holocene river channels in the southern part of the Rhine graben. The eastern Rhine channels of Rothausen (1), Grezhausen (2) and Rimsingen (3), and the location of the present-day and deserted settlements are indicated. ‘Tuniberg’ and ‘Kaiserstuhl’ are prominent hills within the alluvial plain. (adapted from Mackel¨ and Zollinger (1995))

manorial colonization during the medieval climatic optimum (10th to 12th centuries AD)(Mackel,¨ 1998). For the end of the Middle Ages and the beginning of the Modern Period (14th to 16th centuries AD), sedimentary evidence indicates a change to wetter and cooler climatic conditions. More frequent ﬂoods in the area of the eastern Rhine led to the abandonment of villages (see Figure 2). During this time many of the eastern Rhine channels were reactivated, with the exception of those to the east of the Kaiserstuhl, which were impeded by the interim development of the Holocene fan of the Dreisam (Mackel¨ and Zollinger, 1995). High-magnitude ﬂoods in the Upper Rhine Valley during the last three centuries have also caused reactivation of streams of the eastern Rhine system south of the Kaiserstuhl and the Elz north of the Kaiserstuhl.

Loess hill-country The second case study is located in a tributary catchment of the Rhine river. Large parts of the Rhine drainage basin are covered by loess deposits. Most of these regions can be described as rolling hill country, the so called ‘Gau’¨ areas. In these areas, anthropogenic soil erosion has led to extreme truncation of soil profiles and deposition of thick colluvial and alluvial sediments. Colluvial sediments accumulated on the lower slopes have proved to be valuable archives for studying man–landscape interactions over the period of agriculture. The sediments are deposited close to their source areas, so interpretation of results is generally straightforward. However, when looking in detail at such deposits, temporal variation in sediment delivery pathways is obvious. Results obtained at an Early Neolithic settlement in Vaihingen/Enz near Stuttgart serve as an example (Figure 3a; Lang and Honscheidt,¨ 1999). Dating is based on optical techniques, on artefacts and on 14C dating of organic remains incorporated in the sediments. Chronological data allow reconstruction of the depositional history of the colluvium and also the identification of temporary sedimentary sinks along transport pathways. Radiocarbon dates the death of an organism, so it determines the time when organic objects first enter the erosion–transportation–deposition pathway. If final deposition follows quickly thereafter, then the 14Cage provides a close approximation to the time of sediment deposition. In many colluvial environments, however, this is not the case, because reworking of older colluvial sediments occurs. The time of reworking can, however, be estimated by optical dating techniques (Aitken, 1998; Lang et al., 1998). Lang and Honscheidt¨ (1999) have developed a cascade model of colluvium formation (Figure 3b). Colluvial sediments resulting from early soil erosion in the Neolithic to Iron Age periods were mainly deposited on the upper slopes. Significant deposition on the lower slope occurred for the first time during the Iron Age and Roman period. Since then, deposition rates at the lower slope have increased because of more intensive land use. Results from another case study show the interfingering of colluvial sediments with floodplain and wetland deposits. The study site is located in the Amoeneburger Becken (Figure 1), a tectonic basin mainly developed in Tertiary sediments and covered by loess of varying thickness. The area is drained by the Ohm, a tributary of the Lahn, and the Rhine. As in the other loess areas, fertile soils and favourable climatic conditions enabled early settlement and cultivation, as can be seen from archaeological evidence. Results are described in detail by Wunderlich (2000). In brief, a sequence of well-stratified colluvial deposits was recovered by excavation and drilling at the western edge of the Ohm floodplain and on the adjacent slope. The site was chosen close to an archaeological site at which remnants from several settlement phases, extending from the Early Neolithic Bandceramic period to the Middle Ages, were unearthed. An overview of the results is given in Figure 4. Two time slices of the basin’s development are shown in Figure 5. Remnants of luvisols developed in loess were found on the slopes. The degree of truncation of the soil profiles varies with position on the slope. On the upper slopes, soils have been totally stripped and calcareous loess occurs at the surface. At the transition from the lower slope to the floodplain a thick sequence of colluvial deposits occurs. Colluvia interfinger with organic and calcareous muds and peat that have developed at the edge of the floodplain. The majority of the colluvial deposits were laid down in these wetland environments. Most of the material eroded from upslope was trapped here and only a negligible amount was transported to the Ohm river. A chronology of the sediments was obtained by radiocarbon dating of the organic layers and macro-remains buried in the lower part of the colluvial sediments, as well as OSL dating of colluvial sediments from above

0 IR-OSL-age cal. 14C-age 40 potsherd

120 colluvium depth [cm] 160

200 P h

0 1 2 3 4 5 6 78fA age [ka]

Time of first incorporation into the down slope Early Neolithic times sedimentary pathway (e.g.14C-age): ~ 1 ka ~ 5 ka

~ 2 ka ~ 7 ka

Transportation process and resetting source sink 1 sink 2 sink 3 of the OSL-‘clock’ source sink 1 sink 2 sink 3

Late Neolithic times Iron Age / Roman times Early Middle Ages

source sink 1 sink 2 sink 3 source sink 1 sink 2 sink 3 source sink 1 sink 2 sink 3

Late Middle Ages Modern Times today

source sink 1 sink 2 sink 3 source sink 1 sink 2 sink 3 source sink 1 sink 2 sink 3

Copyright  2003 John Wiley & Sons, Ltd. Hydrol. Process. 17, 3321–3334 (2003) CHANGES IN SEDIMENT FLUX AND STORAGE 3327 the groundwater table. The local vegetation history was reconstructed on the basis of plant macro-remains and molluscs (Rittweger, 1999) and by pollen analysis (Bos, 1998). Very few colluvial deposits originate from the Early Neolithic period (Bandceramic), despite the archaeological evidence for extensive settlement activity at the site and the mollusc record indicating forest clearance. The oldest signiﬁcant colluvium was laid down during the Late Neolithic. Part of these sediments may also originate from a failure of the steep slope edge of the ﬂoodplain. During the Bronze Age and the Early Iron Age a dark grey to black colluvium accumulated. Much higher amounts of colluvium were deposited during the Late Iron Age (black) and the Roman period (dark brown). The palaeobotanical record and mollusc analysis show that the black colluvia correspond to large-scale clearing of woodlands and intense land use, including tillage and grassland cultivation. However, the majority of the sediment body formed during medieval times. A more general regional picture of colluvium formation is shown in Figure 6 (Lang, 2003). Optical ages obtained from 60 colluvial sediments in southern Germany are plotted as frequency distributions. Periods of colluvium formation roughly coincide with periods of strong human impact on the environment. Climatic changes seem to play only a minor role.

Landscape dynamics in Germany during the past 1350 years The population density in Germany was reduced drastically during the Dark Ages as a result of disease (e.g. bubonic plague), cold and wet weather conditions resulting in crop failures, famines, ecological catastrophes and migrations from central to southern Europe. Many settlements from the Roman period were abandoned and forest returned to the former agricultural land. By the middle of the 6th century AD, around 90% of Germany was again covered with nearly natural woodland (Bork et al., 1998a). In addition to evidence from historical documents, this can also be seen in a period of rather intense soil formation. In early medieval times, areas with fertile soils and climatic conditions favourable for agriculture were again cleared for agricultural purposes. When population density increased later, the hilly regions in southern, western and central Germany, as well as the lowlands of northern and northeastern Germany, were also cleared and used for agriculture. By the early 14th century, the area covered by woodland had been reduced to only 15% (Figure 7). The effects of land-use change in Germany during the past 1350 years have been modelled with climatic conditions held constant (Bork et al., 1998a). The drastic decrease in total biomass from AD 650 to AD 1300 reduced evapotranspiration (20%) and considerably raised total runoff (C60%). These effects are confirmed by field data. Besides the changes in the average water balance, an increased number of extreme rainfall events characterized the first half of the 14th century, as clearly documented in sediment successions (Bork et al., 1998a). During the first half of the 14th century, widespread colluviation and fan development commenced. The ploughing horizons of fields not densely covered with summer crops were frequently totally eroded. Half of the total hillslope erosion since AD 650 occurred between AD 1310 and 1350. Where shallow fertile soils above stony layers were eroded completely the areas were abandoned, and in many cases they have remained as woodland since that time. Where infertile sands were exposed by soil erosion (Bork et al., 1998b;

Figure 3. Adapted from Lang and Honscheidt¨ (1999). Upper part: age versus depth plot for a colluvium from Vaihingen–Enz. Optically Stimulated Luminescence (OSL) ages and calibrated AMS 14Cages(1 confidence level) of organic remains are plotted against the depth of sampling. The archaeological age of a ceramic fragment is also plotted. The ages of the strata imply that most of the material was brought to its present location not in a single event, but after several pulses of transport with intervening periods of storage. Lower part: model of colluvium formation at Vaihingen–Enz. The first erosion occurred as early as Neolithic times (about 7 ka ago), partly filling sedimentary traps on the upper slope. As soil erosion proceeded ca 5-ka-old material (Late Neolithic) became trapped in the sinks (indicated by the 3030–2700 BC and 2870–2510 BC organic remnants). During the Iron Age/Roman period (around 2Ð5–2 ka ago) upslope sinks were filled (indicated by the 520–370 BC and 760–390 BC organic remnants). During this period, eroded material was for the first time transported all the way down to the lower slope (indicated by the 2170 š 170 years old colluvium). When erosion occurred on the crest and the upper slopes, the erosion of sediment trapped in the depressions (sinks 1 and 2) also started. Around 1Ð5 ka ago the sediment was incorporated in the colluvium that had entered the depositional pathway 0Ð5 to 3 ka earlier. Around 0Ð5 ka ago, Neolithic material was deposited on the lower slope, covering sediments deposited there ca 1 ka ago

3328 A. LANG ET AL. Biostratigraphy AT SB PB SA YD YD BO YD? Hiatus NE 1 2 5 4 3 [m] 1300-980 9880-9520 12130-11730 10300-10040 14610-13390 13410-13160 12590-12370 nterfingering with organic 1 2 5 6 4 3 biostratigraphy, based on pollen [m] 150 m 1340-1260 12800-12580 NO 125 m 1 6 2 5 4 3 [m] 6 m m 20 howing stratified colluvial deposits i 1.3 ± 0.2 2.1 ± 0.3 1.2 ± 0.1 2.0 ± 0.2 0.63 ± 0.07 lly, log plots of adjacent drillings and the local 2790-2750 3350-3260 4530-4420 6660-6500 11010-10990 15 3820-3630 8940-8560 9850-9540 moeneburger Becken. Cross-section s 4570-4410 3690-3620 <2.5 ± 0.3 Organic mud, peat mud, Organic Silty mud Calcareous mud Loess, loess loam Loess, Sand Molluscs 7390-7290 8480-8380 4500-4410 4430-4310 3690-3630 analysis (Bos, 1999), are given. Adapted from Wunderlich (2000) 2.0 ± 0.2 2.4 ± 0.3 2.4 ± 0.3 4.3 ± 0.4 10040-9980 10300-10050 tted according to sampling positions. Additiona 9250-9000 10070-10000 510 44 ± 6 C age range (cal BP) 14 IRSL age (ka) Calibrated Colluvium (period of deposition) Times Middle Ages/Modern Roman Period Iron Age (Hallstatt/Latène) Age/Iron Age (Hallstatt) Bronze Younger Early/Middle Age Bronze Late Neolithic Early Neolithic 0 SW 2.0 ± 0.2 1 4430-4310 0 2 3 4m Figure 4. Soils and sediments recovered at Mardorf–Schweinsberg in the A floodplain deposits. Chronometric results are plo

0 SW NE

Floodplain 5

present day 10 m

Atlantic Period 10 m (Early Neolithic)

0 10 20 30 40 50 m

Loess/ loess loam stagnant water Flood loam Ah prehistoric Peat Al Soil horizons structures Bt Silty mud Bv Colluvium Sand, gravel location of present day surface (lower plot) Figure 5. Two time slices of Holocene landscape development. Schematic SW–NE cross-sections at Mardorf–Schweinsberg for the present and for the Mid-Holocene. Adapted from Wunderlich (2000)

Schatz, 2000), agricultural land use ceased until soils enriched in organic matter were newly developed under woodland. In dells and furrows of sparsely vegetated fields (e.g. in ridge and furrow areas) deep U- or V-shaped gullies were formed. Some gully systems achieved depths of more than 8 m, widths of several decameters and lengths of several hundreds of metres to some kilometres. In some areas the development of extended badlands precluded further agricultural use. The formation of these extended gully systems is attributed, on the basis of detailed analysis of erosional forms, their sedimentary fills and fan sediments, to the occurrence of one (or very few) catastrophic overland flow event(s) (Bork, 1988; Bork et al., 1998a). Where forests returned, these gullies are still present today. In areas subsequently used for farming, the gullies were rapidly filled over periods of a few decades. Extreme weather events, famines, runoff, and floods during the second decade of the 14th century and in July 1342 are reported in contemporary documents. In July 1342 a 1000 year rainfall event was widespread throughout central Europe (Flohn, 1949–50, 1958, 1967; Pfister, 1980, 1985; Alexandre, 1987; Lamb, 1997). River levels for the period from the 19th to 25th July 1342 are by far the highest ever recorded at several sites (Weikinn, 1958; Alexandre, 1987). Most stone bridges over the major rivers were destroyed. In July 1342 the overland flow rates in the catchments of the major central European rivers exceeded 20th century maxima by factors of 50 to 200 (Bork et al., 1998a)! As a result of these catastrophes, and because of the Black Death during the years 1348–50, more than a third of the population died. The woodland area increased by three times between the mid 14th century and the late 15th century. Thus, average rates of transpiration increased and runoff decreased. Soil erosion again was of minor importance in most German landscapes until the mid 18th century. Population density increased and woodland area decreased again during the 16th and 17th centuries. In the late 16th century, a third of the land surface of present-day Germany was covered with woodland—most of it grazed. The size of the forest areas has not changed much since then, although the grazing intensity has been lowered in German forests since the 19th century. Soil erosion rates increased again in many German landscapes during the fifth decade of the 18th century and in others a few decades later. Severe gullying was

Figure 6. Frequency analysis of OSL ages obtained for soil-erosion-derived sediments from several different sites in the loess hills of south Germany. OSL ages are plotted as Gaussian curves and the area below the curves set to one. The individual curves were then summed and the resulting distribution plotted. The inset gives an enlargement for the period 0Ð8ka–3Ð0 ka. Only the number of ages is used for analysis, and the volume of sediment deposited is not considered. Nevertheless, phases of colluviation clearly coincide with phases of strong human impact. The first colluvial sediments were deposited during early Neolithic times. A second small maximum in the distribution occurs towards the end of the Neolithic period. Many colluvia originate in the Iron Age and Roman periods, and the maximum number of optical ages relates to medieval times (adapted from Lang (2003)) common until the end of the 18th century and in some areas until the second or third decade of the 19th century. Based on soil and sediment analysis and on contemporary documents, the occurrence of gullying can clearly be linked to an increased number of rainstorms, although field size and the cropping sequence (namely the presence of fallow land) were also of importance. Hillslope erosion increased by an order of magnitude during the second half of the 18th century and was recognized as a severe problem. After one and a half centuries of low hillslope erosion and an absence of gullying, soil erosion rates increased again significantly in the sixth and seventh decades of the 20th century. On average, rates tripled due to changed crop sequences, increased field sizes and further mechanization. Today, agricultural subsidies and the world market determine crop selection more than site characteristics, and crops with a low vegetation cover density are common in the erosive early summer months (Dikau, 1986). Field areas slightly increased as a result of the rationalization of land tenure. Soil conservation measures, such as terraces and hedges that had existed since the last period of intensive soil erosion (the late 18th century), were removed. The development of large and heavy agricultural equipment led to soil compaction, and thus to reduced infiltration capacities and increased soil erosion.

DISCUSSION The transition to the Holocene caused a change in the ﬂuvial regime of the Rhine. Subsequent human impacts are reﬂected in phases of both colluvial and alluvial sedimentation. Although episodes of sediment generation are broadly consistent with periods of greater human activity, there are slope–channel coupling issues that

100 100

arable land 80 10

60 grass land, fallow land 1 wood land

40 land use %

0.1

average soil erosion [mm/a] average 20

0.01 0 2000 1800 1600 1400 1200 1000 800 600 calendar year a.d.

Soil erosion

Figure 7. Soil erosion and land use in Germany (excluding the Alps) during the last 1400 years (adapted from Bork et al. (1998a)). The average soil erosion in Germany (thick line, left axis, log scale) and the percentages of woodland, fallow and arable land (right axis) are plotted on a calendar scale. Extreme rainfall events during the 14th century coincided with high percentages of both arable and fallow land and caused extreme soil loss mean that the impacts of these sediment-generation phases on the fluvial system are not straightforward. It may be the case that land-use impacts are only expressed in the fluvial system through the agency of climatic events. Climatic events have reactivated channels and reworked both colluvial and alluvial sediments, and determining the relative roles of land use and climate in the Rhine system requires further research. As illustrated by the cases studies, quite large temporal and spatial heterogeneity of responses to land use and climatic forcing exists within the Rhine drainage basin. To account for some of this heterogeneity, a conceptual model of changes in slope–channel coupling during the period of agriculture is outlined below. The geomorphic impact of climatic events, i.e. the mobilization of sediment, is dependent not only on the magnitude of the rainfall event, but also on the sensitivity of the landscape. The sensitivity of the landscape is mainly dependent on the type of agriculture. During the Early Neolithic period, farming was restricted to highly fertile and easy manageable loess soils—which are also highly erodible. Sediment derived from soil erosion did not reach the floodplains and channels in most cases, but was re-deposited on the slopes (Figure 8). The same pattern can also be seen for the Late Neolithic period, although somewhat intensified. During the Bronze Age and Early Iron Age, sediments eroded from the slopes were deposited on the lower slopes for the first time. During the subsequent land-use phases—the pre-Roman Iron Age and the Roman period—the first dramatic soil erosion occurred. This is clearly documented by the widespread occurrence of colluvial and alluvial sediments. Reworking of older colluvial sediments on the slopes occurred during this period. During the medieval period, agricultural activity was widespread and intense for the first time, and thus the sensitivity of the landscape was for the first time high enough to allow for the occurrence of gullying. During extreme rainfall events, e.g. during the 14th and 18th centuries, deeply incised gullies developed. Land degradation during these periods in some regions was so strong that farming activities were no longer possible. In these areas, forest returned and the gullies were preserved. In the more favourable or less-impacted areas, gullies were subsequently filled by soil-erosion-derived sediments and are totally levelled out today. Further, the majority of soil-erosion-derived sediments within the fluvial system apparently originated in this period, with soil erosion during modern times seemingly not producing similar rates of sedimentation. However, it is

Slope-channel coupling

Early Neolithic Period (I) Late Neolithic Period - Bronze Age (II) - first agriculture - agricultural areas extended - only local occurrence - still only small areas - sediment re-destribution - sediment re-destribution limited to the slopes still limited to the slopes

Upper slope Mid-slope Lower slope Floodplain Upper slope Mid-slope Lower slope Floodplain

Iron Age - Roman Period (III) Medieval Period (IV) - widespread deforestation - larger clear cut areas - high sedimentation rates - first deposition on lower - erosion of Neolithic colluvium slopes and on flood-plains

Upper slope Mid-slope Lower slope Floodplain Upper slope Mid-slope Lower slope Floodplain

14th century? (V)

Extreme events Gully

Upper slope Mid-slope Lower slope Floodplain

Modern Times (VI) Modern Times (VII) Forest Field

- Gullies filled up - Gullies preserved - landscape levelled

Upper slope Mid-slope Lower slope Floodplain Upper slope Mid-slope Lower slope Floodplain

Figure 8. Conceptual model of changes in slope–channel coupling during the period of agriculture in central Europe

Copyright  2003 John Wiley & Sons, Ltd. Hydrol. Process. 17, 3321–3334 (2003) CHANGES IN SEDIMENT FLUX AND STORAGE 3333 important to interpret this pattern both in terms of the relative frequency/magnitude of climatic and land-use perturbations to the landscape system, and in a spatial context within the drainage hierarchy. Preston (2001) has demonstrated that continuous agricultural land use has been able to remove the effects of high-magnitude climatic events from headwater basins, reducing drainage density by filling gullies and first-order channels. On the other hand, the effects of climatic events tend to dominate higher order systems, where even smaller rainfall/runoff events have sufficient energy to transport sediment and the continuous soil erosion processes are incapable of masking the morphological responses to climatic impacts. Thus, over the longer term, it seems that agricultural land use has a greater geomorphic effect at small landscape scales, but when the whole fluvial system is considered then the sediment flux is dominated by climatic events. It must be stated, however, that this is a first and preliminary assessment of the impacts of climate and land use within the Rhine system. Future work planned within the Rhein–LUCIFS research framework is aimed at obtaining a better definition of the sources and sinks of particulate sediment, carbon, nitrogen, and phosphorus, determining sediment budgets for the agricultural period, establishing chronological frameworks for these, and quantifying fluxes in the framework of numerical flux models. Comparisons with independent records established for climatic change and land-use change (e.g. as given by Glaser (2001) and Burggraaff (1992)) within a hierarchic modelling framework (Preston and Schmidt, 2003) should facilitate quantitative determination of fluvial system responses to external forces.

CONCLUSIONS In the Rhine river catchment, changes in the fluvial system during the Holocene seem to be dominated by changing human impact. Clearing of woodland and agricultural activities made the drainage basin susceptible to soil erosion by water. These changes influenced the water balance and runoff production. River dynamics changed dramatically due to the high supply of fine sediment mobilized by soil erosion. Today, smaller valleys are filled with several metres of flood loam—a sediment that has been developed mainly since Roman times. The results from the case studies presented clearly show that changes in sediment redistribution similar to those operating over the decadal time scale (Trimble, 1999) also operate over the much longer time spans of centuries and millennia. At the moment we cannot give quantitative estimates. However, it is, clear that, when studying human impact on the landscape, it must be recognized that sediment pathways change through time and also depend on the magnitude of rainfall events. Human impact on the environment, for example, will only be recorded in the slope deposits as long as the climatic impact (rainfall) does not exceed the threshold for gullying, as a result of which the sediment may be redistributed and form part of the alluvial sediment archive in the fluvial system. On the other hand, climatic impacts are only recorded in alluvium if the landscape is sufficiently sensitive (e.g. as influenced by deforestation and human land use) and the rainfall threshold for initiation of large rills or gullies is exceeded. Sediment transport to the channels generally only takes place through linear forms (gullies) that require high-magnitude rainfall events both for their formation and to provide a transport medium. At the same time, gullies and deeper rills incise into older sediments, thus destroying parts of the earlier record. Focusing only on either colluvial or alluvial sediments, therefore, will only provide part of the picture of the impacts of climate and land use on fluvial systems.

ACKNOWLEDGEMENTS The Deutsche Forschungsgemeinschaft is acknowledged for ﬁnancial support.

REFERENCES Aitken MJ. 1998. An Introduction to Optical Dating. Oxford University Press: Oxford; 267. Alexandre P. 1987. Le climat en Europe au Moyen Age. Contribution a l’histoire des variations climatiques de 1000 a 1425, d’apres les sources narratives de l’Europe occidentale. Recherches d’histoire et de sciences sociales 24. Ecole des Hautes Etudes en Sciences Sociales: Paris; 828.

Bork HR. 1988. Bodenerosion und Umwelt—Verlauf, Ursachen und Folgen der mittelalterlichen und neuzeitlichen Bodenerosion. Boden- erosionsprozesse. Modelle und Simulationen. Landschaftsgenese und Landschaftsokologie¨ 13. Selbstverlag Abteilung fuer Physische Geographie und Landschaffsoekologie: Braunschweig; 249. Bork HR, Bork H, Dalchow C, Faust B, Piorr HP, Schatz T. 1998a. Landschaftsentwicklung in Mitteleuropa. Wirkungen des Menschen auf Landschaften. Klett-Perthes: Gotha; 328. Bork HR, Dalchow C, Dotterweich M, Schatz T, Schmidtchen G. 1998b. Die Entwicklung der Landschaften Brandenburgs in den vergan- genen Jahrtausenden. In Geschichte der Landwirtschaft Brandenburgs, Klemm V, Darkow G, Bork H-R (eds). Mezogazda:´ Budapest; 237–258. Bos JAA. 1998. Aspects of Lateglacial–Early Holocene Vegetation Development in western Europe, palynological and palaeobotanical investigations in Brabant (The Netherlands) and Hessen (Germany). LPP Contribution Series 10, Utrecht; 240. Burggraaff P. 1992. Kulturlandschaftswandel am unteren Niederrhein seit 1150. Geschichtlicher Atlas der Rheinlande: maps IV.7.1 and IV.7.1, Publikationen der Gesellschaft fur¨ Rheinisch Geschichtskunde, XII. Abteilung Ib Neue Folge: Koln.¨ Dikau R. 1986. Experimentelle Untersuchungen zu Oberflächenabfluß und Bodenabtrag von Meßparzellen und landwirtschaftlichen Nutzflächen. Heidelberger Geographische Arbeiten 81. Selbstverlag Geographisches Institut: Heidelberg; 195. Flohn H. 1949/50. Klimaschwankungen im Mittelalter und ihre historisch-geographischen Bedeutung. Berichte zur Deutschen Landeskunde 7: 347–358. Flohn H. 1958. Klimaschwankungen der letzten 1000 Jahre und ihre geophysikalischen Ursachen. In Deutscher Geographentag Würzburg, Tagungsberichte und wissenschaftliche Abhandlungen,Budel¨ J, Mensching H (eds). Steiner-Verlag: Wiesbaden; 201–214. Flohn H. 1967. Klimaschwankungen in historischer Zeit. In Die Schwankungen und Pendelungen des Klimas seit Beginn der regelmäßigen Instrumenten-Beobachtung, von Rudloff H (ed). Vieweg: Braunschweig; 81–90. Friedmann A. 1999. Die spät- und postglaziale Landschafts- und Vegetationsgeschichte des südlichen Oberrheintieflands und Schwarzwalds. Habilitationsschrift Geowissenschaftliche Fakultat¨ Universitat¨ Freiburg; 231. Glaser R. 2001. Klimageschichte Mitteleuropas. 1000 Jahre Wetter, Klima, Katastrophen. Primus: Darmstadt; 240. IHP/OHP. 1996. The River Rhine—Development and Management. IHP/OHP-Berichte 9. IHP/OHP: Koblenz; 100. Lamb HH. 1997. Climate, History and the Modern World. Routledge: London; 433. Lang A. 2003. Phases of soil erosion derived colluviation in the loess hills of south Germany. Catena 51: 209–221. Lang A, Honscheidt¨ S. 1999. Age and source of soil erosion derived colluvial sediments at Vaihingen-Enz, Germany. Catena 38: 89–107. Lang A, Nolte S. 1999. The chronology of Holocene sediments from the Wetterau, Germany, provided by optical- and 14C-dating. The Holocene 9: 207–214. Lang A, Rieser U, Habermann J, Wagner GA. 1998. Luminescence dating of sediments. Naturwissenschaften 85: 515–523. Mackel¨ R. 1998. Flußaktivitat¨ und Talgeschichte des Spat-¨ und Postglazials im Oberrheintiefland und Schwarzwald. In Wandel der Geo- Biosphäre in den letzten 15000 Jahren im südlichen Oberrheintiefland und Schwarzwald,Mackel¨ R, Friedmann A (eds). Freiburger Geographische Hefte 54. Selbstverlag des Institutes fuer Physische Geographie: Freiburg; 31–50. Mackel¨ R, Friedmann A. 1999. Holozaner¨ Landschaftswandel im sudlichen¨ Oberrheintiefland und Schwarzwald. Eiszeitalter und Gegenwart 49: 1–20. Mackel¨ R, Zollinger G. 1995. Holocene river and slope dynamics in the Black Forest and Upper Rhine Lowlands under the impact of man. Zeitschrift für Geomorphologie Neue Folge, Supplement Band 100: 89–100. Middlekoop H, Asselmann N. 2003. Impact of climate and land use change on river discharge and the production, transport and deposition of fine sediment in the Rhine basin—a summary of recent results. In Concepts and Approaches to Long Term and Large Scale Modelling of Fluvial Systems, Lang A, Hennrich K, Dikau R (eds). Lecture Notes in Earth Sciences 101. Springer-Verlag: Berlin, Heidelberg, New York; 143–167. Pfister C. 1980. Klimaschwankungen und Witterungsverhaltnisse¨ im schweizerischen Mittelland und Alpenvorland zur Zeit des “Little Ice Age”. In Das Klima, Analysen und Modelle, Geschichte und Zukunft, Oeschger H, Messerli B, Silvar M (eds). Springer-Verlag: Berlin, Heidelberg, New York; 175–190. Pfister C. 1985. Veranderungen¨ der Sommerwitterung im sudlichen¨ Mitteleuropa von 1270–1400 als Auftakt zum Gletscherhochstand der Neuzeit. Geographica Helvetica 4: 186–195. Preston NJ. 2001. Geomorphic response to environmental change: the imprint of deforestation and agricultural land use on the contemporary landscape of the Pleiser Hügelland, Bonn, Germany. PhD dissertation, University of Bonn. URL: http://hss.ulb.uni- bonn.de/ulb bonn/diss online/math nat fak/2001/0170/diss.pdf. Accessed: 27 May 2003. Preston NJ, Schmidt J. 2003. Modelling sediment fluxes at large spatial and temporal scales. In Concepts and Approaches to Long Term and Large Scale Modelling of Fluvial Systems, Lang A, Hennrich K, Dikau R (eds). Lecture Notes in Earth Sciences 101. Springer-Verlag: Berlin, Heidelberg, New York; 49–67. Rittweger H. 1999. Eine boreale–subboreale Molluskensukzession als Spiegel der Vegetationsgeschichte in der Ohmniederung bei Mar- burg/Lahn. Frankfurter Geowissenschaftliche Arbeiten 25: 13–23. Schatz T. 2000. Untersuchungen zur holozänen Landschaftsentwicklung Nordostdeutschlands. ZALF-Bericht 41, ZALF: Muncheberg;¨ 201. Schneider R. 2000. Landschafts- und Umweltgeschichte im Elz-Einzugsgebiet. PhD thesis University of Freiburg. Schneider R, Friedmann A, Mackel¨ R. 2000. Hangsedimente und Kolluvien in den Loßgebieten¨ Sudbadens.¨ Berichte der Naturforschenden Gesellschaft Freiburg im Breisgau 88–89: 1–15. Trimble SW. 1999. Decreased rates of alluvial sediment storage in the Coon Creek Basin, Wisconsin, 1975–93. Science 285: 1244–1246. Wasson RJ (ed). 1996. Land use and climate impacts on fluvial systems during the Period of Agriculture. Recommendations for a research project and its implementation. PAGES report, Series 96-2; 52. Weikinn C. 1958. Quellentexte zur Witterungsgeschichte Mitteleuropas von der Zeitwende bis zum Jahre 1850—Hydrographie, Zeitwende bis 1500 . Akademie Verlag: Berlin; 531. Wunderlich J. 2000. Prahistorische¨ und historische Bodenerosion im Amoneburger¨ Becken—Abgeleitet aus einer Sequenz datierter Kolluvien. Berichte der Kommission für Archäologische Landesforschung in Hessen 5: 9–15.