JOURNAL OF ENVIRONMENTAL GEOGRAPHY Journal of Environmental Geography 10 (1–2), 53–59. DOI: 10.1515/jengeo-2017-0006 ISSN: 2060-467X DATING THE HOLOCENE INCISION OF THE DANUBE IN SOUTHERN HUNGARY Orsolya Tóth1, György Sipos1*, Tímea Kiss1, Tamás Bartyik1 1Department of Physical Geography and Geoinformatics, University of Szeged, Egyetem u. 2-6, H-6722 Szeged, Hungary *Corresponding author, e-mail: gysipos@geo.u-szeged.hu Research article, received 10 April 2017, accepted 20 June 2017 Abstract The alluvial development of the Great Hungarian Plain has greatly been determined by the subsidence of different areas in the Pan- nonian Basin. The temporal variation of subsidence rates significantly contributed to the avulsion and shifting of main rivers. This was the case in terms of the Hungarian Lower Danube when occupying its present day N-S directional course. The considerable role of tectonic forcing is also supported by the presence of different floodplain levels. Although, several channel forms are identifiable on these the timing of floodplain development has been reconstructed up till now mostly by the means of geomorphological analysis, and hardly any numerical dates were available. The main aim of this study is to provide the first OSL dates for palaeo-channels lo- cated on the high floodplain surface of the Hungarian Lower Danube, and to determine the maximum age of low and high floodplain separation on the Kalocsa Plain. For the analysis two meanders were sampled close to the edge of the step slope between the two levels. According to the results, the development of the investigated palaeo-meanders could be rapid. The formation of the older meander was dated to the Late Atlantic, while the possible separation of the high and low floodplain surfaces could start in the be- ginning of the Subboreal Phase. Keywords: Hungarian Lower Danube, floodplain development, Holocene, OSL dating INTRODUCTION The alluvial development of the Great Hungarian Plain (GHP) has greatly been determined by the selective subsidence of different areas in the Pannonian Basin. The significance of tectonic control on fluvial processes has been indicated by several earlier research (e.g. So- mogyi, 1961; Borsy, 1992; Gábris and Nádor, 2007; Kiss et al., 2015). The temporal variation of subsidence rates at different areas lead to the time-to-time avulsion and shifting of the main rivers and related tributaries. Major shifts however were also influenced by geomorphologi- cal processes, namely alluvial fan building and subse- quent sliding off the rivers from these elevated surfaces. Concerning the Danube a significant, but presuma- bly gradual diversion occurred during the Late Würm Period (Pécsi, 1959; Pécsi, 1991; Mezősi, 2011) on its GHP section, when the river had shifted from its earlier NW-SE course to a N-S direction by sliding off its allu- vial fan, the Danube-Tisza Interfluve (Mezősi, 2011). The process can also be explained by the activation of the Kalocsa and Baja Depressions, located near the Hun- garian-Serbian border. According to Jaskó and Krolopp (1991), this subsidence zone attracted the Danube to its present direction at around 30-40 ka, and fault lines related to the zone are still active, though only moderate- ly (Mezősi, 2011). Obviously, in case a river shifts to a depression, let it be of geomorphic or tectonic origin, incision is gener- ated, which propagates upstream and leads to the devel- opment of alluvial fan terraces indicating the major phases of changes (Bridge, 2003; Schumm, 1979). The age of fluvial forms on the abandoned floodplain (now terrace level) corresponds to the maximum, while that of active floodplain forms to the minimum age of incision. The numerical dating of fluvial forms therefore is crucial to reconstruct the timing of terrace formation and conse- quently the phases and sometimes even the rate of sub- sidence (Knighton, 1998). The strath terraces of the Danube on its Hungarian upland section had been intensively studied with classi- cal and more modern dating methods to assess the uplift rate of the adjacent members of the Transdanubian Mountains (Pécsi, 1991; Ruszkiczay-Rüdiger et al., 2005; Gábris, 2013). Concerning the downstream GHP section of the river Pécsi (1959, 1967) investigated the geomorpho- logical units, terrace materials and development of the floodplain and its levels. He concluded that the North- ern part of the present GHP floodplain (Csepel -Solt Plain) is older than the Southern (Kalocsa Plain), and that the upper 10-20 m of sediments is almost com- pletely reworked by the lateral movement of the Dan- ube. Nevertheless, in both areas a low and a high floodplain surface (terrace) can be identified. T he term high floodplain refers to the fact that although this surface lies 1-2 m higher it is still inundated by extreme floods. Consequently, Pécsi (1967) claims that fluvial deposits on the low and high floodplain cannot be separated in age, though the d evelopment of 54 Tóth et al. (2017) the high floodplain can clearly be related to the Late Pleistocene activation of the Baja and Kalocsa De- pressions. So far the timing of floodplain development in the area has been reconstructed mostly by the means of geomorphological analysis, hardly any numerical dates are available concerning the sediments and forms of the high floodplain surface. Two radiocarbon dates from subfossil drift woods placed the age of the right bank side high floodplain surface of the Danube, just opposite to the Kalocsa Plain, to cca. 40 000 and 11 000 BP (Her- telendi et al., 1991). In terms of Late Pleistocene and Holocen terrace surfaces, where the fluvial forms are still identifiable a straightforward method of dating is using optically stimulated luminescence (OSL), which determines the last exposure of sediments to sunlight, i.e. the time of sediment deposition. The method has been extensively applied in studies related to the dating of floodplain surfaces along several rivers (e.g. Kiss et al., 2013; Olszak et al., 2016; Ruszkiczay-Rüdiger et al., 2016; Meng et al., 2015). By considering the above, the main aim of the pre- sent study is to provide the first OSL dates for palaeo- channels located on the high floodplain surface of the Hungarian Lower Danube, and this way to provide a maximum age to the separation of the low and high floodplain levels on the Kalocsa Plain. STUDY AREA AND SAMPLING The study area is located on the Kalocsa Plain, which is situated at the middle section of the Hungarian Lower Danube (Fig. 1). The floodplain area is characterised by two floodplain levels: the elevation of the lower flood- plain surface (A-level) is between 90 and 92 m asl, while that of the higher floodplain surface (B-level) is between Fig. 1 Location of the study area, geomorphological setting, and sampling points Dating the Holocene incision of the Danube in Southern Hungary 55 92 and 94 m asl. Eastwards the study area is adjacent to the Danube-Tisza Interfluve covered by sand dunes and characterised by a relief between 96 and 105 m asl. Westwards loess plains border the Danube plains, with an elevation reaching even 150 m asl (Fig. 1). Sediment samples were collected from two pal- aeo-meanders located on the B-level to determine the maximum age of floodplain surface differentiation, and the time of active B-level floodplain develop- ment. The two selected palaeo-channels belong to two separate meander systems. Based on the geomorpho- logical map (Fig. 1), both systems developed through chute cut-offs and consequent meander growth, being characteristic in case of meanders on the lower flood- plain as well. In case of the eastern meander (width=170 m, R=1,7 km) one drilling was made (AE1) to sample channel sediments, as point bars were hardly recog- nisable on the field (Fig. 1). At the AE1 drilling point 3 OSL samples were collected from 60, 130 and 190 cm (AE1/1, AE1/2, AE1/3). In terms of the western meander (width: 370 m, R=720 m), right on the edge of the B-level, point bars were sampled at three loca- tions (AE2, AE3, AE4), from depths 70, 70 and 110 cm, respectively. OSL samples were mostly taken from layers of pure sand, however in terms of AE1 upper two sam- ples were categorised on the field as sandy silt. Undis- turbed sampling was made with steel cylinders appli- cable to an Eijkelkamp hand drill system. Samples weighed approximately 200 g. Background samples were also collected from above and from below each OSL sample. METHODS The geomorphological map of the study area was compiled on the basis of 1:10 000 scale topographical maps, with 1 m contour line interval, occasionally supplemented by 0.5 m contour lines, The age of sediment samples was determined by optically stimulated luminescence. Samples were either composed of medium sand, or the mixture of fine sand and silt, though containing an adequate amount of medium sand in the latter case as well, thus the so called coarse grain quartz dating procedure was applied. In terms of fluvial samples usually the sand sized quartz fraction is investigated anyway, assuming that it usually has more chance for complete bleaching during sediment transport. The preparation of the samples followed usual laboratory techniques (Aitken, 1998; Mauz et al., 2002). After removing the samples from the cylinders they were dried and the 90-150 μm fractions was separated by sieving. The carbonate and organic material content was removed by repeated treatment in 10% HCl and 10% H 2O2. A Na- polytungstanate (LST Fastfloat) heavy liquid flotation was applied for the separation of the quartz fraction. This step was followed by a 50 min etching in 40% HF, aiming at removing any remaining feldspar con- taminations and the outer layer of quartz. Purified quartz grains were adhered to stainless steel discs of 10 mm diameter by silicone spray. For OSL a Ø6 mm mask for the final measurements Ø2 mm mask was applied to control the number of grains on a disc. A number of aliquots were prepared for luminescence tests and for equivalent dose (De) determination. Measurements were made using a RISOE DA-15 TL/OSL luminescence reader by applying the single aliquot regeneration (SAR) protocol (Wintle and Mur- ray 2006). A preheat test was used for determining optimal heating parameters during the SAR measurements. Preheat temperatures were varied between 180 oC and 300 oC. During the tests 1) SAR recycling ratios (ratio of two sensitivity corrected luminescence signals generated by identical regeneration doses) ; 2) recu- peration (thermal and photo transfer of electrons to OSL traps); and 3) dose recovery (ratios recycling ratio being within 1.00 ± 0.05, De error being lower than 10%, recuperation being lower than 5%) were monitored to determine the best thermal treatment. A combined preheat and dose recovery test was per- formed. Preheat temperature was increased 200 °C SAR measurements were performed on 48–144 ali- quots, depending on the proportion of acceptable measuremets. Acceptability was assessed using the standard rejection criteria for each aliquot. Thresholds of rejection were the following: recycling ratio being within 1.00 ± 0.05, De error being lower than 10%, recuperation being lower than 5%. Possible feldspar contamination was also monitored by measuring IRSL/OSL depletion ratio at the end of the SAR pro- cedure. The first 0.5 s of OSL curves was taken as the signal, the last 10 s as the background of measure- ments. Sample De was calculated from aliquot De using either the minimum age model (MAM) (AE1/1, AE1/2), or the central age model (CAM) (AE1/3, AE2, AE3, AE4) depending on the dispersion De val- ues (Galbraight et. al 1999). Environmental dose rate (D*) was determined by using high-resolution, ex- tended range gamma spectrometry (Canberra XtRa Coaxial Ge detector), using 500 cm3 marinelli beak- ers. Dry dose rates were calculated using the conver- sion factors of Adamiec and Aitken (1998). Wet dose rates were assessed on the basis of in situ water co n- tents. The rate of cosmic radiation was determined by considering burial depth following the method of Prescott and Hutton (1994). RESULTS AND DISCUSSION During the evaluation, a major problem was the low luminescence signal intensity and the low sensitivity of the samples (Fig. 2). Consequently, numerous ali- quots did not pass the necessary criteria and were finally rejected from De calculation. Therefore, a high number of aliquots were measured finally to get an adequate amount of results for the statistical analysis 56 Tóth et al. (2017) of equivalent doses. In average 30-40% of the ali- quots turned to be acceptable for sample De assess- ment (Table 1). In case of preheat tests the problem of low sensi- tivity was a less significant issue, as due to the higher number of grains (Ø6 mm mask ) on the measurement discs luminescence response was considerably higher. The most appropriate preheat temperature for all samples was 200 °C for the SAR measurements (Fig. 3), since the dose recovery ratio at this temperature was well within the limits of acceptability. Nevertheless, the spread of the recycling ratios was wider in some cases and recu- peration was over 5 % in case of samples AE1/2 and AE4. Consequently, the final SAR measurements were carried out with using a hot bleach treatment, i.e. insert- ing a high temperature (280°C) optical bleaching at the end of each measurement cycle. By the application of the right temperature treatment the samples performed ade- quately to retrieve reliable results and ages. Regarding drilling point AE1 the sampled sand lay- ers at 70, 130 and 190 cm depth were dated to 6.7±0.6 ka, 7.2±0.4 ka and 6.1±0.5 ka, respectively. The lowermost sand layer (AE1/3) was characterised almost exclusively by medium sand, and the distribution of individual De results showed a central tendency (Fig. 4), consequently, it is suggested that this sample had undergone the most complete resetting process during transportation and sed- imentation. Therefore, both from a sedimentological, and Fig. 2 Figures A and B show an appropriate shine-down curve (A) and dose response curve (B), meanwhile C and D figures represent the low intensity Table 1 Dose rate, equivalent dose and age data of the investigated samples Sample Aliquots (used/measured) Depth (m) Moisture content (%) U (ppm) Th (ppm) K (%) D* (Gy/ka) De (Gy) Age (ka) AE1/1 30/72 0.6 13±1.3 2.95±0.02 5.05±0.06 1.38±0.04 2.80±0.08 20.27±0.82 6.7±0.6 AE1/2 27/48 1.3 21±2.1 2.86±0.04 5.59±0.07 1.36±0.04 2.46±0.07 24.79±0.86 7.2±0.4 AE1/3 25/96 1.9 33±2 2.76±0.03 6.13±0.07 1.34±0.04 2.37±0.07 18.49±1.61 6.1±0. 5 AE2 26/72 0.7 16±2 2.87±0.02 4.53±0.06 1.22±0.04 2.49±0.07 11.76±0.65 4.7±0.3 AE3 35/64 0.7 7±2 1.80±0.02 3.63±0.04 1.13±0.03 2.27±0.07 11.05±0.5 4.9±0.3 AE4 58/144 1.1 7±2 2.22±0.02 3.68±0.04 1.18±0.03 2.55±0.08 12.87±1.07 5.0±0.5 Dating the Holocene incision of the Danube in Southern Hungary 57 from an OSL point of view this sample can represent the time of major fluvial activity. This is in correspondence with the major findings of Tóth et al. (in press), who measured almost complete resetting in terms of modern coarse grain sediments along the present day Danube. In the meantime the upper two samples yielded somewhat higher ages than AE1/3, which can be ex- plained by their different sedimentology, i.e. the high proportion of fines, which refers to a post formational, lower energy deposition, being less favourable in terms of OSL signal resetting and leading to possible age over- estimation. Based on the results, in case of sample AE1/2, characterised by the finest grain size, the age could be overestimated by more than 1 ka, while in case of the coarser AE1/1 by more than 0.6 ka (Fig. 5). The ages of samples AE2, AE3 and AE4 were very similar and stayed within error, AE4: 4.7±0.3 ka, AE6: 4.9±0.3 ka, AE7: 5.0±0.5 ka (Table 1). As the formation age of point bars cannot be separated, results refer to a relatively fast meander development between 4.7 and 5.0 ka, reflecting the dynamic nature of Danube fluvial ac- tivity. Based on the first age results from the area, the Danube was actively forming its present day high flood- plain up till 5 ka. Consequently, the maximum age of incision and the development of the present day flood- plain can be placed to the beginning of the Subboreal Phase. However, in order to question or reinforce the morphological interpretation of Pécsi (1967), i.e. the formation time of the higher and lower floodplain levels are hard to separate, needs further investigation in the area, especially focusing on lower floodplain palaeo- channels. Nevertheless, the OSL dates presented here give a framework for further studies, and also imply that if there is a separation time between the two levels, it must be in the second half of the Holocene. Fig. 3 Preheat and dose recovery tests of the samples. The red line and the dashed line mark the three criteria and the errors of the criteria Fig. 4 Dose distribution of individual De of AE1/2 (A) and AE1/3 (B). In case of AE1/2 MAM, meanwhile in case of AE1/3 CAM was used for the calculations 58 Tóth et al. (2017) CONCLUSIONS Floodplain development along the Hungarian Lower Danube has previously been reconstructed mainly on the basis of geomorphological evidence. The present study has provided the first OSL dates concerning the fluvial forms of the high floodplain surface on the Kalocsa Plain and therefore on the entire GHP section of the river. The OSL sensitivity of the sampled sediments is low, therefore a high number of measurements is need- ed to get the necessary amount of results for the relia- ble statistical analysis of equivalent doses. However, by the application of the right temperature treatment the samples performed adequately to retrieve reliable re- sults and ages. The development of the investigated palaeo- meanders could be rapid. In case of the older meander the age of channel forming fluvial activity can be in- ferred from the lowermost sample, yielding a 6.1±0.5 ka age placing the development of the meander to the Late Atlantic Phase. In terms of the younger meander, on the edge of the high floodplain the ages of consecu- tive pointbars were in the range of 4.7±0.3 ka and 5.0±0.5 ka, meaning, that the separation of the high and low surfaces could start in the beginning of the Sub- boreal Phase or later. Acknowledgements The research was funded by projects NKFI K 119309, NKFI K 119193 and HURO/1101/126 ENVIARCH References Adamiec, G., Aitken, M. 1998. Dose-rate conversion factors: update. Ancient TL 16 (2), 37–50. Aitken, M. J. 1998. An Introduction to Optical Dating. Oxford Univer- sity Press. London. Borsy, Z. 1992. Általános természetföldrajz. Nemzeti Tankönyvkiadó, Budapest. (In Hungarian) Bridge, J. S. 2003. Rivers and Floodplains. Form, Processes and Sedi- mentary Record. Blackwell Science LtD. Galbraith, R. F., Roberts, R. G., Laslett, G. M., Yoshida, H., Olley, J. M. 1999. Optical dating of single and multiple grains of quartz from Jinmium rock shelter, northern Australia. Part I: Experi- mental design and statistical models. Archeometry 41, 339–364. DOI: 10.1111/j.1475-4754.1999.tb00987. Gábris, Gy. 2013. A folyóvízi teraszok hazai kutatásának rövid áttekin- tése – A teraszok kialakulásának és korbeosztásának új magya- rázata. Földrajzi Közlemények 137 (3), 240–247. (in Hungarian). Gábris, Gy., Nádor, A. 2007. Long-term fluvial archives in Hungary: response of the Danube and Tisza rivers to tectonic movements and climatic changes during the Quarternaly: a review and new synthesis. Quarternaly Science Reviews 26, 2758-2782. DOI: 10.1016/j.quascirev.2007.06.030 Hertelendi, E., Petz, R., Scheuer, Gy., Schweitzer, F. 1991. Radiocar- bon age of the formation in the Paks Szekszárd depression. – In: Pécsi, M., Schweitzer, F. (eds) Quaternary environment in Hun- gary: contribution of the Hungarian National Committee to the XIIIth INQUA Congress Beijing, China. Akadémiai Kiadó, Bu- dapest, 85–89. Jaskó, S., Krolopp, E. 1991. Negyedidőszaki kéregmozgások és folyó- vízi üledékfelhalmozódás a Duna-völgyében Paks és Mohács között. A Földtani Intézet Évi Jelentése 1989-ről, 65–84. Kiss, T., Hernesz, P., Sümeghy, B., Györgyövics, K., Sipos, Gy. 2015. The evolution of the Great Hungarian Plain fluvial system – Fluvial processes in a subsiding area from the beginning of the Weichselian. Quaternary International 388 (3), 142–155. DOI: 10.1016/j.quaint.2014.05.050 Kiss, T., Sümeghy, B., Hernesz, P., Sipos, Gy., Mezősi, G. 2013. Az Alsó-Tisza menti ártér és a Maros hordalékkúp késő-pleisztocén és holocén fejlődéstörténete. Földrajzi Közlemények 137: (3) 269– 277. (In Hungarian) Knighton, D. 1998. Fluvial forms and processes. Hodder Arnold Publi- cation, London. Mauz, B., Bode, T., Mainz, E., Blanchard, H., Hilger, W., Dikau, R., Zöller, L. 2002. The luminescence dating laboratory at the Uni- versity of Bonn: Equipment and procedures. Ancient TL 20, 53– 61. Meng, Y.M., Zhang, J.F., Qui, W.L., Fu, X., Guo, Y.J., Zhou, L.P. 2015. Optical dating of the Yellow River terraces in the Mengjin area (China). Quaternary Geochronology, 30, 219–225. DOI: 10.1016/j.quageo.2015.03.006 Fig. 5 The results of the two sample site. A: The AE1 drilling point, B: The ages of AE1 drilling point, C: The ages of the second site point bars Dating the Holocene incision of the Danube in Southern Hungary 59 Mezősi, G., 2011. Magyarország természetföldrajza. Akadémiai kiadó, Budapest (in Hungarian) Olszak, J., Kukulak, J., Alexandersonm H. 2016. Revision of river terrace geochronology in the Orawa-Nowy Targ Depression, so- uth Poland: insights from OSL dating, Proceedings of the Geo- logists' Association, 127 (5), 595–605. DOI: 10.1016/j.pgeola.2016.09.004 Pécsi, M. 1959: A magyarországi Duna-völgy kialakulása és felszín- alaktana. Akadémiai kiadó, Budapest, 345. (in Hungarian) Pécsi, M. 1967. A dunai Alföld. Akadémiai kiadó, Budapest (in Hun- garian) Pécsi, M. 1991. A magyarországi Duna-völgy teraszai és szintjei. In: Pécsi, M., Geomorfológia és domborzatminősítés. MTA FKI, Budapest, 36–47. (in Hungarian) Prescott, J. R., Hutton, J. T. 1994. Cosmic ray contributions to dose rates for luminescence and ESR dating: large depths and long- term time variations. Radiation Measurements 23, 497–500. DOI: 10.1016/1350-4487(94)90086-8 Ruszkiczay-Rüdiger, Zs., Dunai, T. J., Bada, G., Fodor, L., Horváth, E. 2005: Middle to late Pleistocene uplift rate of the Hungarian Mountain Rnage at the Danube Bend, (Pannonian Basin) using in situ produced 3He. Tectonophysics, Vol. 410, Iss. 1-4. 173- 187. pp. Ruszkiczay-Rüdiger, Zs., Braucher, R., Novothny, Á., Csillag, G., Fodor, L., Molnár, G., Madarász, B., ASTER Team. 2016. Tectonic and climatic control on terrace formation: Coupling in situ produced 10Be depth profiles and luminescence approach, Danube River, Hungary, Central Europe. Quaternary Science Reviews 131, 127-147. DOI: 10.1016/j.quascirev.2015.10.041 Somogyi, S. 1961: Hazánk folyóhálózatának fejlődéstörténeti vázlata. Földrajzi Közlemények 9 (85), 25–50. (In Hungarian) Schumm, S. A. 1979: Geomorphic thresholds: The concept and its applications. Transactions of the Institute of British Geograp- hers 4, 485–515. DOI: 10.2307/622211 Tóth, O., Sipos, Gy., Kiss, T., Bartyik, T.,: Variation of OSL residual doeses in terms of coarse amd fine grain modern sediments along the Hungarian section of the Danube. Geochronometria (in press) Wintle, A. G., Murray, A. S. 2006. A review of quartz optically stimu- lated luminescence characteristic and their relevance in single- aliquot regeneration dating protocols. Radiation Measurement 41, 369–391. DOI: 10.1016/j.radmeas.2005.11.001