Netherlands Journal of Geosciences Geologie en Mijnbouw | 87 4 | 323 - 337 | 2008

The age and origin of the Gelderse IJssel

B. Makaske1,*, G.J. Maas1 & D.G. van Smeerdijk2

1 Alterra, Wageningen University and Research Centre, P.O. Box 47, 6700 AA Wageningen, the Netherlands.
2 BIAX Consult, Hogendijk 134, 1506 AL Zaandam, the Netherlands.
* Corresponding author. Email: bart.makaske@wur.nl

Manuscript received: February 2008; accepted: September 2008

Abstract

The Gelderse IJssel is the third major distributary of the Rhine in the Netherlands and diverts on average ~15% of the Rhine discharge northward.
Historic trading cities are located on the Gelderse IJssel and flourished in the late Middle Ages. Little is known about this river in the early
Middle Ages and before, and there is considerable debate on the age and origin of the Gelderse IJssel as a Rhine distributary. A small river
draining the surrounding Pleistocene uplands must have been present in the IJssel valley during most of the Holocene, but very diverse opinions
exist as to when this local river became connected to the Rhine system (and thereby to a vast hinterland), and whether this was human induced
or a natural process. We collected new AMS radiocarbon evidence on the timing of beginning overbank sedimentation along the lower reach of
the Gelderse IJssel. Our data indicate onset of overbank sedimentation at about 950 AD in this reach. We attribute this environmental change
to the establishment of a connection between the precursor of the IJssel and the Rhine system by avulsion. Analysis of previous conventional
radiocarbon dates from the upper IJssel floodplain yields that this avulsion may have started ~600 AD. Our results contradict earlier suppositions,
based on interpretation of archaeological data and historical accounts, that the Gelderse IJssel is much older and originated as a canal, dug under
supervision of the Roman general Drusus.

Keywords: avulsion, Drusus, Gelderse IJssel, overbank deposits, Rhine, radiocarbon age determination.

Introduction While the history of the IJssel during and after the late
Middle Ages is well known, the importance and development
The Gelderse IJssel is a major Rhine distributary (Fig. 1) that of the river in the early Middle Ages is still obscure. Especially
ranks third in size after the Waal and the Nederrijn-Lek, and the age and origin of the IJssel as a Rhine distributary has
on average discharges about 300 m3/s. Especially in the late been a subject of much discussion among historians, archaeolo-
Middle Ages, the Gelderse IJssel (henceforth called IJssel) was gists, geographers and geologists for more than a century.
important as a shipping route playing an important role in the Because of the scarcity of historical sources from this earlier
rise of Hansa cities like Kampen, Deventer, Zutphen and period, many of the arguments put forward in the discussion
Doesburg (Fig. 1). At that time, the IJssel was still a natural, rely on interpretations of archaeological finds and geological
freely meandering river. In the course of the 16th century AD data.
the IJssel started silting up, which ultimately led to engineering Geologically, the IJssel channel is scoured into the Late-
works upstream in the beginning of the 18th century AD to Pleistocene and Holocene fill of a valley of (largely) Saalian
improve water supply and navigation depth. At present, the glacial origin (Busschers et al., 2007). The Holocene floodplain
IJssel channel is trained with groynes and bank reinforce- is bordered by Pleistocene uplands (Fig. 1), comprising river
ments, and a system of waterworks cause diversion of a fixed terraces, aeolian (coversand) dunes and plains, and glacial ice-
portion (~15%) of the total Rhine discharge into the IJssel. pushed ridges with associated periglacial landforms. During

Netherlands Journal of Geosciences Geologie en Mijnbouw | 87 4 | 2008 323

 Fig. 1. Map of the IJssel valley showing
localities mentioned in the text, sampling
locations, the locations of Figs 2, 3 and 8,
and the location of the Roman castellum
referred to in the text. Distinction between
Holocene and Pleistocene deposits based
on the Geomorphological Map of the
Netherlands (Koomen & Maas, 2004).

most of the Holocene, a small river must have existed in the The objectives of this paper are: 1) to evaluate the existing
IJssel valley, fed by the Berkel and a few minor streams draining data on the age and origin of the IJssel, 2) to present new
the surrounding Pleistocene uplands. The question is when and evidence on the age of the IJssel, and 3) to discuss this new
how a connection between the Nederrijn (or a precursor Rhine evidence in relation to previous work and geomorphological
branch) and this small river in the IJssel valley came into observations.
existence.

324 Netherlands Journal of Geosciences Geologie en Mijnbouw | 87 4 | 2008

 Previous research Radiocarbon age determinations

Weichselian and early-Holocene palaeogeographical Numerous radiocarbon dates relating to the age of Rhine dis-
developments tributaries are available from the Rhine-Meuse delta (summarized
by Berendsen & Stouthamer, 2001). In general, the age of begin-
During the Early Weichselian, the Rhine flowed northward ning overbank sedimentation is taken to represent the age of
through the IJssel valley, entering it from the southeast, formation of the stratigraphically associated channel belt. For
roughly following the valley of the present Oude IJssel dating the beginning of overbank sedimentation, radiocarbon
southeast of Doetinchem (Van de Meene, 1979; Fig. 1). In the age determination is carried out of the top of a peat bed that is
Middle Weichselian (Middle Pleniglacial) the Rhine changed its non-erosively overlain by overbank deposits (e.g. Berendsen,
course westward, to the present Rhine-Meuse delta, near 1982, pp. 114-118). From five locations on the upper reach of the
Doesburg and abandoned its course through the lower IJssel IJssel (Fig. 1) radiocarbon age determinations from the top of a
valley (Busschers et al., 2007). In the Late Weichselian, the peat bed underlying overbank deposits are available (Table 1).
Rhine gradually abandoned its course through the Oude IJssel An often cited radiocarbon age determination suggesting a
valley. At that time, the Rhine entered the Netherlands Roman age of the IJssel is GrN-5491: 2000 65 14C yr BP
through its present more southerly course (in the Gelderse (Vogel & Waterbolk, 1972; Lanting & Mook, 1977; Teunissen,
Poort area; Fig. 1) (Busschers et al., 2007). Climatically induced 1980). The dated peat sample was collected in 1967 by J.N.B.
stepwise incision and associated terrace formation (e.g. Poelman in the Giesbeekse Broek area (location 1 in Fig. 1)
Berendsen et al., 1995) continued in the lower Rhine and Meuse near a borehole cross-section that was published by Poelman
valleys in the Late-Weichselian and Early-Holocene periods, & Harbers (1966). This cross-section reveals a strongly decom-
ultimately leading to relatively narrow incised channel belts posed, dark top of the dated peat bed. This dark top horizon is
with surface levels below wide abandoned Weichselian terraces discontinuous and of variable thickness (up to at least 30 cm
(Cohen, 2003; Gouw & Erkens, 2007) . (Poelman 1981b, p. 174)). The boundary between the sampled
Following abandonment by the Rhine, a local drainage dark peat and the overlying clay is sharp (Poelman 1981b, p.
system had established in the IJssel valley north of Doesburg. 174). These phenomena suggest that: 1) the top of the peat
A low divide, consisting of abandoned Middle-Weichselian has been eroded, at least in parts of the cross-section; 2) there
Rhine terraces, existed across the IJssel valley near Doesburg. has been a time gap between the end of peat-growth and the
We suppose (see section The IJssel avulsion) that this divide beginning of fluvial clay deposition during which the observed
remained intact during most of the Holocene, with the Oude soil formation in the peat could take place. Thus, the top of the
IJssel flowing from Doesburg to the southwest and joining the peat may be older than the beginning of overbank deposition
Rhine near Arnhem, thereby following the overall gradient of from the IJssel. Teunissen (1981, 1990) disputed the existence
the shallow Late-Weichselian subsurface. of a significant time gap based on his palynological analyses
Primarily under influence of rapid relative sea-level rise, and proposed that rapid peat decomposition could have resulted
aggradation of the incised Rhine and Meuse channel belts from the change to eutrophic circumstances with first clay
occurred in the Middle Holocene, starting in the west and deposition. Additional age determinations (carried out later)
gradually occurring more and more upstream. Approximately of material from various depths at the same location (Table 1)
4000 BP the upstream limit of onlap of Holocene deposits on show stratigraphic inconsistency: GrN-5491 is obviously too
Weichselian terraces had reached the area near Arnhem and old to fit into the series and should therefore be rejected.
Westervoort (Gouw & Erkens, 2007) (Fig. 1). Herewith, one From the Lathumse Broek (location 2 in Fig. 1), a younger
condition promoting a northward avulsion of the Rhine into date from the top of the peat bed was obtained: 1720 25 14C
the IJssel valley was satisfied: especially aggrading rivers with yr BP (GrN-7525). According to Poelman (1981b, p. 174), the
lateral floodplains are prone to avulsion. Another favourable top of the (slightly clayey) peat bed was dark and decomposed
condition is an energetic (usually gradient) advantage of and sharply overlain by overbank clay. Teunissen (1990, p. 100),
the avulsion route over the existing channel. Because the however, stated that his pollen diagram and organic matter
Westervoort-Doesburg reach of the IJssel valley dipped south- analyses indicate a gradual peat-clay transition, without a
westward (the direction of Late-Weichselian Rhine flow in this depositional hiatus. Because of the stratigraphic consistency
part of the IJssel valley), substantial additional aggradation in of the series of dates from this location (Table 1) GrN-7525 was
the area Arnhem-Westervoort was needed to reverse the accepted by Trnqvist (1993, p. 150) as representing the age
direction of slope and create a northward gradient advantage, of the IJssel. Following Trnqvist, Berendsen & Stouthamer
facilitating avulsion. (2001, app. 1) also assumed start of IJssel activity at 1700 BP.
The selection of GrN-7525 is quite arbitrary, however, because
from three other locations on the upper IJssel floodplain
younger dates of the upper peat bed exist (Table 1).

Netherlands Journal of Geosciences Geologie en Mijnbouw | 87 4 | 2008 325

 At location 3 (Fig. 1) the top of the peat bed was dated at deposition near Zwolle after Roman times (i.e., ~250 AD), to
1575 35 14C yr BP (GrN-10293). Because this peat was strongly some extent supported by palynological analysis. Ente (1971,
decomposed, beginning IJssel overbank deposition could have 1973-1974) reconstructed the genesis of the IJssel delta using
been younger. Strong changes in organic matter content and many deep boreholes and concluded that delta formation near
pollen composition directly above the dated level (Teunissen, Kampen started in the 12th century. This age estimate relied
1990, pp. 111-112) also suggest a time gap. Indeed, a some- on stratigraphic correlation of an archaeological find (a
what younger date is available from location 4: 1420 30 14C sarcophagus cover) in the lacustrine deposits north of the
yr BP (GrN-12604). This is an unpublished date from Teunissen delta with the early delta deposits. Partly based on historical
of a thin (<10 cm thick) peat bed on top of Pleistocene gravelly reflections, however, he considered it likely that the upper
sand grading upward into humic clay (field observations first IJssel (reach Westervoort-Doesburg) already existed in the 8th
author). Strong decomposition of this shallow peat bed or 9th century.
probably is due to recent drainage. An even younger date is Van de Meene (1979) and Harbers & Mulder (1981) investi-
available from location 5, which is closer to the IJssel: 1265 gated the soils and stratigraphy of the upper IJssel floodplain
45 14C yr BP (GrN-14407). The latter date represents the clayey between Westervoort and Doesburg. They believed that the
top of a 2.5-m-thick peat bed of which two deeper dates exist origin of the upper reach of the IJssel may date back to late
that are stratigraphically consistent (Table 1). One of these, Subboreal (Harbers & Mulder, 1981) or early Subatlantic times
GrN-14410, is believed to be a few centuries too old for paly- (Van de Meene, 1979). With the Subboreal-Subatlantic transi-
nological reasons (Teunissen, 1990, p. 109). tion at 900 BC, this is much earlier than suggested by radio-
Given the present radiocarbon data set, the start of IJssel carbon dates from the same area discussed above. Harbers &
activity could be considerably younger than 1700 (Trnqvist, Mulder (1981) argued that the upper IJssel could initially have
1993; Berendsen & Stouthamer, 2001) or 2000 BP (Teunissen, existed in this area as a small stream without depositing over-
1980). It should also be noted that all radiocarbon dates bank sediments. In contrast, Van de Meene (1977) supposed
available are conventional dates of bulk peat samples. Such late Subboreal initial overbank deposition marking increasing
dates in general are less accurate than AMS dates of selected influence of the Rhine, with upper IJssel channel formation
terrestrial macrofossils (Trnqvist et al., 1992). starting in Subatlantic times.

Stratigraphic analysis and indirect dating The Drusus canal

Widely variable ages of the IJssel have been proposed by various In classic Roman sources (Tacitus, annales II, 8; Suetonius,
authors based on interpretation of stratigraphy and indirect Claudius I, 2) mention is made of military strategic waterworks
dating. Hamming et al. (1965) studied IJssel deposits in an in the lower Rhine area, including a canal dug under super-
excavation near Zwolle and supposed onset of IJssel overbank vision of the Roman general Drusus, leading (northward) into

Table 1. Radiocarbon age determinations from the upper IJssel floodplain. Italics indicate for each core the date closest to the top of the upper peat bed.

Core (Fig. 1) Laboratory number 14C-age (yr BP) Depth below surface (cm) Material Reference
1 GrN-11871 1590 60 103.5 - 107 clay Teunissen (1988, 1990)
1 GrN-5491 2000 65 106 - 108 peat1 Vogel & Waterbolk (1972)
1 GrN-11870 1820 50 117 - 120 clayey peat Teunissen (1988, 1990)
1 GrN-11869 1910 60 124.5 - 127.5 clayey peat Teunissen (1988, 1990)
1 GrN-11868 2810 70 142 - 144.5 peat Teunissen (1990)
2 GrN-11126 1495 25 95 - 97 humic clay Teunissen (1988, 1990)
2 GrN-7525 1720 25 102 - 110 sl. clayey peat1 Teunissen (1980, 1990)
2 GrN-11894 1830 110 110 - 115 sl. clayey peat Teunissen (1988, 1990)
2 GrN-11289 2230 80 117 - 121 sl. clayey peat Teunissen (1988, 1990)
2 GrN-10129 2520 70 140 - 144 clayey peat Teunissen (1990)
3 GrN-10293 1575 35 140 - 144 peat2 Teunissen (1988, 1990)
4 GrN-12602 1420 30 83 - 86 sandy peat2 Teunissen (unpubl.)
5 GrN-14407 1265 45 162 - 165 clayey peat Teunissen (1988, 1990)
5 GrN-14410 1915 50 201 - 203 peat Teunissen (1988, 1990)
5 GrN-14411 2140 50 233 - 235 peat Teunissen (1988, 1990)
1 Decomposed (Poelman, 1981b, p. 174).
2 Strongly decomposed.

326 Netherlands Journal of Geosciences Geologie en Mijnbouw | 87 4 | 2008

hostile territory. This canal was used for a military campaign away from the IJssel channel. In both areas borehole cross-
in 12 BC. The exact position of the Drusus canal has been a sections were constructed more or less perpendicular to the
subject of much debate among archaeologists and historians, IJssel River to facilitate selection of optimum sampling
with the upper IJssel floodplain between Westervoort and locations.
Doesburg being considered as a likely location by a large
number of authors since the 16th century (see references in Wapenveld study area
Teunissen, 1975, 1980). A canal in this area could have con-
nected the Rhine with a drainage system in the middle and Cross-section Wapenveld 1 (Fig. 4) shows the fill of an aban-
lower IJssel valley and as such have been the precursor of the doned IJssel channel in its central part, which consists of silty
present upper IJssel. In this view, the Oude IJssel is considered and sandy clay. The natural levee deposits of the present IJssel
to have been a tributary of the IJssel valley drainage system channel occur more to the east as a wedge of silty and sandy
in Roman times, instead of a Rhine tributary (see section The clay hosting the dyke that bounds the embanked section of
IJssel avulsion). the IJssel floodplain that occurs to the east. The floodbasin
Data that seem to support this hypothesis are radiocarbon deposits of the IJssel appear as a wedge of clay, thinning away
date GrN-5491 (2000 65 14C yr BP) and the remains of a from the present IJssel channel, encasing the upper part of
Roman castellum near Arnhem (Fig. 1) discovered in 1979 the channel-fill, and overlying a bed of peat and clayey peat.
(Teunissen, 1980; Willems, 1980, 1981, 1986). Willems (1980) The clastic IJssel deposits belong to the Echteld Formation,
argued that the existence of a canal would very well explain whereas the peat belongs to the Nieuwkoop Formation (cf.
the location of this fortification. We feel that GrN-5491 cannot Westerhoff et al., 2003).
serve as an argument in favour of a Drusus canal between The clastic deposits below the peat consist mainly of sand
Westervoort and Doesburg, because this date is stratigraphically that may be gravelly. The sandy deposits are draped by a clay
inconsistent and should therefore be rejected (see above). bed (locally sandy or silty) that thins to the west. The top of
Poelman (1981a,b) disputed the conclusions of Teunissen (1980) the sandy deposits shows a step of about 1 m, which is taken
and Willems (1980, 1981) for pedological and palynological to represent the boundary between two buried river terraces.
reasons. The sandy deposits are interpreted to predominantly belong to
the Kreftenheye Formation (Westerhoff et al., 2003), which
Collection of new radiocarbon data occurs throughout the IJssel valley as the product of
Weichselian braided rivers that were part of the Rhine system
Selection of study areas (Van de Meene, 1979; Verbraeck, 1990). However, because of
the long exposure at the surface after abandonment by the
The ambiguity of the radiocarbon ages obtained hitherto, and Rhine, the original braidplain has been reworked by local
the strongly divergent views on the age and origin of the rivers, and aeolian and periglacial processes (e.g. Busschers et
IJssel, were the motive for new field research. Erosion of the al., 2007), of which the sedimentary products belong to the
peat underlying the upper IJssel overbank deposits between Boxtel Formation (Westerhoff et al., 2003) and are hard to
Westervoort and Doesburg was considered a potential cause of distinguish in the field from the Kreftenheye deposits. We
the variability of the radiocarbon dates from this area, and expect a local component (e.g. material derived from the
therefore the focus of this new study was on the lower reaches glacial ice-pushed ridges) to be present in the basal sandy
of the IJssel from which no radiocarbon dates indicating the deposits of Cross-section Wapenveld 1 too. The capping clay
age of the IJssel were available. on top of the sandy deposits is classified as Boxtel Formation
Only at a few locations in the middle and lower IJssel valley and is taken to represent the overbank deposits of a meandering
substantial Holocene peat has developed. A soil map (Stiboka, local river. Thickening of these deposits towards the east
1966) and the Geomorphological Map of the Netherlands suggest that the former channel has been located near the
(Koomen & Maas, 2004) suggest the floodbasin near Wapenveld present IJssel River that perhaps rejuvenated this old course.
(Fig. 2) as one of very few suitable locations. Approximately 5 The local meandering river was probably mainly fed by the
km downstream of Wapenveld, the IJssel valley narrows with Berkel. East of Zutphen, the Berkel shows large palaeomeanders
the glacial ice-pushed ridge on the west-side approaching the (Maas & Makaske, in prep.) that may have been active until
river. Downstream of this narrow, a suitable location (Eilander early-Holocene times (Bos et al., 2005).
& Heijink, 1990) is just northwest of Zwolle (Fig. 3), near the The fill of the palaeochannel in the central part of the cross-
former village of Westenholte. This location is well upstream section is fine-grained, contrasting with the coarse deposits
of the IJssel delta. Considerable floodplain width at these two into which the channel base has scoured. This suggests a
locations allowed sampling the top of the peat beneath IJssel relatively sudden abandonment, possibly following an avulsive
overbank deposits outside the natural levee zone (Figs 2 and 3) shift of the IJssel channel to its present location. Channel-fill
where erosion of underlying peat is more likely than farther width is ~110 m, somewhat less than the historical width of

Netherlands Journal of Geosciences Geologie en Mijnbouw | 87 4 | 2008 327

Fig. 2. Geomorphological map of
the Wapenveld study area (see Fig.
1 for location), with the locations
of the borehole cross-sections
Wapenveld 1 and Wapenveld 2. This
map is an updated detail from the
Geomorphological Map of the
Netherlands (Koomen & Maas,
2004). The legend only explains the
most important map units that are
marked with codes. It shows the
large floodbasin east and southeast
of Wapenveld. An abandoned
channel (code 2R11) exists to the
east between this floodbasin and
the present IJssel channel belt. The
provincial boundary between
Gelderland (west) and Overijssel
(east) partly follows this abandoned
channel. To the west, the floodplain
is bounded by glacial and periglacial
landforms, locally covered by
coversand ridges and (more recent)
inland dune complexes.

328 Netherlands Journal of Geosciences Geologie en Mijnbouw | 87 4 | 2008

Fig. 3. Geomorphological map of the
Westenholte study area (see Fig. 1
for location), with the location of the
borehole cross-section Westenholte.
This map is an updated detail from
the Geomorphological Map of the
Netherlands (Koomen & Maas, 2004).
The legend only explains the most
important map units that are marked
with codes. It shows the right
margin of the IJssel channel belt,
with a floodbasin extending to the
northeast between (partly buried)
aeolian riverdunes. The channel belt
is bounded by a natural levee of
irregular width. North of the aeolian
riverdunes, crevasse splays exist in a
peaty plain.

the present IJssel channel (before normalisation), which was compaction of the underlying peat. However, erosion of peat
~150 m. A remarkable aspect of the palaeochannel is that it will also have occurred, because the peat bed is totally lacking
hardly has associated natural levee deposits of silty and sandy in the extreme eastern part of the cross-section. Likewise,
clay, whereas these are well developed along the present IJssel near the palaeochannel erosion may have taken place. An
channel. Whether the palaeochannel really preceded the present undisturbed top of the peat bed, suitable for sampling, was
IJssel channel is unknown. The IJssel may also initially have expected in the extreme western part of the cross-section.
been an anastomosing system in this reach. The fact that the Cross-section Wapenveld 2 (Fig. 5), located ~350 m down-
present provincial boundary between Gelderland and Overijssel, stream of cross-section Wapenveld 1 (Fig. 2), reveals a detailed
which dates back to medieval times, still partly follows this picture of the lithostratigraphy of the distal IJssel overbank
palaeochannel instead of the present IJssel channel, as it deposits and the underlying peat. This location is charac-
does elsewhere, underscores the former importance of the terized by a gradual peat-clay transition. We sampled the very
palaeochannel. top of the peat (Fig. 5), directly below the transition to peaty
The peat bed in cross-section Wapenveld 1 thickens in a clay (classification of organic sediments according to De Bakker
westward direction. Deposition of the thick natural levee near & Schelling, 1966). The botanical macroremains in the sampled
the present IJssel channel will undoubtedly have resulted in peat were analysed using a binocular microscope and suitable

Netherlands Journal of Geosciences Geologie en Mijnbouw | 87 4 | 2008 329

Fig. 4. Borehole cross-section Wapenveld 1 (see Fig. 2 for location). Borehole data were selected from the DINO database of TNO B&O. The borings were
carried out in 1994 by the Geological Survey of the Netherlands, now part of TNO B&O. Sediment texture description is according to NEN 5104 (Nederlands
Normalisatie-instituut, 1989) and the depth of bed boundaries was identified with a resolution of 10 cm. Lithostratigraphical units (formations)
according to Westerhoff et al. (2003).

remains were selected for radiocarbon age determination (for the R.J. van de Graaff Laboratory (Utrecht University) yielded
the selection procedure of macroremains, see Trnqvist et al., 1026 43 14C yr BP (Table 2). Calibration of this date (Table 3)
1992). Our sample WAP-I consisted of seeds and fruits of Alisma gives 970 - 1046 AD (1 range) as an age estimate for the
plantago-aquatica, Carex, and Typha from a 1-cm-thick slice of beginning of IJssel overbank sedimentation at this location.
a 3.5-cm-wide core. AMS age determination of this sample in

Fig. 5. Borehole cross-section Wapenveld 2 (see Fig. 2 for location) showing the position of radiocarbon-dated sample WAP-I (white dot). Note that
horizontal and vertical scale is different from cross-section Wapenveld 1 (Fig. 4). Borings were carried out for this study using an Edelman auger and a
gouge. Sediment properties were described every 10 cm, following Berendsen & Stouthamer (2001, pp. 29-32). The elevation of all borehole locations was
measured relative to an absolute datum (NAP = Dutch Ordnance Datum). Lithostratigraphical units (formations) according to Westerhoff et al. (2003).

330 Netherlands Journal of Geosciences Geologie en Mijnbouw | 87 4 | 2008

Table 2. Radiocarbon age determinations from the lower IJssel floodplain carried out for this study.

Sample Laboratory number 14C age Depth below surface Co-ordinates and elevation of Mass C Material
(14C years BP) (cm) core location relative to NAP1 (mg)
(m)
WAP-1 UtC-14774 1026 43 55 - 56 203.397/492.597/+1.43 0.49 fruits and seeds2
WH-I-1 UtC-14771 1142 37 60 - 62.5 199.162/503.853/+0.07 2.08 wood and bark
WH-I-2 UtC-14772 1143 55 60 - 67.5 199.162/503.853/+0.07 0.23 fruits3
1 Co-ordinates according to the Dutch co-ordinate system; NAP = Dutch Ordnance Datum.
2 12 Alisma plantago-aquatica fruits, 10 Carex fruits, 2 Typha seeds.
3 4 Carex fruits, 1 Oenanthe aquatica mericarp, 2 Ranunculus fruits, 1 Scirpus lacustris fruit, 4 Urtica dioica fruits, 1 Apiaceae fruit (mericarp).

Table 3. Calibration of radiocarbon ages.

Sample Laboratory number 14C age Median cal. age1 1 cal. age range1 2 cal. age range1
(14C years BP) (years AD) (years AD) (years AD)
WAP-1 UtC-14774 1026 43 1007 970 - 1046 930 - 1098
WH-I-1/WH-I-2 UtC-14771/UtC-14772 1142 312 896 866 - 938 822 - 962

Velperbroek GrN-12602 1420 30 621 594 - 654 566 - 682

Vbroek Circuit GrN-14407 1265 45 770 726 - 814 686 - 862
1 The Groningen calibration program (version CAL25) was used (Van der Plicht, 1993). The degree of smoothing of the calibration curve was based on Trnqvist &
Bierkens (1994). For the bulk samples Velperbroek and Velperbroek Circuit, s = 200 was applied. For the samples WAP-1 and WH-I-1/WH-I-2 that consisted of
wood and terrestrial macroremains, s = 60 was applied.
2 Weighted mean of UtC-14771 and UtC-14772.

Westenholte study area facies indicating much local reworking, and aeolian and local
fluvial deposition. Such local deposits belong to the Boxtel
Cross-section Westenholte (Fig. 6) shows surficial clayey over- Formation. With the present limited data, we leave open the
bank deposits on the right bank of the IJssel. In the south- possibility of a local origin of the basal deposits in Cross-section
western part of the cross-section a 300-m-wide wedge of silty Westenholte.
and sandy clay occurs, representing the natural levee of the A package of fine, poorly sorted sand rests on top of the
IJssel. The dyke, that bounds the embanked part of the IJssel stiff sandy clay in the central part of the cross-section. Sand
floodplain that extends to the west, rests on these natural body geometry and its location on the border of a terrace
levee deposits. Farther to the east, a clay bed extends into the suggest an aeolian riverdune origin, and therefore it was
floodbasin. The IJssel overbank deposits overlie a peat bed of classified as Boxtel Formation (Delwijnen Member) (Westerhoff
variable thickness. The transition from peat (Nieuwkoop et al., 2003). A thin bed of (humic) clay draping the lower
Formation) to clay (Echteld Formation) is generally gradual, level of the terraced surface in the south-western part of the
through intervening layers of clayey peat, peaty clay and cross-section is taken as a local river deposit analogous to the
humic clay. much thicker Boxtel Formation clay bed in the Wapenveld
Most boreholes ended in a blue-grey, stiff, very sandy clay, study area.
with an admixture of relatively coarse material. The charac- The top of the peat below the IJssel overbank deposits was
teristics of this material are very different from the Holocene sampled well outside the natural levee zone in order to
material and indicate a long period of non-deposition with minimize the risk of sampling below overbank deposits that
weathering and soil formation. Only in one borehole, the sand erosively overlie the peat. At the sampling location there is a
that presumably underlies this bed of sandy clay everywhere gradual transition from peat to clay, through peaty clay and
in the cross-section was reached. The stiff clayey deposits seem humic clay. Like in cross-section Wapenveld 2, the very top of
to have a terraced upper surface, with a step of about 1 m in the peat, directly below the transition to peaty clay, was
the south-western part of the cross-section, and could simply sampled (Fig.6). Analysis of botanical macroremains in the
be interpreted as the top of the Kreftenheye Formation. A sample, yielded abundant wood and bark fragments, but only
study in a pit nearby (Hamming et al., 1965), however, revealed few seeds and fruits. We composed two subsamples for radio-
a rather complicated picture of the stratigraphy between the carbon age determination: WH-I-1 consisting of wood and bark
Holocene peat and the top of the Kreftenheye deposits, with remains from a 2.5-cm-thick core interval (diameter 3.5 cm),

Netherlands Journal of Geosciences Geologie en Mijnbouw | 87 4 | 2008 331

Fig. 6. Borehole cross-section from the Westenholte study area (see Fig. 3 for location) showing the position of radiocarbon-dated subsamples WH-I-1
and WH-I-2 (white dot). Borings were carried out for this study using an Edelman auger and a gouge. Sediment properties were described every 10 cm,
following Berendsen & Stouthamer (2001, pp. 29-32). The elevation of all borehole locations was measured relative to an absolute datum (NAP = Dutch
Ordnance Datum). Lithostratigraphical units (formations) according to Westerhoff et al. (2003).

and WH-I-2 consisting of seeds and fruits of various terrestrial 2006) confirmed this AMS bulk peat radiocarbon date. We
plant species from a 7.5-cm-thick core interval that includes agree with Kooistra et al. (2006) that the clay bed directly
and extends downward from the interval from which WH-I-1 above the dated peat at this location belongs to the River
was taken (Table 2). The scarcity of seeds and fruits necessi- Vecht and not to the IJssel. The floodplain site investigated
tated combining remains from a relatively thick interval in by Kooistra et al. (2006) is separated from the IJssel floodplain
order to reach a sufficiently large sample size. AMS age determi- west of Westenholte by aeolian riverdunes and a coversand ridge.
nation of the subsamples in the R.J. van de Graaff Laboratory Clay texture as described by Kooistra et al. (2006) (almost
yielded 1142 37 14C yr BP (WH-I-1) and 1143 55 14C yr BP 100% <2 m) differs strongly from grain size characteristics of
(WH-I-2) (Table 2). A weighted mean of these dates (1142 31 IJssel overbank deposits (e.g. Stiboka, 1966, p. 114-115).
14C yr BP) was calibrated at 866-938 AD (1 range) (Table 3),

suggesting that overbank sedimentation at this location started Interpretation of the new radiocarbon data
about 100 years earlier than at the sampling site in the
Wapenveld study area. The older age obtained in the Westenholte For methodological reasons, the new radiocarbon dates presented
study area, may have been influenced by sampling from a above are superior to previous radiocarbon dates indicating
relatively thick core interval (subsample WH-I-2), thereby the beginning of IJssel overbank sedimentation. The new
including deeper, and thus older, material. On the other hand, dates are AMS dates of selected botanical macroremains,
the almost similar ages of the subsamples WH-I-1 and WH-I-2 which are generally more accurate than conventional dates of
suggest that this ageing effect is very small. bulk peat samples (Trnqvist et al., 1992), such as the previous
Only 2.6 km north-northeast of our Westenholte sampling dates from the upper IJssel floodplain. Also, the new dates
site, the top of a bed of (clayey) peat overlain by clay was radio- can be expected to relate closer to beginning IJssel overbank
carbon-dated at 1540 30 14C yr BP (474-538 AD; 1 range) sedimentation, because only pure, intact peat directly beneath
by Kooistra et al. (2006). Dendrochronological dating of gradual peat-clay transitions was sampled. All previous samples
associated oak and ash tree remains (Sass-Klaassen & Hanraets, were taken from less optimal contexts, i.e., from decomposed

332 Netherlands Journal of Geosciences Geologie en Mijnbouw | 87 4 | 2008

or clayey peat beds, from below sharp peat-clay boundaries or local IJssel valley drainage system and the Rhine. This connec-
from too far below the very top of the peat. Probably because tion transformed the local river with a small catchment
of methodological improvements and a more precise sampling delivering little sediment into a much larger river with a vast
procedure our radiocarbon dataset is much more consistent hinterland delivering large amounts of sediment.
than the dataset available hitherto. A point of discussion remains whether there has been a
We also consider the lower IJssel floodplain to be a more limited time lag between formation of the upper IJssel
suitable sampling area than the upper IJssel floodplain. Van channel, and sedimentation on the lower IJssel floodplain
de Meene (1977, p. 63) described Rhine floodwaters to have near Wapenveld and Westenholte. The available radiocarbon
passed through the Oude IJssel valley (Fig. 1) since Subboreal data suggest so. As argued above, the older part of the upper
times. These have inundated and deposited clay on the upper IJssel floodplain radiocarbon data set (GrN-5491 (2000 65 14C
IJssel floodplain. Therefore, at locations 1 and 2 (Fig. 1) only yr BP), GrN-7525 (1720 25 14C yr BP) and GrN-10293 (1575
clayey peats exist (Teunissen, 1990). Moreover, in the whole 35 14C yr BP) (Table 1)) has low indicative value for the start
Rhine-Meuse delta clastic sedimentation increased after 2000 of IJssel channel formation, and is probably subject to errors
BP and peat formation came to an end (Pons, 1957; De Boer & related to sampling material and sampling site conditions.
Pons, 1960; Berendsen, 1990; Berendsen & Stouthamer, 2001, However, the younger part of the data set (GrN-12602 (1420
p. 89). In this context, the fact that a number of radiocarbon 30 14C yr BP) and GrN-14407 (1265 45 14C yr BP) (Table 1))
dates indicate end of peat formation on the upper IJssel flood- needs serious consideration. The latter date is certainly some-
plain in the Roman Period (12 BC - 425 AD; Willems, 1986) is what too young to indicate initiation of IJssel channel
not surprising, and may not have been caused by the formation formation because it represents clayey peat on top of peat;
of the IJssel channel, but rather by a regional change in sedi- i.e., IJssel clay deposition had already started when the dated
mentation style of the Rhine system that probably already material formed. The former date represents a sample
started ~3000 cal yr BP (Gouw & Erkens, 2007). Therefore, collected ~500 m farther from the IJssel channel belt, at a
radiocarbon dates from the lower IJssel floodplain, located location where thin peat grades upward into humic clay. Field
well outside the zone of clastic overbank deposition of Rhine inspection of the sampling site for the present research
branches other than the IJssel, are more indicative of the time yielded an apparently non-erosive peat-clay transition. With
of IJssel channel formation. the sample having been taken almost directly below the humic
Calibrated ages from the Wapenveld and Westenholte study clay bed (unpubl. description D. Teunissen), GrN-12602 (1420
areas differ about 100 years (Table 3). Taking into account the 30 14C yr BP) can be interpreted to indicate beginning of
resolution associated with the methodology of 14C dating upper IJssel overbank sedimentation at about 600 AD (Table 3)
Holocene channel belts (estimated to be about 100 years by (note that the above-mentioned methodological resolution,
Berendsen & Stouthamer (2000, p. 320)), this means that the about 100 years (Berendsen & Stouthamer, 2000, p. 320),
dates from both areas are essentially similar. Considering the also applies to this estimate). This would imply a time lag of
overlap of the 2 ranges of the calibrated ages (Table 3), onset ~350 years between onset of upper and lower IJssel floodplain
of overbank sedimentation in the lower IJssel area can be sedimentation. Interesting in this context is that IJssel-delta
estimated to have started ~950 AD. To which extent is this progradation was interpreted by Ente (1973 - 1974) to have
important environmental change on the lower IJssel flood- started in the 12th century AD, i.e., another ~200 years later.
plain a response to channel formation upstream in the reach If, in spite of some remaining uncertainties, the younger
Westervoort-Doesburg? part of the upper IJssel radiocarbon dataset (GrN-12602 and
Some authors argued that the upper IJssel could have GrN-14407) is accepted, one could propose a model of IJssel
existed for a very long period (~1000 year), without widespread channel-belt progradation, as a part of a relatively slow avulsion
deposition of overbank sediments (Harbers & Mulder, 1981). process (cf. Smith et al., 1989; Farrell, 2001), accounting for
This notion is in conflict with the concepts underlying the a downstream decreasing age of first IJssel sedimentation.
extensive palaeogeographical work carried out by Berendsen
(1982), Berendsen & Stouthamer (2000, 2001) and many co- The IJssel avulsion
workers, leading to excellent reproducible results over the last
three decades. Moreover, this notion is contradicted by research In the course of the Holocene, many avulsions have occurred
in modern fluvial environments, showing that the formation in the Rhine-Meuse delta (e.g. Stouthamer & Berendsen, 2000;
of a new channel by avulsion is accompanied by widespread Makaske et al., 2007) as a result of aggradation of channel
sedimentation stopping peat growth (Smith et al., 1989). belts above the surrounding floodplain (Berendsen &
Therefore, we exclude the possibility that the upper IJssel Stouthamer, 2001). The steep cross-levee slopes developing
existed very long before the 10th century AD, and rather near the channel usually play an important role in the early
consider the onset of lower IJssel overbank sedimentation as stages of the avulsion process (Slingerland & Smith, 1998),
a response to the establishment of a connnection between the when a crevasse in the levee develops and routes water and

Netherlands Journal of Geosciences Geologie en Mijnbouw | 87 4 | 2008 333

sediments to the floodplain. Farther away, on the floodplain, lining the northern boundary of the Late-Weichselian terrace
gradients approach the general valley slope. An energetically that underlies the Oude IJssel valley. Because the aeolian dunes
favourable avulsion route on the floodplain must exist for an are essentially Late-Weichselian features (Van de Meene, 1977,
avulsion to be completed, i.e., to result in the formation of new pp. 44-45) the crevasse channels are younger, probably Holocene
channel belt (cf. Makaske, 2001, p. 158). in age. A Holocene age is suggested by poorly developed soils
In the case of the IJssel avulsion, the first reach of the and clay admixture in the associated sandy crevasse splays
avulsion route must have had a relatively low gradient, which (Stiboka, 1975). The crevasse splays are non-calcareous
may have been a reason for initial slow avulsion-belt progra- (Stiboka, 1975) indicating that they predominantly consist of
dation. Nowadays, there is hardly any difference between locally eroded Pleistocene material. This suggests that the
channel-belt gradients for various reaches of the IJssel (Fig. 7). crevasse splays formed before the IJssel avulsion introduced
Originally, however, a low divide existed near Doesburg con- large amounts of fresh calcareous sediments (i.e., the material
sisting of the top of the Middle-Weichselian Rhine terrace that found in the IJssel channel belt today) into the area. We feel
was formed before abandonment of the IJssel valley by the that the Oude IJssel played a role in crevasse splay formation.
Rhine. A digital elevation model of the Doesburg area (Fig. 8) The present Oude IJssel is a small local river that seems
shows intact parts of this divide at the present surface. From incapable to form crevasses of the size observed on its own.
this divide, the Middle-Weichselian terrace dips to the north We hypothesize that Rhine floodwaters initiated crevasse
largely being exposed at the surface in juxtaposition to (and splay formation. When the Rhine started to spill into the Oude
slightly above) the modern IJssel channel belt. North of IJssel valley in the Subboreal (Van de Meene, 1977; Fig. 1),
Deventer, the terrace is covered with Holocene IJssel deposits infilling of the Oude IJssel valley with Holocene sediments
and it presumably corresponds to the upper terrace levels gradually reduced the relief along the divide and forced extreme
identified in the subsurface in the Wapenveld and Westenholte Rhine floods to seek an outlet across the divide eroding gaps
study areas (Figs 4 and 6). From the divide near Doesburg the into it. One of these crevasses near Doesburg, in the lowest
terrace also dips to the southwest, as indicated by the top level part of the divide, may have caused a northward avulsion of
of terrace remnants. In this area, however, southwest dipping the Oude IJssel. Possibly, a small pre-existing stream draining
Late-Weichselian terraces constitute most of the Pleistocene seepage water from the ice-pushed ridge that borders the IJssel
subsurface below the Holocene peat and alluvium. Late- valley to the west, guided the Oude IJssel avulsion route
Holocene aggradation in this area had to compensate for the northward down the Middle-Weichselian terrace, linking it to
reverse slope in the Pleistocene surface to enable water to spill the lower IJssel valley drainage system. Before that time, the
over the divide near Doesburg into the steeper middle IJssel Oude IJssel more likely flowed southwestward into the Rhine-
valley. An important question is to which extent this divide Meuse delta joining the Rhine near Arnhem, thereby following
was still intact when the IJssel avulsion was taking place. the overall gradient of the shallow Late-Weichselian subsurface.
In Figure 8 a few crevasse channels crossing the divide can Increasing aggradation in the Arnhem area during the late
be identified east of Doesburg. They dissect aeolian dunes Holocene must have hampered drainage of the Oude IJssel in

Fig. 7. Gradient lines of the

IJssel channel belt and the
Middle-Weichselian terrace.
Data points for the IJssel
channel belt represent highest
surface levels of early IJssel
point bars as identified on a
digital elevation model (AHN).
Data points for the Middle-
Weichselian terrace represent
AHN surface levels near the
IJssel channel belt, and data
from Pons (1953) and Figs 4,
5 and 6 where the terrace is
in the subsurface. Gradient
lines were fitted by the eye.

334 Netherlands Journal of Geosciences Geologie en Mijnbouw | 87 4 | 2008

Fig. 8. Digital elevation model (AHN 5 5 m; Van Heerd et al., 2000) of the area around Doesburg (see Fig. 1 for location), showing the former divide
between the IJssel valley catchment to the north and the Rhine catchment to the south. Colours in order of increasing elevation: green-yellow-red-white.
The former divide approximately follows the southern edge of the Middle-Weichselian terrace (white line). Directly west of Doesburg the IJssel broke
through the divide (presumed former edge of the Middle-Weichselian terrace indicated by dashed white line) and eroded the terrace. A high (field-
checked) terrace remnant remains west of the IJssel (white circle). East of Doesburg crevasse channels in the edge of the Middle-Weichselian terrace can
be seen (white arrows).

this direction, contributing to a northward avulsion. Continuing triggered by a single extreme Rhine flood, as proposed by
Rhine channel belt aggradation, ultimately reversed the surface Cohen & Lodder (2007), based on preliminary observations on
gradient between Westervoort and Doesburg and led to the IJssel crevasses and splays in the IJssel valley.
avulsion that perhaps reactivated (parts of) the abandoned
lower Oude IJssel course. Conclusions
Anyway, the geomorphological data suggest that the IJssel
avulsion has been a complex process taking significant time, New AMS radiocarbon age determinations indicate that over-
which would explain how upper IJssel sedimentation may bank sedimentation along the lower reach of the Gelderse IJssel
have started earlier than lower IJssel sedimentation. It can also started ~950 AD. This environmental change most likely resulted
be expected that incision of the middle IJssel channel into the from an avulsion of the Rhine into the IJssel valley. Prior to
generally stiff, consolidated Middle-Weichselian terrace has this avulsion, only local streams carrying minor amounts of
been a slow process, initially limiting the amounts of sediment sediment existed in the IJssel valley during most of the
that could be transported to the lower IJssel floodplain. This Holocene. Reinterpretation of previous conventional radiocarbon
concept fits with the typically very gradual transition from data from the upper IJssel floodplain leads to the conclusion
peat to floodplain clay observed in the lower IJssel floodplain that upper IJssel channel formation may have started ~600
[Hamming et al. (1965) even distinguished finer old IJssel AD. Our results contradict earlier suppositions, based on inter-
clay and sandier younger IJssel clay] and also would explain pretation of archaeological data and historical accounts, that
the apparent time lag between initial channel formation and the Gelderse IJssel is much older and originated as a canal,
IJssel-delta progradation (Ente, 1973-1974). The gradual dug under supervision of the Roman general Drusus.
transition from peat to floodplain clay speaks against the Geomorphological and soil data suggest that the IJssel
hypothesis that the IJssel avulsion was a catastrophic event, avulsion was a slow process, retarded by a low gradient of the

Netherlands Journal of Geosciences Geologie en Mijnbouw | 87 4 | 2008 335

first reach of the avulsion route, and an erosion-resistant Cohen, K.M., 2003. Differential subsidence within a coastal prism; Late-Glacial
subsurface (the strongly consolidated top of the Middle- Holocene tectonics in the Rhine-Meuse delta, the Netherlands. Nederlandse
Weichselian terrace) in the middle reach of the avulsion route. Geografische Studies 316: 172 pp.
These particular conditions caused the IJssel avulsion to have Cohen, K.M. & Lodder, Q.J., 2007. Paleogeografie en veiligheid tegen over-
been different in process and rate from many other Holocene stromen; de bruikbaarheid van inzichten in de ontwikkeling van de
avulsions in the Rhine-Meuse delta, which might explain the Nederlandse delta in de laatste 5000 jaar voor het kwantitatief begrenzen
apparent time lag between the beginning of upper and lower van overstromingsmagnitudes en -frequenties. RWS RIZA rapport 2007.016,
IJssel overbank sedimentation. Rijkswaterstaat RIZA (Lelystad): 80 pp.
The early-medieval age of the Gelderse IJssel suggests that De Bakker, H. & Schelling, J., 1966. Systeem van bodemclassificatie voor
the IJssel avulsion has been an important trigger for the rise Nederland; de hogere niveaus. Pudoc (Wageningen), 217 pp.
of medieval trading cities on the IJssel like Zutphen, Deventer, De Boer, Th.A. & Pons, L.J., 1960. Bodem en grasland in de Vijfheerenlanden.
Kampen and Doesburg. Verslagen van landbouwkundige onderzoekingen 66.6, Pudoc (Wageningen):
72 pp.
Acknowledgements Ente, P.J., 1971. Sedimentary geology of the Holocene in Lake IJssel region.
Geologie en Mijnbouw 50: 373-382.
We wish to thank Arjan Koomen for managing the project GKN Ente, P.J., 1973-1974. De IJsseldelta. Kamper Almanak (Kampen): 137-164.
(Geomorphological Map of the Netherlands), in the framework Eilander, D.A. & Heijink, W., 1990. Bodemkaart van Nederland, schaal 1 : 50.000;
of which the present work was carried out. Klaas van der Borg toelichting bij de kaartbladen 20 West Lelystad (gedeeltelijk), 20 Oost
(R.J. van de Graaff Laboratory, Utrecht University) is thanked Lelystad en 21 West Zwolle. Staring Centrum (Wageningen): 165 pp.
for carrying out high-quality radiocarbon age determinations. Farrell, K.M., 2001. Geomorphology, facies architecture, and high-resolution,
TNO-B&O kindly supplied borehole data for Fig. 4. Frans Bunnik, non-marine sequence stratigraphy in avulsion deposits, Cumberland Marshes,
Joop Kalis and Hans van der Plicht were very helpful in the Saskatchewan. Sedimentary Geology 139: 93-150.
search for details about radiocarbon dates used by the late Gouw, M.J.P. & Erkens, G., 2007. Architecture of the Holocene Rhine-Meuse
Dr D. Teunissen. We appreciated the thoughtful comments of delta (the Netherlands); a result of changing external controls. Netherlands
the journal reviewers Kim Cohen and Henk Weerts. Journal of Geosciences Geologie en Mijnbouw 86: 23-54.
Hamming, C., Knibbe, M. & Maarleveld, G.C., 1965. Afzettingen van de IJssel,
References nabij Zwolle. Boor en Spade 14: 88-103.
Harbers, P. & Mulder, J.R., 1981. Een poging tot reconstructie van het Rijnstelsel
Berendsen, H.J.A., 1982. De genese van het landschap in het zuiden van de in het oostelijk rivierengebied tijdens het Holoceen, in het bijzonder de
provincie Utrecht, een fysisch-geografische studie. Utrechtse Geografische Romeinse tijd. K.N.A.G. Geografisch Tijdschrift 15: 404-421.
Studies 25: 259 pp. Kooistra, M.J., Kooistra, L.I., Van Rijn, P. & Sass-Klaassen, U., 2006. Woodlands
Berendsen, H.J.A., 1990. River courses in the Central Netherlands during the of the past; the excavation of wetland woods at Zwolle-Stadshagen (the
Roman Period. Berichten van de Rijksdienst voor het Oudheidkundig Netherlands): reconstruction of the wetland wood in its environmental
Bodemonderzoek 40: 243-249. context. Netherlands Journal of Geosciences Geologie en Mijnbouw 85:
Berendsen, H.J.A. & Stouthamer, E., 2000. Late Weichselian and Holocene 37-60.
palaeogeography of the Rhine-Meuse delta, the Netherlands. Palaeogeography, Koomen, A.J.M. & Maas, G.J., 2004. Geomorfologische Kaart Nederland (GKN);
Palaeoclimatology, Palaeoecology 161: 311-335. achtergronddocument bij het landsdekkende digitale bestand. Alterra-
Berendsen, H.J.A. & Stouthamer, E., 2001. Palaeogeographic development of rapport 1039, Alterra (Wageningen): 38 pp.
the Rhine-Meuse delta, The Netherlands. Van Gorcum (Assen): 268 pp. Lanting, J.N. & Mook, W.G., 1977. The pre- and protohistory of the Netherlands
Berendsen, H.J.A., Hoek, W.Z. & Schorn, E.A., 1995. Late Weichselian and in terms of radiocarbon dates. (Groningen): 247 pp.
Holocene river channel changes of the rivers Rhine and Meuse in the Makaske, B., 2001. Anastomosing rivers: a review of their classification, origin
Netherlands (Land van Maas en Waal). Paloklimaforschung/Palaeoclimate and sedimentary products. Earth-Science Reviews 53: 149-196.
Research 14: 151-171. Makaske, B., Berendsen, H.J.A. & Van Ree, M.H.M., 2007. Middle Holocene
Bos, J.A.A., Van Geel, B., Groenewoudt, B.J. & Lauwerier, R.C.G.M., 2005. avulsion-belt deposits in the central Rhine-Meuse delta, the Netherlands.
Early Holocene environmental change, the presence and disappearance of Journal of Sedimentary Research 77: 110-123.
early Mesolithic habitation near Zutphen (the Netherlands). Vegetation Nederlands Normalisatie-instituut, 1989. NEN 5104: Geotechniek; classificatie
History and Archaeobotany 15: 27-43. van onverharde grondmonsters. Nederlands Normalisatie-instituut (Delft):
Busschers, F.S., Kasse, C., Van Balen, R.T., Vandenberghe, J., Cohen, K.M., 23 pp.
Weerts, H.J.T., Wallinga, J., Johns, C., Cleveringa, P. & Bunnik, F.P.M., Poelman, J.N.B., 1981a. De gracht van Drusus? Westerheem 30: 20-23.
2007. Late Pleistocene evolution of the Rhine-Meuse system in the southern Poelman, J.N.B., 1981b. Facta non fabula. Westerheem 30: 173-176.
North Sea basin: imprints of climate change, sea level oscillation and glacio- Poelman, J.N.B. & Harbers, P., 1966. Heeft Lathum altijd aan de rechter IJssel-
isostacy. Quaternary Science Reviews 26, pp. 3216-3248. oever gelegen? Boor en Spade 15: 128-130.

336 Netherlands Journal of Geosciences Geologie en Mijnbouw | 87 4 | 2008

Pons, L.J., 1953. De bodemgesteldheid van een gedeelte van de Lijmers (het Van Heerd, R.M., Kuijlaars, E.A.C., Teeuw, M.P. & Van t Zand, R.J., 2000.
binnendijkse land van de gem. Westervoort, Duiven (ged.) en Zevenaar Productspecificatie AHN 2000. Rapport MDTGM 2000.13, Rijkswaterstaat,
(ged.)). Rapport 343, Stichting voor Bodemkartering (Wageningen): 58 pp. Adviesdienst Geo-informatie en ICT (Delft): 22 pp.
Pons, L.J., 1957. De geologie, de bodemvorming en waterstaatkundige ontwik- Verbraeck, A., 1990. De Rijn aan het einde van de laatste ijstijd: de vorming
keling van het Land van Maas en Waal en een gedeelte van het Rijk van van de jongste afzettingen van de Formatie van Kreftenheye. K.N.A.G.
Nijmegen. Mededelingen van de Stichting voor Bodemkartering, Bodemkundige Geografisch Tijdschrift 23: 328-339.
Studies 3, Stichting voor Bodemkartering (Wageningen): 156 pp. Vogel, J.C. & Waterbolk, H.T., 1972. Groningen radiocarbon dates X. Radio-
Sass-Klaassen, U. & Hanraets, E., 2006. Woodlands of the past; the excavation carbon 14: 6-110.
of wetland woods at Zwolle-Stadshagen (the Netherlands): growth pattern Westerhoff, W.E., Wong, T.E. & De Mulder, E.F.J., 2003. Opbouw van de
and population dynamics of oak and ash. Netherlands Journal of Geosciences ondergrond. In: De Mulder, E.F.J., Geluk, M.C., Ritsema, I.L., Westerhoff,
Geologie en Mijnbouw 85: 61-71. W.E. & Wong, Th.E. (eds): De ondergrond van Nederland. Wolters-Noordhoff
Slingerland, R. & Smith, N.D., 1998. Necessary conditions for a meandering- (Groningen/Houten): 247-352.
river avulsion. Geology 26: 435-438. Willems, W.J.H., 1980. Arnhem Meinerswijk: een nieuw castellum aan de Rijn.
Smith, N.D., Cross, T.A., Dufficy, J.P., & Clough, S.R., 1989. Anatomy of an Westerheem 29: 334-348.
avulsion. Sedimentology 36: 1-23. Willems, W.J.H., 1981. Over een fossa, een castra en een tabula. Westerheem
Stiboka, 1966. Bodemkaart van Nederland, schaal 1:50.000; toelichting bij 30: 168-171.
kaartblad 27 Oost Hattem. Stichting voor Bodemkartering (Wageningen): Willems, W.J.H., 1986. Romans and Batavians; a regional study in the Dutch
126 pp. eastern river area. PhD thesis, University of Amsterdam (Amsterdam): 491 pp.
Stiboka, 1975. Bodemkaart van Nederland, schaal 1 : 50.000; toelichting bij de
kaartbladen 40 West Arnhem en 40 Oost Arnhem. Stichting voor Bodem-
kartering (Wageningen): 197 pp.
Stouthamer, E. & Berendsen, H.J.A., 2000. Factors controlling the Holocene
avulsion history of the Rhine-Meuse delta (the Netherlands). Journal of
Sedimentary Research 70: 1051-1064.
Teunissen, D., 1975. De wordingsgeschiedenis van het natuurlijke landschap
van de Duffelt. Numaga 22: 79-94.
Teunissen, D., 1980. Enkele gebeurtenissen uit het leven van de Romeinse
veldheer Drusus in het licht van enige nieuwe geologische gegevens uit
oostelijk Gelderland. Westerheem 29: 321-334.
Teunissen, D., 1981. Nogmaals Drusus. Westerheem 30: 167-168 pp.
Teunissen, D., 1988. De bewoningsgeschiedenis van Nijmegen en omgeving,
haar relatie tot de landschapsbouw en haar weerspiegeling in palynologische
gegevens. Mededelingen van de Afdeling Biogeologie van de Sectie Biologie
van de Katholieke Universiteit van Nijmegen 15: 108 pp.
Teunissen, D., 1990. Palynologisch onderzoek in het oostelijk rivierengebied:
een overzicht. Mededelingen van de Afdeling Biogeologie van de Sectie Biologie
van de Katholieke Universiteit van Nijmegen 16: 161 pp.
Trnqvist, T.E., 1993. Fluvial sedimentary geology and chronology of the Holocene
Rhine-Meuse delta, the Netherlands. Nederlandse Geografische Studies 166:
169 pp.
Trnqvist, T.E. & Bierkens, M.P.F., 1994. How smooth should curves be for
calibration of radiocarbon ages? Radiocarbon 36: 11-26.
Trnqvist, T.E., De Jong, A.F.M., Oosterbaan, W.A. & Van der Borg, K., 1992.
Accurate dating of organic deposits by AMS 14C measurement of macrofossils.
Radiocarbon 34: 566-577.
Van de Meene, E.A., 1977. Toelichtingen bij de geologische kaart van Nederland
schaal 1 : 50.000; blad Arnhem Oost (40O). Rijks Geologische Dienst (Haarlem):
147 pp.
Van de Meene, E.A., 1979. Het ontstaan van de Geldersche IJssel. K.N.A.G.
Geografisch Tijdschrift 13: 202-210.
Van der Plicht, J., 1993. The Groningen radiocarbon calibration program. Radio-
carbon 35: 231-237.

Netherlands Journal of Geosciences Geologie en Mijnbouw | 87 4 | 2008 337

Wageningen UR - Alterra - Nieuws & Agenda Page 1 of 2

Log in Zoek Links

Uitgebreid
Deze Site Zoek
zoeken

Onderzoek Publicaties Alterra Producten Nieuws & Agenda Over Alterra Werken bij Contact

wageningen ur (home) > alterra (home) > nieuws & agenda > nieuws > de ijssel is jonger dan gedacht

De IJssel is jonger dan gedacht

Nieuws 12 jan 2009
Onderdeel: Alterra Meer over dit
Persvoorlichting onderwerp
Dossiers Over de oorsprong van de IJssel is weinig bekend. Lange tijd
stroomde in het IJsseldal een klein riviertje, dat echter niet Dit onderzoek (The
Archief verbonden was met de Rijn. De bovenloop van de IJssel zou rond age and origine of the
het begin van de jaartelling zijn gegraven door de Romeinse Gelderse IJssel) is
RSS veldheer Drusus, zo wilden oude geschriften ons doen geloven. Dat gepubliceerd in het
is echter onjuist, zo blijkt nu. Netherlands Journal of
Agenda Geosciences.
Alterra- Contact
onderzoekers
Bart Makaske Alterra, onderdeel van
en Gilbert Maas Wageningen UR
hebben samen
dr. A. (Bart) Makaske
met Biax
Consult de 0317 481609
ouderdom van bart.makaske@wur.nl
de IJssel
vastgesteld met
behulp van ing. G.J. (Gilbert) Maas
bodemonderzoek 0317 481649
en gilbert.maas@wur.nl
radiokoolstofdateringen.
Op twee meer Contact
plaatsen in het
IJsseldal
bepaalden zij
de ouderdom van een veenlaag onder de oudste kleilagen van de IJssel.
Het onderzoek toont aan dat de IJssel pas in de vroege Middeleeuwen is
ontstaan, tussen 600 en 950 na Chr. Daarvoor was het huidige IJsseldal
een moerassig gebied waarin een klein riviertje stroomde dat werd
gevoed door lokale beken, die water afvoerden van de hoger gelegen
zandgronden. Een verbinding met de Rijn was er toen nog niet.

Een populaire theorie zegt dat de bovenloop van de IJssel rond het begin
van de jaartelling is gegraven door de Romeinse veldheer Drusus.
Stroomafwaarts van de splitsing van de Rijn en de Waal, niet ver van de
Limesweg die vanuit Duitsland tot Katwijk langs de zuidoever van de Rijn
liep, zouden de soldaten van Drusus rond 12 v. Chr. een kanaal hebben
aangelegd. Een goede vaarroute naar het noorden was namelijk van
groot belang, omdat de Romeinen toen plannen hadden hun rijk naar die
kant uit te breiden. Deze theorie is gebaseerd op klassieke Romeinse
geschriften en was eeuwenlang een punt van discussie tussen
geschiedkundigen, archeologen en geografen.

Bart Makaske: Met ons onderzoek weerleggen we deze theorie definitief.

Ons werk werpt nieuw licht op de vroegste geschiedenis van de IJssel als
belangrijke vaarroute. Het ontstaan van de IJssel, als spontane aftakking
van de Rijn, was een langzaam proces dat pas begon rond 600 na Chr.,
in de buurt van Arnhem. Langzamerhand ging de nieuwe loop steeds
meer Rijnwater afvoeren en rond 950 na Chr. raakte het veengebied bij
Zwolle bedekt met de afzettingen van de nieuwe rivier. De opkomst van
de Middeleeuwse handelssteden Doesburg, Zutphen, Deventer en
Kampen volgt snel op het ontstaan van de IJssel.

Meer informatie:
dr. Bart Makaske, 0317 481609 of bart.makaske@wur.nl
ing. Gilbert Maas, 0317 481649 of gilbert.maas@wur.nl

Dit onderzoek (The age and origine of the Gelderse IJssel) is

gepubliceerd in het Netherlands Journal of Geosciences. Zie
www.njgonline.nl.

http://www.alterra.wur.nl/NL/nieuwsagenda/nieuws/De_IJssel_is_jonger_dan_gedacht... 17-6-2009
Wageningen UR - Alterra - Nieuws & Agenda Page 2 of 2

Print nieuwsbericht

Disclaimer Contact Alle content 2009 Wageningen UR. Alle rechten voorbehouden.

http://www.alterra.wur.nl/NL/nieuwsagenda/nieuws/De_IJssel_is_jonger_dan_gedacht... 17-6-2009