Reply: Fluvial incision and channel downcutting as a response to Late-glacial and Early Holocene climate change: the lower reach of the River Meuse (Maas) The Netherlands

Reply: Fluvial incision and channel downcutting as a response to Late-glacial and Early Holocene climate change: the lower reach of the River Meuse (Maas) The Netherlands

   JOURNAL OF QUATERNARY SCIENCE (2000)  15  (1) 95–100 CCC 0267-8179/2000/010095–06$17.50Copyright  © 2000 John Wiley & Sons, Ltd. Correspondence Reply: Fluvial incision andchannel downcutting as aresponse to Late-glacial andEarly Holocene climate change:the lower reach of the RiverMeuse (Maas) The Netherlands In their comment on our contribution on Late-glacial dynam-ics of the River Meuse, Kasse  et al  . (1999) question thevalidity of our  14 C data and disagree on our reconstructionof Late-glacial gradient lines and interpretations concerningthe presence and age of a transitional river system in theNorth Limburg study area. Before discussing the details, wefirst address the comments in their first paragraph, in whichKasse  et al  . wonder why we reinvestigated the Meuse valleyin North Limburg and came up with a new interpretation.We deliberately selected this area to relate the bulk geo-chemistry of Late-glacial fine-grained residual channel infil-lings to Late-glacial climate change and to obtain inde-pendent and absolute dates (Tebbens  et al  ., 1998; Tebbens,1999). Previously, most residual channels had been datedonly using pollen stratigraphy. We discovered new sites inaddition, and datable material in the undated residual chan-nels of the transitional river system of Kasse  et al  . (1995)and Huisink (1997, 1999). On arrival of the  14 C results andafter careful consideration (see below), it appeared that seve-ral dated residual channels did not fit into the existingregional pollen-based stratigraphy. Because some major dis-crepancies could not be reconciled with previous interpret-ations of Kasse  et al  . (1995) and Huisink (1997, 1998, 1999),we published a different interpretation for the same area.The following sections will present our thoughts to addressthe criticism of  Kasse  et al  . (1999), which in our view is notalways justified. 14 C dating of bulk samples Kasse  et al  . (1999) legitimately refer to the hardwater andreservoir effects that are inherent to bulk  14 C dates obtainedfrom (partly) calcareous gyttja deposits. With these ageingeffects, they illustrate their doubts of our entire chronostrati-graphical framework and subsequent interpretations. Thesrcinal section on  14 C chronostratigraphy (Tebbens  et al  .,1999) indicates that we were fully aware of the ageingeffects during interpretation of our bulk  14 C results.Kasse  et al  . (1999) state that ‘a  combination  of pollenanalytical results  and   bulk  14 C dates can give a more reliable estimate   of the obtained radiocarbon ages’. From the formerpart, we conclude that an independent set of complementary 14 C dates was actually most welcome, because it enablesthe reader to critically evaluate the pollen stratigraphy aswell. Pollen stratigraphy should not be overvalued. Afterall, vegetation developments lag climate changes by severalhundreds of years. Thus, biostratigraphical pollen zones aretime-transgressive and have to be ‘pinpointed’ in time too.Consequently, there is no alternative of using the radiocar-bon chronology as the basic tool for correlation betweendifferent archives (Wohlfahrt, 1996). Indeed, Hoek (1997)has dated the Late-glacial pollen zones of The Netherlandsboth conventionally and with AMS in order to obtain aregional bio- and chronostratigraphical framework. Obvi-ously, any dating (both pollen and  14 C) on an oldest residualchannel  infilling   is by definition younger than the period of channel downcutting and formation of the correspondingterrace morphology (e.g. meander pointbars) itself. Forexample, a point of concern in using pollen zones forreconstructing fluvial activity of the Meuse system inresponse to climate change is that pollen zones from theMeuse headwaters will be of greater importance than thosefrom the southern Netherlands. This is because the Ardenneslow-mountain range with its many tributaries generates themajor share of run off (discharge) and has the highesthillslope erosion rates (sediment supply), especially in aperiglacial climate. Thus, the balance between sedimentsupply and river transport capacity and the timing of fluvialresponse to climate change will be determined there andnot in the southern Netherlands.‘Estimating radiocarbon ages’ by means of pollen stratigra-phy can be considered as a second-best alternative toobtaining absolute chronostratigraphical ages. Nevertheless,we followed exactly this procedure to assess our  14 C dates.In our srcinal contribution, the bulk  14 C dates have beencompared extensively with available pollen data. We exten-sively discussed those cases that did not match in order toprovide the readers with food for thought to construct theirown interpretation. Thus, the statement of  Kasse  et al  . (1999)in their first and last paragraph that ‘pollen analyses % havehardly been used’ is simply not true. Our dates should beconsidered as providing independent and complementaryinformation to their pollen stratigraphy. Therefore, we agreethat a combination of pollen analyses and  14 C dates is veryuseful to date or assess the dating of residual channel infill-ings, but we reject the view that any  14 C date that is not inaccordance with an  expected   age based on biostratigraphicalpollen zones is necessarily wrong.Kasse  et al  . cite Hoek (1997; p. 112), to exemplify a bulk 14 C date from a calcareous sample being 800–1200 yr olderrelative to its AMS age and  expected   pollen age. However,this sample referred to lake marl containing 90% CaCO 3 .Moreover, the ageing effect was calculated for the authigeniccalcium carbonate fraction, which is removed during pre-treatment of a bulk sample. Likewise, the organic detritusof  Bo¨ttger  et al  . (1998) srcinated from lake marl with  40%CaCO 3 . Assuming all CaO srcinated from CaCO 3  (ignoringCaO bound in clay minerals and feldspars), we can estimatefrom our CaO–XRFS bulk measurements that the most cal-careous sample (site 1, Beugen) contained 23% CaCO 3  atmaximum, whereas most other dated samples contained 0  96 JOURNAL OF QUATERNARY SCIENCE to 8%. It would have been more appropriate to mention theageing effect for the bulk organic fraction, which amountedto between 500 and 800 yr (Hoek, 1997). Its very low   13 Cvalue ( − 36.42‰) already hinted in that direction, whereasour samples were in the  − 27 to  − 31‰ range. With respectto assessing the bulk  14 C dates by means of their   13 Cvalues, Kasse  et al  . assert that we ‘proposed to correct thesamples having   13 C values of     − 30‰, simply by(subtracting) 600 years’. Our words have been misquotedhere. We evaluated the  possible   magnitude of the ageingeffects as our text clearly reads: ‘Our   13 C values are compa-rable to those measured by To¨rnqvist  et al  .  %  and for thisreason we  expect   ageing effects will be  in the same order of magnitude  , i.e. 0.2 to 0.6 kyr’, ‘ % These samples  possibly  suffer from a hardwater effect of   over   0.6 kyr (p. 63)’ andin case of only one sample we stated: ‘allowing a  maximum  correction of 0.6 kyr, an  approximate   age of 11.9 kyr BP isfound’ (Tebbens  et al  ., 1999; p. 68).Performing AMS analyses on terrestrial macrofossils mighthave prevented this discussion on the use of bulk  14 C dates.Our srcinal contribution provides the reason we could notdo this. However, before raising the AMS analyses to the‘Golden Standard’, it should be mentioned that these toomight suffer from comparable dating uncertainties as thebulk sediment  14 C dates. For example, some 20 dates froma set of 51 AMS dates of Late-glacial samples were anoma-lously  younger   (several hundreds to thousands of radiocarbonyears), compared with their expected and well-establishedvarve and biostratigraphical ages (Wohlfahrt  et al  ., 1998).This was attributed to microbial contamination during samplepreparation and/or storage. We performed and published anAMS date on terrestrial macroremains from the Beugenresidual channel (Fig. 1), in order to assess its disputedconventional date on calcareous gyttja. Ironically (andunfortunately), the AMS date (8390 yr BP) was 3500 yryounger compared with the expected pollen age (11900 yrBP) and 4000 yr younger with respect to the conventionaldate (12330 yr BP). In dialogue with our Amsterdam col-leagues, we decided to reject this date because typicalHolocene pollen was absent above the sampled level. Thisexperience gives us reason to doubt the Bosscherheide AMSdate of Huisink (1998; 12390 yr BP and 900 radiocarbonyears younger), for it was prepared in the same laboratory.However, if we accept this AMS date, then we can onlyguess why our bulk date of 13280 yr BP should be tooold and be disputed: the lime-free, non-gyttja bulk sampleconsisted of purely organic material (  13 C:  − 28.12‰) from ashallow palaeochannel in a well-protected setting. Terrestrialmacroremains (small twigs,  Betula nana  and  Dryas Octopet- ala  leaves) were clearly distinguishable and have been sub-mitted in bulk quantity to the laboratory. Furthermore, se-veral other conventional bulk dates from a comparablesetting at a distance of several hundreds of metres in thesame sandpit (Bohncke  et al  ., 1993) are not disputed at all.Therefore, we still see no reason to reject our Bosscher-heide dating.Kasse  et al  . question the interpretation in which we placedrapid channel downcutting and meandering river activity inthe Late Bølling period. We based this on bulk  14 C datesof the only and best-developed meanders in the area (sites1c, 7, 11a and Dubbroek). Even if the conventional date of the oldest Beugen meander gyttja (Fig. 1: 12330 yr BP) wasto be doubted, then pollen still indicate that  infilling   of thismeander started ‘ca. 11900 yr BP’, i.e. ‘the very beginningof the Allerød’ according to Kasse  et al  . (1999). So, althoughthe bulk date from our most calcareous sample might be400 yr too old (as we anticipated), palynology also indicates Copyright  ©  2000 John Wiley & Sons, Ltd. J. Quaternary Sci., Vol. 15(1) 95–100 (2000) that the well-developed neck cut-off meander had alreadybeen  formed and abandoned   then. Loamy overbank sedi-ments in the Bosscherheide pit and  upstream   of this meanderwere deposited before 12200 yr BP (Tebbens  et al  ., 1999).Likewise, overbank loam was deposited ‘in a period rangingover the Bølling till late in the Allerød Interstadial’ on aLate Pleniglacial terrace and locally before 12210 yr BP(Teunissen, 1983). Other bulk dates of the earliest infillingsof the deep Meerlo (12500 yr BP), Keuter (11780 yr BP)and Dubbroek (11830 yr BP) meander scars consistentlyhinted at major channel downcutting and indicated meander-ing activity  before   the end of the Bølling. All evidence pointsto a Meuse that was certainly meandering during the LateBølling. Pastre  et al  . (1997) reconstructed strong incision andmeandering activity before 12400 yr BP for the Rivers Oiseand Marne, which have their upper and middle courses inthe same climatic and tectonic setting as the River Meuse.This contrasts with the River Warta in Poland (Vandenberghe et al  ., 1994), the response of which to Late-glacial warmingmight have been influenced by strong local isostatic crustalrebound effects due to disappearance of the ice-cover inthat region (e.g. fig. 3, Emery and Aubrey, 1985). The rapidchannel downcutting was the response to rapid climateamelioration after 14500  14 C yr BP, for it was related tothe radiocarbon time-scale of  Guiot and Couteaux (1992),and not to the calibrated time-scale of Hoek (1997), asKasse  et al  . (1999) wrongly postulate. Hoek (1997) was citedin the same sentence in order to illustrate the general effectof climate amelioration on permafrost and vegetation.The discrepancy between the oldest  14 C date (12500 yrBP,   13 C:  − 30.03, CaCO 3  ca. 7%) from the Meerlo channeland its Younger Dryas pollen assemblage might be attributedto the offset in sampling locations. We emphasise here thata Bølling/Allerød gyttja section was not represented in thepalynologically sampled coring, whereas we did find aclayey gyttja section. Kasse  et al  . (1999) mention the pres-ence of reworked pollen in the lowest sandy–clayey part of their core. This alternatively might indicate re-use of thedeep channel, which in turn unfortunately might have ledto erosion of the older material at their coring site. Transitional phase and gradient linereconstructions With respect to the discussion on distinguishing between a‘low-sinuosity and high-sinuosity meandering system’, werefer to Kasse  et al  . (1995), who introduced and used theseterms throughout (pp. 131–134). Surprisingly, Kasse  et al  .(1999) now claim that the sinuosity of meandering channels cannot   be used to distinguish between the different fluvialgenerations. However, this was an important criterion intheir srcinal conceptual model. How else then should wedistinguish the palaeofloodplain of their transitional phase(Huisink, 1999),  in casu   low-sinuosity meandering system(Kasse  et al  ., 1995: p. 131) from the Pleniglacial braidedsystem and Allerød high-sinuosity meandering system (p.132; Huisink, 1999) respectively?In fact, our whole discussion started there. We attributeda large terrace section between Meerlo and Boxmeer (Fig.1) to a Late Allerød meandering river system. Kasse  et al  .(1995) and Huisink (1997, 1999) correlated the same—butundated—terrace section on the basis of its elevation andits low-sinuosity channel pattern to a dated (Early Bølling)  97CORRESPONDENCE Figure 1  Late-glacial terraces in the North Limburg area (after Huisink, 1999). The added  14 C dates and pollen stratigraphies highlight thequestionable correlation of the braidplain from a fossilised Rhine branch (indicated between thick dashed lines) with the meanderingchannels of the Meerlo–Boxmeer section to the srcinal Bølling age Vierlingsbeek terrace (sections V-1 to V-4). Both pollen stratigraphyand  14 C dates indicate formation of the Beugen meander during the Late Bølling instead of the Allerød (Broekhuizen terrace, see questionmarks). palaeofloodplain northwest of Boxmeer. Next, the Meerlo–Boxmeer terrace section was thought to be part of theirEarly Bølling low-sinuosity transitional system (‘Vierlingsbeekterrace’, Fig. 1). Several facts show that both the inferredBølling age and the correlation cannot be maintained. Firstly,three well-curved channels and their infillings gave LateAllerød ages (Fig. 1: 11280, 11260 and 11170 yr BP).Obviously, these dates are far too young to attribute theMeerlo–Boxmeer section to a transitional system of EarlyBølling age. This is complicated even more if the bulk dates Copyright  © 2000 John Wiley & Sons, Ltd. J. Quaternary Sci., Vol. 15(1) 95–100 (2000) suffer from the aforementioned ageing effects as well. Re-use of the channels is unlikely, because two channelsshowed well-preserved almost purely organic gyttja sectionsat their very bottoms. Moreover, the lowermost gyttja of another curved channel laterally cutting off these two chan-nels indeed showed a younger age (11170 BP).Secondly, the correlation of the northern and southernparts of the Vierlingsbeek terrace (Huisink, 1999) on basisof both their channel patterns and terrace elevation is ratherdoubtful. The northern part (Fig. 1: sections V-1 and V-2) is  98 JOURNAL OF QUATERNARY SCIENCE clearly the western extension of a fossilised Rhine branch(Van de Meene and Zagwijn, 1978) and has straight, braidedchannels. Teunissen and De Man (1981) and Teunissen(1990) dated the earliest channel infillings in this area usingboth pollen stratigraphy and conventional  14 C to ca. 12700yr BP (Early Bølling). The southern part (Fig. 1: sectionsV-3 and V-4; Meerlo–Boxmeer) on the other hand, is clearlyof Meuse srcin and has well-curved meandering channels.Their earliest infillings were dated conventionally  and   bypollen stratigraphy to ca. 11200 yr BP (Tebbens  et al  ., 1999).Kasse  et al  . (1999) corroborate that the Holthees channel isa ‘ true meandering channel  ’. However, with this channelthey exemplify the end of their Bølling-age  transitional   phase.Apart from the confusing question of whether this channelbelongs either to a meandering or to a transitional system,it had an age of 11170 yr BP (Fig.1: site 14). Together withthe other residual channels of sites 4 and 13 (Fig. 1), ittherefore should be attributed to a Late Allerød meanderingriver system. The correlation with the clearly braided, Rhine-influenced palaeofloodplain dating from before 12700 yr BPnorthwest of Boxmeer is thus highly improbable oruntenable.Soil and geomorphological maps (Stiboka, 1976; Buiten-huis and Wolfert, 1988) locally indicate coversands andfrequently show anthropogenic horizons (0.5 m up to 1.2 m)on top of the Meerlo–Boxmeer terrace section. Kasse  et al  .(1995) and Huisink (1997, 1999) obtained their altitudepoints from the highest terrace parts and did not correct forthese horizons. Therefore, we have reason to doubt theirterrace level elevations and gradient reconstructions. Forexample, high surface elevation has led Van den Broek andMaarleveld (1963) and Kasse  et al  . (1995) to believe thatthe Looiveld area was of Pleniglacial age. Our  14 C date(10950 yr BP) from a buried channel indicated a muchyounger, Late Allerød age. A discussion of whether theoverlying cover is aeolian and/or anthropogenic is notimportant: in either case they should be subtracted from theterrace elevation before calculating its gradient line andbefore it is compared to other terrace levels lacking thesehorizons. Of course, correcting terrace elevations for suchhorizons is not the most elegant way to obtain terracegradients. However, we feel that larger errors are made if one does not correct for these overlying materials, whenknowing from the literature, soil maps and geomorphologicalmaps that they are widespread on the older terraces butlacking on the youngest ones. How else can one distinguishbetween genetically different terraces if their elevations differsome one or two metres and if they have (or lack) overlyinganthropogenic horizons varying from 0.5 to 1.2 m thick?We extensively studied and sampled a well-exposed typi-cal fining-upward sequence in the Lottum construction pitand the fluvio-aeolian sequences of the sandpits Panheeland Grubbenvorst (Tebbens  et al  ., 1998). Thus, our col-leagues could at least have given us some credit that we areable to distinguish aeolian coversands from fluvial meanderpointbar or overbank sediments. Indeed, the Broekwegcoring in the Holthees channel pointbar (the star in Fig. 1)shows a 3 m fining-upward sequence, but this  ends   at 3 mbelow surface, where a hiatus is indicated (Huisink, 1999;pp. 100 and 102–103). The same coring also shows that atleast the upper 1.2 m overlies a buried soil and thus mostprobably is an anthropogenic soil, as we had established inthe field. The 1:10000 topographical map indicates typicalaeolian relief in the immediate vicinity of the Broekwegcoring. Moreover, fieldwork indicated well-sorted, well-rounded, yellowish fine sand with very little clay in it untila depth of some 3 m. This clearly differed from loamy Copyright  © 2000 John Wiley & Sons, Ltd. J. Quaternary Sci., Vol. 15(1) 95–100 (2000) deposits and less well-sorted, sharp and greyish fine sandfrequently containing loamy layers in several other cored orstudied true fluvial pointbars and overbank deposits. So,only after very careful consideration and several field checks,were terrace altitude points corrected for overlying material.Therefore, we most strongly contradict the statement of Kasse et al  . in their last paragraph, suggesting that (all) ‘upper partsof fluvial fining-upward sequences have been interpretedincorrectly as aeolian deposits’.Considering the fact that Kasse  et al  . (1995, 1999) andHuisink (1997, 1998, 1999) interpreted asynchronous terrace remnants to belong to the same genetic lithostratigraphicallevel, we were not surprised that our gradient line recon-structions differed from theirs. We correlated only thoseremnants from which we were absolutely certain to whichgeneration they belonged, based on our  14 C dates andadditional pollen information of the palaeochannels. Inaddition, we averaged some 35 1:10000 altitude points perpointbar or terrace remnant (we should have mentionedthis), corrected them for overlying material only where geo-morphological and soil maps combined with field infor-mation gave cause for this and then plotted this point per-pendicular to the mid-valley palaeo-axis. For this reason ournumber of data points seems to be much smaller than inHuisink (1998, 1999: p. 94), for which by the way abouthalf of the points stemmed from the fossilised Rhine branch(see Fig. 1: sections V-1 and V-2). In our opinion, thesepoints should not be involved in reconstructing gradientlines for the River Meuse. Actually, it is not the number of data points that is important, but their quality. For example,the gradient reconstruction for the transitional level(Vierlingsbeek terrace) incorporates uncorrected altitudepoints obtained from: (i) the fossilised Rhine branch olderthan 12700 yr BP (Fig. 1: sections 1–2), (ii) possibly cover-sand–overblown remnants with anthropogenic horizons nearBoxmeer (sections 2–3, Buitenhuis and Wolfert, 1988) and(iii) the ca. 11200 yr BP Meerlo–Boxmeer section (section4), with anthropogenic soils up to 1.2 m thick.In any case, the different palaeofloodplain gradients cannothave been caused by the ‘initial manipulation’ of our mid-valley palaeo-axis. The fact that our Holocene gradient of 11.6 cm km − 1 compares well with the Holocene gradient of Huisink (1999: 11.99 cm km − 1 , no overlying material!) showsthat we have used a very similar projection axis. The gradi-ents of the Milsbeek and Gennep terraces—located in thefossilised Rhine branch now occupied by the River Niers(Fig. 1)—are much steeper compared with their Late-glacialMeuse equivalents (Huisink, 1999). The gradients of Berendsen  et al  . (1995) and Verbraeck (1984) relate to down-stream parts of this branch, where Rhine influence cannotbe excluded. The Rhine has a much higher discharge andhigher sediment load and, therefore, their gradients aresteeper than our pure Meuse gradients. Van den Broek andMaarleveld (1963) and ourselves have studied the Venlo–Boxmeer terraces. In 1963, the terrace remnants for thisparticular area were correlated by their altitudes instead of relating them to  14 C or pollen dated palaeochannels. Forexample, the entire Meerlo–Boxmeer section, the Looiveldarea and the Meerlo meander scar were assumed to be of Pleniglacial age (‘Terrace 1’) in 1963. Currently,  14 C datingand pollen stratigraphy indicate Late Allerød ages and aBølling age (or, when following Kasse  et al  .: Allerød) respect-ively. Thus, the 1963 terrace remnant correlations and gradi-ent lines are literally outdated and that is why we adaptedthem. Correlation of asynchronous terrace remnants to onegenetic level must have led to different, steeper gradientlines in 1963 compared with ours.  99CORRESPONDENCE In conclusion, our bulk dates were chronologically andspatially coherent and generally matched the pollen stratigra-phy. It has still not been proven that all our bulk dates weretoo old. Thus, we see no reason to reject any of them andargue that they should be seen as relevant, complementaryand independent dates. We agree that AMS measurementsmight shed more light on the local chronostratigraphy, oncondition that sample preparation and measurement are per-formed according to Wohlfahrt  et al  . (1998).Our interpretations aimed at reconstructing the downcut-ting history of the Meuse to provide a consistent strati-graphical framework for its residual channel infillings. Wehave never  denied   the  existence   of a transitional river system,as Kasse  et al  . (1999) claim on page 91 and thus ourinterpretations are not contradictory to our own findings.Instead, we showed that the main source of disagreementin the gradient line reconstructions arose from questionableterrace remnant correlations. The major, southern part of thesrcinal Bølling-age Vierlingsbeek transitional system  sensu  Vandenberghe  et al  . (1994), Kasse  et al  . (1995) and Huisink(1997, 1999) appears to be a relict Late Allerød floodplainformed by a (low-sinuosity?) meandering Meuse. The north-ern part clearly is the palaeofloodplain of a fossilised, braid-ing Rhine branch that presumably was in use until the LatePleniglacial. Therefore, this part should neither be correlatedto an Allerød meandering system, nor should it be used inreconstructing River Meuse gradient lines or River Meuseresponse to climate change. Nevertheless, altitude pointsfrom this area were included to reconstruct a gradient forthe transitional Vierlingsbeek Meuse terrace (Huisink, 1999).Gradient lines were reconstructed using highest terrace alti-tude points, but these were not critically examined for knownoverlying material of varying thickness. Consequently, sev-eral altitude points srcinating from non-fluvial surfaces maybe too high. Thus, the gradient lines of   both  higher andlower levels of Kasse  et al  . (1995) and Huisink (1999) needto be changed, as either incorrect correlations were madeor higher levels might now incorporate remnants that shouldbe assigned to lower and genetically younger levels.Finally, both a basic assumption to distinguish the tran-sitional system from other fluvial generations and its inferredage have now been cancelled. Thus, we are left wonderinghow the transitional system should then be defined andwhether the Meerlo–Boxmeer terrace section concerned trulyreflects River Meuse response to early Late-glacial climaticwarming. Instead, a response to gradual cooling during theAllerød seems more likely. We never intended to presentthese interpretations as the absolute truth and had an openmind to reconcile new data with published data. However,if these new data—after careful scrutiny—are still seriouslyconflicting with a conceptual stratigraphical model or evenfalsify it, then a new interpretation is inevitable. We greatlyappreciate it that our colleagues apparently share this viewwith respect to their own work too. Their adapted terracemap (Fig. 1, page 92) shows that at least one residualchannel/meander pointbar combination (out of three), for-merly belonging to the Bølling age Vierlingsbeek transitionalterrace, is now attributed to the Allerød age Broekhuizenmeandering terrace. References Berendsen HJA, Hoek WZ, Schorn EA. 1995. Late Weichselian andHolocene river channel changes in the rivers Rhine and Meuse Copyright  © 2000 John Wiley & Sons, Ltd. J. Quaternary Sci., Vol. 15(1) 95–100 (2000) in the Netherlands (Land van Maas en Waal). In Pala¨ oklimaforschung/Palaeoclimate Research,  Vol. 14, Frenzel B,Vandenberghe J, Kasse C, Bohncke S, Gla¨ser B (eds). Fischer:Stuttgart; 151–171.Bohncke SJP, Vandenberghe J, Huijzer AS. 1993. Periglacial environ-ments during the Weichselian Late Glacial in the Maas valley,the Netherlands.  Geologie en Mijnbouw   72 : 193–210.Bo¨ttger T, Hiller A, Junge FW, Litt T, Mania D, Scheele N. 1998.Late Glacial stable isotope record, radiocarbon stratigraphy, pollenand mollusc analyses from the Geiseltal area, Central Germany. Boreas   27 : 88–100.Buitenhuis A, Wolfert HP. 1988.  Geomorfologische kaart van Neder- land (Geomorphological map of the Netherlands)  . Scale 1:50,000Blad 46 Gennep, Toelichting op kaartblad 46 Gennep.Emery KO, Aubrey DG. 1985. Glacial rebound and relative sealevels in Europe from tide-gauge records.  Tectonophysics   120 :239–255.Guiot J, Couteaux M. 1992. Quantitative climate reconstructionfrom pollen data in the Grand Duchy of Luxembourg since 15000yr BP.  Journal of Quaternary Science   7 : 303–309.Hoek WZ. 1997. Palaeogeography of Lateglacial vegetations.Aspects of Lateglacial and Early Holocene vegetation, abioticlandscape, and climate in The Netherlands. PhD Thesis, FreeUniversity Amsterdam; 146 pp.Huisink M. 1997. Late-glacial sedimentological and morphologicalchanges in a lowland river in response to climatic change; theMaas, southern Netherlands.  Journal of Quaternary Science   12 :209–223.Huisink M. 1998. Tectonic versus climatic controls on the RiverMaas dynamics during the Late Glacial. In  Palaeohydrology and Environmental Change  , Benito G, Baker V, Gregory KJ (eds).Wiley: Chichester; 99–109.Huisink M. 1999. Lateglacial river sediment budgets in the Maasvalley, The Netherlands.  Earth Surface Processes and Landforms  24 : 93–109.Kasse C, Vandenberghe J, Bohncke SJP. 1995. Climatic change andfluvial dynamics of the Maas during the Late Weichselian andEarly Holocene. In  Pala¨ oklimaforschung/Palaeoclimate Research,Vol. 14 , Frenzel B, Vandenberghe J, Kasse C, Bohncke S, Gla¨serB (eds). Fischer: Stuttgart; 123–150.Kasse C, Huisink M, Hoek WZ, Vandenberghe J. 1999. Comment:Fluvial incision and channel downcutting as a response to Late-glacial and Early Holocene climate change: the River Meuse(Maas), The Netherlands.  Journal of Quaternary Science   15 :91–94.Pastre J-F, Fontugne M, Kuzucuoglu C, Leroyer C, Limondin-LozouetN, Talon M, Tisne´rat N. 1997. The evolution of river beds in theNorth-East region of Paris (France) during the Lateglacial and theHolocene. Relation to the paleoenvironments and to the humanimpact on the slopes.  Ge ´ omorphologie: relief, processus,environnement   4 : 291–312.Stiboka, 1976.  Bodemkaart van Nederland (Soil map of the Netherlands)  . Scale 1:50,000, Blad 46 West/Oost Vierlingsbeek.Tebbens LA. 1999.  Late Quaternary evolution of the Meuse fluvial system. A reconstruction based on bulk sample geochemistry and forward modelling  . PhD thesis, Wageningen University; 155 pp.Tebbens LA, Veldkamp A, Kroonenberg SB. 1998. The impact of climate change on the bulk and clay geochemistry of fluvialresidual channel infillings: the Late Weichselian and Early Holo-cene River Meuse (Maas) sediments (The Netherlands).  Journal of Quaternary Science   13 : 345–356.Tebbens LA, Veldkamp A, Westerhoff W, Kroonenberg SB. 1999.Fluvial incision and channel downcutting as a response to Late-glacial and Early Holocene climate change: the lower reach of the River Meuse, The Netherlands.  Journal of Quaternary Science  14 : 59–75.Teunissen D. 1983. The development of the landscape of the naturereserve De Hamert and its environs in the northern part of theprovince of Limburg, The Netherlands. In  Developments in Physi- cal Geography—a Tribute to J.I.S. Zonneveld  , Terwindt JHJ, VanSteijn H (eds).  Geologie en Mijnbouw   62 : 569–576.Teunissen, D. 1990.  Palynologisch onderzoek in het Oostelijk rivier- engebied: een overzicht  . Mededelingen van de Afdeling Biogeolo-