Global and Planetary Change 98–99 (2012) 109–121

Contents lists available at SciVerse ScienceDirect

Global and Planetary Change

j ourna l homepage: www.e lsev ie r .com/ locate /g lop lacha

A systematic overview of the coincidences of river sinuosity changes and tectonicallyactive structures in the Pannonian Basin

Judit Petrovszki a,c,⁎, Balázs Székely a,b, Gábor Timár a

a Department of Geophysics and Space Sciences, Eötvös University, Budapest, Hungaryb Institute of Photogrammetry and Remote Sensing, Vienna University of Technology, Vienna, Austriac Water Management Research Group of the Hungarian Academy of Sciences, Budapest, Hungary

⁎ Corresponding author.E-mail address: geojudit@gmail.com (J. Petrovszki).

0921-8181/$ – see front matter © 2012 Elsevier B.V. Allhttp://dx.doi.org/10.1016/j.gloplacha.2012.08.005

a b s t r a c t

a r t i c l e i n f o

Article history:Received 9 January 2012Accepted 16 August 2012Available online 23 August 2012

Keywords:Pannonian Basinsinuosityriversneotectonicsclassification

As tectonic movements change the valley slope (low-gradient reaches of valleys, in sedimentary basins), the allu-vial rivers, as sensitive indicators, respond to these changes, by varying their courses to accommodate this forcing.In our study sinuosity values, a commonly used characteristic parameter to detect river pattern changes,were stud-ied for the major rivers in the Pannonian Basin in order to reveal neotectonic influence on their planform shape.Our study area comprises the entire Pannonian Basin (330,000 km2) located in eastern Central-Europe,bounded by the Alps, Carpathians and Dinarides. The studied rivers were mostly in their natural meanderingstate before the main river regulations of the 19th century. The last quasi-natural, non-regulated river plan-forms were surveyed somewhat earlier, during the Second Military Survey of the Habsburg Empire. Using thedigitized river sections of that survey, the sinuosities of the rivers were calculated with different sample sec-tion sizes ranging from 5 km to 80 km. Depending on the bank-full discharge, also a ‘most representative’section size is given, which can be connected to the neotectonic activity.In total, the meandering reaches of 28 rivers were studied; their combined length is 7406 km. The placeswhere the river sinuosity changed were compared to the structural lines of the “Atlas of the present-daygeodynamics of the Pannonian Basin” (Horváth et al., 2006). 36 junctions along 26 structural lines were iden-tified where the fault lines of this neotectonic map crossed the rivers. Across these points the mean sinuositychanged. Depending on the direction of the relative vertical movements, the sinuosity values increased or de-creased. There were some points, where the sinuosity changed in an opposite way. Along these sections, therivers belong to the range of unorganized meandering or there are lithological margins.Assuming that the rivers indicate on-going faulting accurately, some placeswere found, where positioning of thefaults of the neotectonic map could be improved according to the sinuosity jumps. However, some significantsinuosity changes cannot be correlated to known faults. In these cases other factors may play a role (e.g., hydro-logical changes, increase of sediment discharge also can modify sinuosity). In order to clarify the origin of thesechanges seismic sections and other geodynamical information should be analyzed to prove or disprove tectonicrelationship if hydrological reasons can be excluded.

© 2012 Elsevier B.V. All rights reserved.

1. Introduction

In the last decade fluvial geomorphological analysis is increasinglyused in order to detect youngest neotectonic activity across faultzones. Over time river geomorphology is sensitive to changes in hydro-logical and sediment transport conditions (e.g. due to climate changeand humans), but in space the rivers can respond spectacularly tochanges in valley gradient, topography and substratewhich are affectedby neotectonics. If a fault causes a displacement that has a vertical com-ponent, the rivers of the area tend to change their courses and theirstyle, often good indication of fault activity. According to Schumm(1972, 1977), Schumm and Khan (1972) and Schumm et al. (1972),

rights reserved.

stream patterns are sensitive indicators of valley slope change. The re-duction of valley slope will lead to a reduction of sinuosity, the ratio ofchannel length to valley length or valley slope to channel slope.

This response is assumed to be detectable, if the fault is acompletely new fault, or if the fault reactivated recently, after a longperiod of inactivity (Holbrook and Schumm, 1999). If the tectonicshas been persistent for long time, but there are other, much strongerprocesses, the channel response may remain unclear (Holbrook andSchumm, 1999).

The hydrology of the river system (baseflow vs. peakflow,floodwave propagation through valley and floodplain) interactswith the associated groundwater system of the tectonic basin filland the flatter topography allows for wider floodplains. The geomet-ric and substrate differences are also controls on vegetation on riverbank and floodplain, which co-control fluvial style.

110 J. Petrovszki et al. / Global and Planetary Change 98–99 (2012) 109–121

Taking the valley-slope reducing effects as given (as implicitlydone by Schumm in most of his papers), it is useful to look at the sin-uosity variations along these systems and, where systems of the sameage are compared and observed differences are not explainable as thegrowth of discharge due to joining tributaries, and there one can in-terpret sinuosity change as neotectonic response-signal in the fluvialgeomorphology.

This work aims to analyze the geomorphological response of riversto active deformations within our study area which is the PannonianBasin where the aforementioned complexity of the phenomenon isalso observable. The basin is located in eastern Central Europe, boundedby the Alps, Carpathians and Dinarides, and it is unique in some geo-morphic and tectonic senses. The inner part of the basin is characterizedby a mosaic of hilly landforms (typically uplifting) and twomajor areasof prevailing subsidence that form conspicuously flat alluvial plains: theLittle Hungarian Plain and the Great Hungarian Plain (GHP). During thePliocene, tectonic reactivation interrupted the previously widespreadsubsidence and the filling up of the Pannonian basin (e.g., Bada et al.,2007). This recent activity hasmanifested itself in the uplift of thewest-ern and eastern flanks, and continuing subsidence of the central part ofthe basin creating a patchwork pattern of differential crustal uplift. Inthe Late Pliocene, extensional basin formation came to an end, and com-pressional inversion of the Pannonian basin is in progress since(Horváth and Cloething, 1996). In response to the filling up andrestructuring of sediment supply paths, a river system also formed inthat time (Nádor et al., 2003; Gábris and Nádor, 2007).

During the Quaternary, the subsidence rate increased in the deepestpart of the basin (Horváth and Cloething, 1996). Within the basin therate of subsidence changed often, so that the rivers often changedtheir channels (Franyó, 1992; Joó, 1992; Székely, 2009). In the modernsituation, differential subsidence is still on-going, but the subsidingareas with fastest subsidence rates nowadays coincide with areas ofconsiderable hydrocarbon and water extraction (Franyó, 1992; Joó,1992). The tectonically subsiding areas of the Pannonian Basin are char-acterized by the alluvial rivers with low or very low gradient settings;the prevailing process is aggradation. Alluvial terraces, if they exist atall, have very low relief as well. Their deformation by the tectonic forc-ing is considered to be below the detection limits.

Based on industrial seismic sections and earlier studies numerousfaults were identified and presented in the integrative neotectonicmap of Horváth et al. (2006). The noisy upper 50–100 mof the seismicsections, however, typically cannot be evaluated and inmost cases theneotectonic activity of these faults could not be assessed. In order to beable to classify them in the alluvial, low-relief area, we have carriedout the classic analysis of river channel pattern (more specificallythe sinuosity analysis) with some methodological extensions.

The classic experimental studies were performed by, amongothers, Ouchi (1985) to study the effects of substrate warping onchannels through alluvial basins (Fig. 1). Any deformation that

Fig. 1. The illustration of the river responses to the active vertical movements based on the1a: Meandering pattern with normal sinuosity, 1b: meandering pattern with high sinuosity, 1or braided pattern, 3: anastomosing pattern.

changes the slope of a river valley resulted in a corresponding coun-teractive change in sinuosity, back to equilibrium channel slope.Where the sinuosity changes downstream along a river the rates ofmeander migration and floodplain reworking accelerate accordingly.

Assuming this model being valid, the increasing slope is a very im-portant property in this consideration. The river has to stay in a givenchannel–slope range to do the self-organized meandering. Schummand Khan (1972) in their classic paper studying the phenomenon influme experiments found that the sinuosity increases only to a criticaldip (at a given discharge). If the slope increases above this threshold,the river sinuosity decreases. Considering the varying discharge,Leopold andWolman (1957) and Ackers and Charlton (1971) studiedhow the channel pattern changes with the changing slope andbankfull-discharge. Fig. 2 summarizes these threshold dependenciestogether with the calculated and measured values of the rivers stud-ied in this work. Most of the river reaches belong to the self-organizedmeandering domain.

The discharge (bankfull, average or mean-annual) determines to alarge extent channel dimensions (width, depth, meander amplitude,meander wavelength), but the quantity of water moving through achannel does not affect the basic pattern (Schumm, 1977). The type ofsediment load transported by the river also plays a role: channels thatcarry decent amounts offines (slit, clay) as suspended load ismore likelyto be sinuous (Schumm, 1977), whether they also carry significantamounts of bed load or not (Kleinhans and van den Berg, 2011).

Besides geomorphological response to along-river differences inhydrology, sediment load and active tectonics, alluvial channels havebeen reacting to human activities since prehistoric times, and directand lagging responses to climatic variations such as the Pleistocene–Holocene transition 15,000–10,000 years ago (e.g. Vandenberghe,1995) also play a role. Consequently, it may be difficult to separatethe causes of the river course modification, especially if reaches areconsidered individually only. Observing a downstream change in nat-ural channel pattern on its own is an insufficient indication to infer ac-tive tectonics. Nevertheless, if there is an independent indication forthe presence of fault zones e.g., based on geological or geophysical rea-sons, identification of anomalous reaches that were neither affectedby in regulation/engineering construction, nor have tributary influ-ences may reasonably be assumed to be the result of active tectonics(Schumm, 1986).

Nowadays, as the low plains of the Pannonian Basin are highlyflood-endangered, most of the rivers are regulated. The main rivercontrol works were carried out in the second half of the 19th century.During these works, the lengths of the rivers decreased, consequentlythe channel slope increased, and the shape of the rivers becamestraighter. The confinement of their course by the regulating earth-works influences or reshapes their original free meandering course,thus, if a fault is recently active, the rivers cannot change their sinuos-ity in the natural way to react to the valley slope changes.

flume experiments of Ouchi (1985) and the theories of Holbrook and Schumm (1999).c: meandering pattern with low sinuosity (even including straight pattern), 2:wandering

Fig. 2. a, b: a) The connection between the slope and sinuosity— according to the flume experiments of Schumm and Khan (1972). b) The channel pattern changes — depending onthe slope and bankfull discharge (Leopold and Wolman, 1957; Ackers and Charlton, 1971; Miall, 1977). The small dots refer to the slope and discharge values of the studied rivers.

111J. Petrovszki et al. / Global and Planetary Change 98–99 (2012) 109–121

The sinuosity-based method was successfully used worldwide in agreat number of cases. It was used by Adams (1980) on the Mississippiand Missouri rivers; Burnett and Schumm (1983) verified the uplift ofthe Monroe and Wiggins anticlines in Louisiana and Mississippi;Marple and Talwani (1993, 2000) used it on the rivers of the coastalplain of the Carolinas (USA); and later Aswathy et al. (2008) used thismethod on the Pannagon River, India. Concerning the Pannonianbasin, Timár (2003) demonstrated the recent activity of theMid-Hungarian Shear Zone (Csontos et al., 1992); Pišut (2006) studiedthe Lower Morava River (Slovakia); Kovács (2009, 2010) and Zámolyiet al. (2010) analyzed the river planforms of the Little HungarianPlain; and Petrovszki and Timár (2010) analyzed the Körös River Sys-tem, in Hungary.

In this study we intend to give an overview about the related sin-uosity studies in the region, providing indications of correlations be-tween sinuosity and tectonic pattern. Furthermore, we present asinuosity spectrum-based automatic segmentation method that ro-bustly separates river sections that appear to be significantly differentfrom the point of view sinuosity pattern. This method analyzes theriver courses very objectively, because it is free of the window size se-lection problem and its results are corresponding well with those ofthe classic sinuosity analysis.

2. Data and methods

2.1. The Second Military Survey of the Habsburg Empire

Similar to the aforementioned works studying the sinuosity of therivers in the Pannonian Basin, the watercourses were digitized fromthe map sheets of the Second Military Survey of the Habsburg Empire(Hofstätter, 1989; Jankó, 2001; Timár, 2004; Timár et al., 2006), be-cause they were surveyed before or during river control implementa-tion. So usually the original river beds were surveyed, but at placesthe planned cut-offs also were marked.

These maps were surveyed between 1806 and 1869. The maingeodetic reference points were defined by astronomical methodsand the work had a solid geodetic datum (Zach–Oriani hybrid ellip-soid, a=6,376,130 m, f=1/310). The scale of the maps is 1:28,800(Hofstätter, 1989; Varga, 2002). The survey was planned to be donebased on the Vienna projection origin. However, later the distortions

at the western and eastern frontiers of the empire turned to be ex-tremely high for areas far from the Vienna meridian. Consequently,in case of most of the provinces eventually local starting pointswere used. The estimated accuracy of the fitting of the map sheetsto the modern system is 50–100 m. Fortunately, for the territory oflarge parts of the Hungarian Kingdom, the survey was made withthe projection origin at the Stephansdom in Vienna; therefore the ac-curacy is better than 30 m (Hofstätter, 1989; Jankó, 2001; Timár,2004; Timár et al., 2006). The georeferencing of the survey sectionsused in our study was made by using the database of Timár et al.(2006).

2.2. Digitalization

The meandering, pre-regulated river beds were digitized fromthe georeferenced map sheets of the Second Military Survey of theHabsburg Empire, in the Hungarian National Grid (EOV) coordinatesystem. Polylines were digitized along the middle line of the chan-nels. The studied, digitized channel sections are enlisted in Table 1.These river sections are altogether 7500 km long. (Fig. 3) For thesake of simplicity for each considered section a capital letter notationhas been introduced, so the features revealed by the analysis can bereferenced easier. This notation is also included in Fig. 3.

The spacing of the vertices of these polylines varies according tothe curvature of the stream: the higher the curvature, the more verti-ces were digitized. To calculate the sinuosity it is advantageous tohave equal vertex spacing; after this conversion, the spacing was setto 50 m.

2.3. Sinuosity calculations

The sinuosity values of the river sections were calculated using thedefinition of Schumm (1963): it is the ratio of channel length to valleylength. The channel length was the window size, which was changedsuccessively from 5 km to 80 km, with a step of 5 km. The valleylength was the distance between the two endpoints of the window,along a straight line. Using this method, the sinuosity was calculatedfor all the point of the channels. These lengths were exported fromthe GIS and the calculations were made in Excel worksheets.

Table 1The results of the sinuosity calculations for each river section.

Northwestern rivers (Fig. 4) km Low Medium-low Medium Medium-high High Place of fault line Fault line

Vág C 163 C2, C5, C7 – C1, C3 –- C4, C6 C5|C6 1aVág-Duna B 124 B2, B5, B7 – B4, B6 B1, B3 –

Mosoni-Duna A 122 A5 A1, A4 A3 A2 –

Garam D 233 D3, D5, D10, D1, D9 D2, D4, D6, D8,D11, D13

D7, d14 D12 D7|D8; D10|D11;D13|D14

1b, 1a, 2

Ipoly Y 227 Y5, Y11, Y20 Y2, Y4, Y6,Y12, Y19

Y1, Y3, Y7, Y10,Y13, Y18

Y9, Y15, Y17 Y8, Y14, Y16 Y19|Y20 2

Danube (Fig. 5) km Low Medium-low Medium Medium-high High

Danube E 676 E1, E6, E8, E13 E5, E7, E9, E2, E4, E10, E12 – E3, E11 E1|E2|E3; E4|E5|E6;E7|E8|E9; E11|E12;

10, 12, 18, 20,19, 21, +

Baracskai-Duna F 76 F3, F7 F4 F2, F6 – F1, F5 F2|F3; 19

Croatia (Fig. 5) km Low Medium-low Medium Medium-high High

Sava T 308 T1, T7, T9, T11,T16, T18, T20, T22

T2, T4, T8, T10,T15, T21

T5, T12, T14, T17 T6 T3, T13, T19 T11|T12T13; T14|T15 22a, 22b

Drava S 311 S2, S4, S6, S8, S12 S1, S3, S7, S11 S5, S9, – S10 S2|S3(?);S5||||S6;

23

Upper-Theiss region (Fig. 6) km Low Medium-low Medium Medium-high High

Ondava G 140 G2, G4, G10 – G1, G3, G5, G7,G9, G11

G6 G8 G7|G8; 5

Laborc H 115 H1, H6 H2 H3, H5, H7 – H4 H3(?)|H4; 5Ung I 107 I2, I4 – I1, I3, I5 – I6 I4|I5; 5Latorca J 165 J1, J3 J2, J4, J7 J5 J6, J8 J3|J4; J1|J2; 5, 6Bodrog K 134 K2, K6 – K1, K3, K5 K4 –

Tisza L 1222 L1, L3, L6, L8, L10, L12 L2, L4, L7, L9, L11 – L5Sajó M 158 M1, M6 M2, M4, M7 M5 M3 – M6|M7 3–4 (3a, 3b)Hernád N 111 N4 N1, N3 N2 N5 –

Rivers inTransylvania (Fig. 7) km Low Medium-low Medium Medium-high High

Kraszna O 180 O7 O6, O8 O1, O3, O5 – O2, O4Kis-Szamos P 103 P1 P3 P2, P4, P6, P8, P10 P7, P9 P5 P4|P5; 8Szamos Q 491 Q1, Q4, Q6, Q9 Q2, Q7 Q3, Q8, Q10, Q12, Q14 Q13 Q5, Q11 Q7|Q8; 7Maros R 758 R1, R7, R15 R5, R9, R11,

R13, R17, R19R{even number} – R3, R21 R6|R7 9

Olt W 421 W1, W3, W5,W8, W10, W13

W2, W4, W12 W6, W11 W9 W7

Feketeügy X 100 X2, X4, X7 X6, X10 X1, X3, X8 X9 X4

Körös River System (Fig. 7) km Low Medium-low Medium Medium-high High

Berettyó V 270 V3 – V2, V4, V6 – V1, V5,V7 V5V7; 11Fehér-Körös U 173 U1, U3, U5 – U4, U11, U13 – U2, U12, U14 U13|U14; 13Fekete-Körös U 137 U8 – U6, U9, U11 – U7, U10 U6|U7; U10|U11; 16, 17Hármas-Körös U 236 – – U19, U23 – U20, U21, U22 U20U22; 12Sebes-Körös U 145 U15 – U19 – U16, U18 U15|U16; U17|U18; 14, 13Total 7406

112 J. Petrovszki et al. / Global and Planetary Change 98–99 (2012) 109–121

The ‘sinuosity spectrum’, a spectrum-like diagram (van Balenet al., 2008) is the best way to represent all the sinuosity values, com-puted for different window sizes. In the spectrum, the horizontal axisgives the distance along the river; the vertical axis indicates the win-dow size. To illustrate the sinuosity values, gray-scale images wereused (Fig. 4).

2.4. Classification

In order to improve the possible human interpretation error, we in-troduced a new automatedmethod to separate river sections character-ized by considerably different sinuosity properties (Fig. 5). Eachsinuosity calculation that was performed for a given window size, hasbeen considered as one band (one channel) of a multichannel“image”. Then, the sinuosity spectrums which became multichannelimages are of size 1×N where N represents the length of the actualriver in pixels (actually representing the originally resampled vertices).During the sinuosity analysis we introduced a classification step. An

automated segmentation of the rivers was performed, based on the sin-uosity spectrum: using this multichannel input unsupervised ISOCLASSclassification was carried out on these data, using ER Mapper software.The requested number of classes was set to 5.

Accordingly for each point of the river a resulting class wasassigned, except for the first and the last 25 km of the channel, be-cause the sinuosity spectrum data are, by definition, not complete inthese leading and trailing sections, the information is incomplete todo the classification.

This classification was done for each river spectrum individually.Depending on the rivers (size, discharge, slope, the lowest and thehighest sinuosity values), the sinuosity values can change in differentscales: Some cases from 1.5 to 3, but along the River Körös, Berettyóand Tisza we can find much higher values, up to 15–20. If we performthe classification for all rivers including all sections at once, the resultis different (Fig. 6), because the classes 4 and 5 will contain only thepoints with extreme high sinuosity values: they are subordinatelypresent in all input data.

Fig. 3. An overview of the analyzed river sections (black lines), dark gray lines represent the river sections that are not considered in this study (white lines: present-day stateborders).

113J. Petrovszki et al. / Global and Planetary Change 98–99 (2012) 109–121

In most cases, there is a clear and quite linear correlation betweenthe classes and the sinuosity values. Typical sinuosity values of theclasses were shown in Fig. 7 for some rivers with lower and highermaximum sinuosity values.

3. Results

The sinuosity spectra and section classifications were created forall studied river sections (Fig. 4). In the following we use the nameof the rivers used at the time of the surveying of the Second MilitarySurvey, indicating that the course of the river may considerably differfrom that of the present river course. However, at the first mentioningwe also give the present-day name equivalent for the respective riverfor reference.

In the analysis the individual features are referred to as a capitalletter-number combination where the capital letter refers to the ana-lyzed river (see Table 1 and Fig. 3) while the number is a sequentialnumber for the respective feature.

The sinuosity results were correlated to the maps of “Atlas of thepresent-day geodynamics of the Pannonian Basin” (Horváth et al.,2006) and with some seismic sections. Our results are in agreementwith both the faults of the neotectonic map and the faults of the seis-mic sections (Fig. 13; in this figure, not the whole spectrum displayedalong the rivers, but one window size, chosen by the bankfull-discharge). For the sake of brevity, instead of using the names of faultsestablished in the tectonic literature, a brief notation is used: ‘FaultN’

is used where N refers to the corresponding fault in Fig. 13. This in-tentional notation reduces the necessity of naming the faults thathave different sections named often differently in the literature, ormultiple names exist for overlapping sections. Our aim here is notthe discussion on existence or position of any faults: here we treatthem as input data as Horváth et al. (2006) considered them. The in-terested reader is kindly asked to verify their names, courses and tec-tonic style in Horváth et al. (2006), as well as in previous andsubsequent individual and integrative works, e.g., Bada et al. (2007),or Lenhardt et al. (2007).

3.1. Sinuosity patterns

In the following an evaluation is given for the individual faults thatcross the analyzed rivers. The evaluation focuses on whether there isa change observable in the sinuosity of the river in question, and ifyes, how far it matches the assumed crossing of the fault(s). It is im-portant to note that the surface projection of e.g., a low-angle normalfault has, by definition, a relatively large horizontal error. Further-more the faults are thought to be rather fault zones that are com-posed of various fault sections, that are typically bending andundulating; in addition, they are almost never perfectly continuous.

3.1.1. Northwestern part of the Pannonian Basin (Figs. 8, 13)Where Fault1 crosses the River Vág (present-day name: Váh) and

the River Garam (present-day: Hron), the sinuosity of the rivers is

Fig. 4. The sinuosity spectra for the studied river sections: the different colors mean different sinuosity values. The capital letter refers to the analyzed river while the number is a sequential number for the respective feature (see Table 1 andFig. 3).

114J.Petrovszkiet

al./Globaland

PlanetaryChange

98–99

(2012)109

–121

Fig. 6. a) The results of the classification for all spectra at once. b) Class 4 and class 5 contains only the points with extreme high sinuosity values. (cf. Fig. 7e).

Fig. 5. Visualization of the sinuosity spectrum. The sinuosity values displayed using a color scale/grayscale: the sinuosity values were imported like a single channel image. The sizeof the first image was N×10 pixels. (N — the length of the river in 50 m units; and 10 refers to the 10 different window size.) To enhance the visualization, the vertical size of thepixels was increased for 100 times.

115J. Petrovszki et al. / Global and Planetary Change 98–99 (2012) 109–121

increasing (C6 and D11). Furthermore, where Fault2 crosses the RiverGaram, the sinuosity is increasing again (D14). However, along theRiver Ipoly (present-day: Ipel’), the sinuosity is decreasing (Y20)after crossing Fault2. This may be due to various causes. The most ob-vious reason is the changing direction of the vertical movementsalong the fault. This explanation may have a counter-argument thatGaram and Ipoly are too close to each other to allow such a change

Fig. 7. Statistic of the classification: the mean sinuosity values of the classes with the 10 diffeclasses (7a — River Latorca, 7b — River Dráva, 7c — River Tisza); except the River Ung (7d), wthe classification was made for all rivers at once (7e), higher sinuosity values belonged to huosity values — less than 0.5% of the whole data.

in the tectonic regime. There is only 10 km distance between thesetwo fault crossings. Another reason could be that the channel slopeand bankfull-discharge is too high, so the river gets to the range ofunorganized meandering. The rivers of this group respond to thevertical movements just the opposite way, like the rivers in therange of self-organizing meandering. But here, both rivers belong tothe self-organizing meandering group. Likely, the opposite sinuosity

rent window-sizes — some typical examples. The more sinuous points belong to higherhere no clear correlation was detected between the sinuosity values and the classes. Ifigher classes, but class 4 and class 5 contained only the points with extreme high sin-

Fig. 8. The NWpart of the studied area (cf. Fig. 3): the sinuosity values are displayed along the river courses. Colors refer to the sinuositymeasured for a givenwindow size determined bythe bankfull-discharge. The different colorsmean different sinuosity values (see Fig. 4). The capital letter refers to the analyzed riverswhile the numbers are the sequential numbers for therespective features (see Table 1 and Fig. 3).

116 J. Petrovszki et al. / Global and Planetary Change 98–99 (2012) 109–121

changing was caused by the bedrock-alluvial change. The Ipoly, alongthe Slovakian–Hungarian border is not an alluvial river but bedrockcontrolled. It incises into the rocks of the Börzsöny Hills. After leavingFault2, the Ipoly becomes an alluvial river. Fault2 is a lithological mar-gin. Fault3 is crossing the River Garam, and the sinuosity valueschange from medium (D6) to medium-high (D7), and then medium(D8) again. The neotectonic map shows, it is an uncertain fault, butsinuosity values indicate its plausibility.

Fig. 9. The north-eastern part of the studied area (cf. Fig. 3): the sinuosity values are displarefer to the analyzed rivers while the numbers are the sequential numbers for the respecti

3.1.2. Northeastern part of the Pannonian Basin (Figs. 9, 13)To evaluate the crossing of Fault4a and Fault4b a seismic profile

(MK-21) was also studied that runs parallel to the River Sajó to anoffset of about 10 km (Fig. 10). There are two remarkable faultlines on this seismic section that are also indicated in theneotectonic map. From the seismic section, it can be derived thatthe northern block is uplifted with respect to the southern block.The river had low sinuosity upstream of the faults (M6), and

yed along the river courses. For explanation see the caption of Fig. 8. The capital lettersve features (see Table 1 and Fig. 3).

Fig. 10. The seismic section (MK-21) is parallel to a part of the River Sajó (indicated in green in the bottom left inset). The fault on the seismic section is the same as the fault on the neotectonicmap. Sinuosity reacts to the change in slope. The river has low sinuosity upstream of the faults, and downstream, at the southern side of the faults, the sinuosity increases. The change of thesinuosity appears right as the river crosses the fault; the coincidence is a clear indication that it is caused by the fault, therefore the fault is still likely to be active. In the seismic section it canbe seen that the northern block is uplifted with respect to the southern block. This increase happens upstream to the confluence with the River Hernád, so not the change is independent ofwater or sediment discharge caused by the confluence.

117J. Petrovszki et al. / Global and Planetary Change 98–99 (2012) 109–121

downstream, at the southern side of the faults, the sinuosityincreased (M7). The sinuosity changed similar to that of the flumeexperiments of Ouchi (1985). If the Sajó did not flow into theTisza (Theiss) then the sinuosity might increase more. The changeof the sinuosity appears right as the river crosses the faults, whichis a clear indication that it is related to the faults. Therefore thefaults are still likely to be active.

Fault5 crosses all the tributaries of the River Bodrog, and after thesinuosity values were increased (J4, I5, H4, G8). Where the fault linemeets the River Laborc (today: Laborec), on the neotectonic map, itis between sections H4 and H5. But according to the rivers, the placesof the sinuosity changes, and the values of the sinuosity, this structuralline does not coincide with that place. It should cross the River Laborcbetween sections H3 and H4.

The north-eastern end of the Fault6 meets with the River Latorca(today: Latorica). Here the increase of sinuosity is not so significantlike at the other faults, but it is present (J2). This may imply thateither this fault line has not been active recently, or the verticalmovement component is less expressed here than at the other places.

3.1.3. Middle and southeastern part of the Pannonian Basin (Figs. 11, 13)Fault7 crosses the River Szamos (today the Romanian section is

Someş) between sections Q7 and Q8, increasing the sinuosity values.The medium low values of the section Q7 increase to medium insection Q8.

There are only a few faults and also a few suitable rivers to studyin this region. Fault8 crosses the River Kis-Szamos (today: SomeşulMic). Upstream to the fault, the river had medium sinuosity (P4)and below the fault the sinuosity increased, and became high (P5).

Fault9 crosses the River Maros (today the Romanian section isMureş), and the medium sinuosity values (R6) decrease to low values(R7). So the direction of the vertical movements, comparing to theriver flowing direction is just the opposite of that at Fault8.

Fault10 crosses the River Danube (Fig. 12b), and increases thesinuosity of the river (E2). This fault also crosses the River Tiszaincreasing the sinuosity of this river too. It is important to note thatthese two rivers are the largest ones in the region.

Fault12 is almost parallel to Fault10. This fault crosses the RiverDanube too and also theHármas-Körös. Along theDanube, the sinuosity

Fig. 11. The eastern and middle part of the studied area (cf. Fig. 3): the sinuosity values are displayed along the river courses. For explanation see the caption of Fig. 8. The capitalletters refer to the analyzed rivers while the numbers are the sequential numbers for the respective features (see Table 1 and Fig. 3).

118 J. Petrovszki et al. / Global and Planetary Change 98–99 (2012) 109–121

decreases (E3: high, E4:medium). Along theHármas-Körös, themediumsinuosity values become high after the river crosses the fault (U20, U22).This is consistent with the phenomena along the River Danube. At theSW part of the fault, the sinuosity values are medium, and at the NWpart of the fault, the values are high.

In the Körös Region, Fault11, Fault13, Fault14, Fault16 and Fault17are also increasing the sinuosity of the rivers. V5, V7, U14, U16, U18,U10 and U7 are the high sinuosity sections of these rivers, indicatingthe neotectonic activity of the structural lines.

The U14 and U18 sections with high sinuosity values are not exactlyalong the Fault13. But according to some seismic section and the signif-icant changes of the sinuosity, the neotectonic map could also be some-what modified to become congruent with this observations.

Fig. 12. The southern part of the studied area (cf. Fig. 3): the sinuosity values are displayed ato the analyzed rivers while the numbers are the sequential numbers for the respective fea

Where Fault15 crosses the River Sebes-Körös (today in RomaniaCrişul Repede), the sinuosity does not change significantly, but thefluvial pattern does: Anastomosed river pattern appears instead ofthe meandering (Petrovszki and Timár, 2010).

3.1.4. Southwestern part of the Pannonian Basin (Figs. 12, 13)Fault18, Fault19, and Fault20 also cross the River Danube, with a co-

inciding decrease in sinuosities. Fault19 crosses the Baracskai-Duna,too, and decreases its sinuosity, too. (Fault18: E4—medium, E5—low;Fault19: F2—medium, F3—low and E7—medium-low, E8—low;Fault20: F1—high, F2—medium). After the River Danube crossedFault21, its sinuosity increases a bit (E8: low, E9: medium-low).

long the river courses. For explanation see the caption of Fig. 8. The capital letters refertures (see Table 1 and Fig. 3).

Fig. 13. The faults interpreted to be active (black lines) according to the sinuosity changes of the rivers revealed in this study. The colors displayed along the rivers refer to onewindow size, chosen by the bankfull-discharge.

119J. Petrovszki et al. / Global and Planetary Change 98–99 (2012) 109–121

Fault25 is a strike–slip fault. However, it also coincides with thechanged sinuosity of the Danube. A great bend appears by the fault,with high sinuosity values (E11), but this can be attributed to theshifting of the blocks, and not to the varying slope.

Where Fault22a crosses the River Sava, the sinuosity of the riverincreased (T12, T13). According to the neotectonic map, this fault isa normal fault, and the northern block was uplifted with respect tothe southern block. The river proves this by increasing the sinuosity.Fault22b made just the opposite, the sinuosity decreased (T15).

Fault23 is rather a fault system also as it appears in the map. At theplace where the River Dráva crosses these faults its sinuosity is medium(S5). Downstreamof this fault system the sinuosity of theDráva is decreas-ing (S6). Fault24 seems to also increase the sinuosity of the Drava (S3).

Fault26 is a normal fault, and it increases the sinuosity of the RiverSava from low (T18) to high (T19).

3.2. Classification

The classification was made for all digitized river sections(Fig. 14). Because of the sinuosity classification method that shortensthe river course considered in the classification, in some cases, thedisplayed section of the river does not reach to the respective faultrepresented on the neotectonic map. For example in the case of theRiver Vág (C) the interpreted section does not cross Fault1, and thesame applies to the River Garam (D) and to the River Ipoly in caseof Fault2, etc. In the other cases, where the faults are crossing the riv-ers, the results are corresponding with the results of the sinuosityspectrum (Table 2, Fig. 14).

The points of the River Garam upstream of Fault1 belong to theclass 2, and that downstream of the fault belongs to the class 3.Along the River Danube, the neotectonic activity of Fault10, Fault12,Fault18, Fault20 and Fault21 was also confirmed by the classification.In the Upper-Theiss (Tisa, Tisza) region, the rivers are crossing Fault5and the points of the rivers are also changing their class. Except forthe River Ung (today: Uzh in Ukraine and Uh in Slovakia), this riveris terminated before crossing the fault. The same phenomenon canbe seen at the following crossings: along the River Kis-Szamos withFault8, along the River Maros with Fault9, along the River Drau(Drava/Dráva) and Sava, and in the Körös River System (Fig. 14.)

4. Discussion

In general we conclude that according to our observations in mostof the cases the sinuosity of the rivers changes along the known struc-tural lines as they crossed these fault lines. The downstream sinuositychanges along the rivers of the Pannonian Basin are significant, andthe changes at the indicated locations (see details above) can be relat-ed to the differences in vertical movements across the known faults.These faults are very young and neotectonically active during the Ho-locene, so analyzing the industrial seismic sections their neotectonicactivity was not evident. At some places (e.g. Fault5 and Fault13)this sinuosity change is somewhat offset to the structural lines indi-cated in the neotectonic map (see the horizontal error in Section 3.1).

The fault lines of the neotectonicmap and the places of the significantsinuosity changes are correlated the most in the southern part of theGreat Hungarian Plain, in the area of the Danube and the Körös River

Fig. 14. The faults (black lines) interpreted to be active — according to the classification of the sinuosity changes of the rivers.

120 J. Petrovszki et al. / Global and Planetary Change 98–99 (2012) 109–121

System. Note that this area is characterized by very low relief. Here, lotsof fault lines are indicated on the map, and the rivers verified theneotectonic activity of most of them. Of course those fault lines, whichdo not cross the analyzed rivers, cannot be evaluated.

The most successful identification in this area is the tectonic line(Fault12), of which neotectonic activity was verified by 4 different riv-ers. The Danube, the Hármas-Körös, the Kettős-Körös and the Berettyóare also crossing this fault and downstream to the crossing, their sinuos-ity changed considerably. However, the style of the change (decrease orincrease) cannot be determined, because the River Danube comes fromnorth, and downstream the fault, its sinuosity decreased, while theother rivers get to the fault from the opposite direction, and their sinu-osity increase. What we can state is that in the southern part of Fault12the rivers flowwith lower sinuosity than in the northern side with oneexception: the River Hármas-Körös. The reason is that Fault12 is astrike–slip fault; its main displacement component is horizontal.

Table 2The results of the classification for those rivers which cross the faults represented on the n

River The classes (“|” indicates the place of a fault l

Garam (D) 1 2 3 4 3 2 3 4 3 2 3 4 3 4 3 4 5 4 3 2 3 1 2 | 3Ondava (G) 3 1 2 5 4 5 4 5 4 2 5 | 2 4 2 5 4 2 4 5Laborc (H) 1 2 3 4 3 5 4 | 2Latorca (J) 1 | 2 1 2 | 3 4 5 4 5 4 5 4 5 4 5Kis-Szamos (P) 2 1 2 1 2 4 5 4 5 4 2 4 3 4 | 5 2 1Maros (R) 1 2 1 2 | 1 2 1 2 1 4 1 2 3 2 3 2 1 2 4 2Hármas-Körös (U) 1 | 2 3 2 1 2 3 5 3 2 1 | 2 3 4 3 | 2 1 2 1Sebes-Körös (U) 1 2 3 5 | 4 5 4 3 2Dráva (S) 2 | 3 2 3 4 | 3 2 3 2 | 1 2 3 2 3 4 5 4Száva (T) 1 2 3 2 1 2 3 2 3 2 1 2 1 2 1 | 2 3 4 3 4 3 4 3 2Duna (E) 1 2 | 3 4 5 4 | 3 4 | 3 | 2 3 2 | 3 4 5 4 | 3 2 1 3

Along the strike–slip faults, flower structures can be seen in the seismicsections. The different blocks in the flower structures move to differentdirections, up or down, or even theymay remain neutral. This effect ap-pears in this section.

There is one more structural line, of which the neotectonic activitywas verified by many rivers. Fault5 crossing all the four digitized trib-utaries of the River Bodrog (the Ondava, the Laborc, the Ung and theLatorca), and their sinuosities increase. So, along these two fault lines,the river sinuosity changes are indicated the neotectonic activity notonly in a certain point, but along hundreds of kilometers. The otherstructural lines are shorter, crossing only one or sometimes twofaults, or the rivers they crossing are not meandering (e.g. the upperpart of the Danube).

There are some places, where the changes of the sinuosity are signifi-cant, but according to the neotectonic map, no fault lines are there. Atthese places, seismic sections should be analyzed to decide whether is

eotectonic map.

ine) Fault line

4 3 4 5 4 5 4 3 4 5 4 3 4 5 4 5 4 3 4 3 1556, 58912, 12, 121324, 23, 25

1 2 3 2 1 | 2 3 2 3 5 3 2 1 | 2 22a, 26, 261 2 3 2 1 2 3 2 1 2 1 2 3 4 3 2 1 10, 12, 18, 20, 21, 25

121J. Petrovszki et al. / Global and Planetary Change 98–99 (2012) 109–121

there any fault line or something else could cause these changes. Alterna-tively, the sinuosity changes at these locations could be of different origin,for example due to changes in the sediment load or the sediment type inprehistoric to historic times. At some places the rivers are not alluvial, butbedrock controlled, so their sinuosity cannot reflect self-organizationof the meanders in equilibrium with imposed slopes, as in alluvialmaterial. The applicability of the methods of this paper is provided only,if the channel slope and the bankfull-discharge stay in a given range(Fig. 2), otherwise the river course becomes straight, unorganizedmeandering, wandering or braided and in those cases sinuosity no longerreveals tectonic control.

5. Conclusions

The connection between the river sinuosity changes and theneotectonic activity was studied on the rivers of the PannonianBasin. The rivers are responding to the tectonic forcing, if the move-ments have vertical component.

In the Pannonian Basin, 28 rivers were studied. Along 26 faultlines, at 36 points, the places of the significant sinuosity changesand the faults on the neotectonic map correlate. Using this methodthe neotectonic activity of 26 faults was found to be correlated. Theactivity of two faults can be verified by more than one river. Atsome places, according to the rivers, the faults are not running atthe same place on the neotectonic map where the sinuosity changeoccurs; there is some offset to be observed.

We introduced a new evaluationmethod, the classification of mul-tiple window-size based sinuosity spectrum. If the river is longenough for the analysis, the classification can be at least as useful, asthe sinuosity spectrum, but sometimes it is more straightforward.Furthermore, for the classification, there is no need for the main hy-drological parameters of the river, e.g. the bankfull discharge.

Analyzing the results of the classification, it is important to note thatthemethod typically splits the river course into contiguous sections thatbelong to the same class. Boundaries of these classes can be considered asshort transition zones of considerable change in the river course, becausethe method uses statistically relevant amount of data of the river coursein a robust way to detect changes. Some specific classes or their bound-aries seem to be correlated to tectonically active zones.

Acknowledgments

The authors are grateful for the constructive comments of the re-viewers that helped to improve the manuscript. This research hasbeen carried out in the frame of project OTKA-NK83400 (SourceSinkHungary). The European Union and the European Social Fund alsohave provided financial support to the project under the grant agree-ment no. TÁMOP 4.2.1./B-09/1/KMR-2010-0003. The presented seismicsection was provided by the Hungarian Geological Service.

References

Ackers, P., Charlton, F.G., 1971. The slope and resistance of small meandering channels.Inst. Civil Engineers Proc. Supp. XV, Paper 73625.

Adams, J., 1980. Active tilting of the United States midcontinent: geodetic and geomorphicevidence. Geology 8, 442–446.

Aswathy, M.V., Vijith, H., Satheesh, R., 2008. Factors influencing the sinuosity ofPannagon River, Kottayam, Kerala, India: an assessment using remote sensingand GIS. Environmental Monitoring and Assessment 138 (1–3), 173–180.

Bada, G., Horváth, F., Dövényi, P., Szafián, P., Windhoffer, G., Cloetingh, S., 2007. Present-day stress field and tectonic inversion in the Pannonian basin. Global and PlanetaryChange 58, 165–180.

Burnett, A.W., Schumm, S.A., 1983. Alluvial river-response to neotectonic deformationin Louisiana and Mississippi. Science 222, 49–50.

Csontos, L., Nagymarosi, A., Horváth, F., Kovač, M., 1992. Tertiary evolution of the Intra-Carpathian area: a model. In: Ziegler, P.A. (Ed.), Geodynamics of Rifting, Vol. 1. CaseHistory Studies on Rifts: Europe and Asia: Tectonophysics, 208, pp. 221–241.

Franyó F. (1992):Magyarországnegyedidőszaki üledékeinek vastagságtérképe (Thicknessmap of Quaternary sediments in Hungary). Map, scale 1:500.000, Magyar ÁllamiFöldtani Intézet, Budapest.

Gábris, Gy, Nádor, A., 2007. Long-term fluvial archives in Hungary: response of the Danubeand Tisza rivers to tectonic movements and climatic changes during the Quaternary: areview and new synthesis. Quaternary Science Reviews 26, 2758–2782.

Hofstätter, E., 1989. Beiträge zur Geschichte der österreichischen Landesaufnahmen, I.Teil, Bundesamt für Eich- und Vermessungswesen, Wien. 196 pp.

Holbrook, J., Schumm, S.A., 1999. Geomorphic and sedimentary response of rivers to tectonicdeformation: a brief review and critique of a tool for recognizing subtle epeirogenicdeformation in modern and ancient settings. Tectonophysics 305, 287–306.

Horváth, F., Cloetingh, S., P. L., A., 1996. Stress-induced late-stage subsidence anomaliesin the Pannonian basin. In: Cloetingh, S.A.P.L., Ben Avraham, Z., Sassi, W., Horváth,F. (Eds.), Dynamics of Extensional Basins and Inversion Tectonics: Tectonophysics,266, pp. 287–300.

Horváth, F., Bada, G., Windhoffer, G., Csontos, L., Dombrádi, E., Dövényi, P., Fodor, L.,Grenerczy, Gy, Síkhegyi, F., Szafián, P., Székely, B., Timár, G., Tóth, L., Tóth, T.,2006. A Pannon-medence jelenkori geodinamikájának atlasza: Euro-konformtérképsorozat és magyarázó. Magyar Geofizika 47 (4), 133–137.

Jankó, A., 2001. A második katonai felmérés. Hadtörténeti Közlemények 114, 103–129.Joó, I., 1992. Recent vertical surface movements in the Carpathian Basin. Tectonophysics

202, 129–134.Kleinhans, M.G., van den Berg, J.H., 2011. River channel and bar patterns explained and

predicted by an empirical and a physics-based method. Earth Surface Processesand Landforms 36, 721–738, http://dx.doi.org/10.1002/esp. 2090.

Kovács, G., 2009. Structural lineaments derived from planforms of watercourses in theSecond Military Survey of the Habsburg Empire. In: Gartner, G., Ortrag, F. (Eds.),Cartography in Central and Eastern Europe. Springer, Berlin Heidelberg. ISBN:978-3-642-03293-6, pp. 525–533, http://dx.doi.org/10.1007/978-3-642-03294-3(Print) 978-3-642-03294-3 (Online).

Kovács, G., 2010. The advantages of using the second military survey maps in fluvialstudies. Acta Geodaetica et Geophysica Hungarica 45 (1), 64–70.

Lenhardt, W.A., Švancara, J., Melichar, P., Pazdírková, J., Havíř, J., Sýkorová, Z., 2007.Seismic activity of the Alpine–Carpathian–Bohemian Massif region with regard togeological and potential field data. Geologica Carpathica 58 (4), 397–412.

Leopold, L.B., Wolman, M.G., 1957. River channel patterns; braided, meandering andstraight. USGS Prof. Paper 282B, pp. 1–73.

Marple, R.T., Talwani, P., 1993. Evidence of possible tectonic upwarping along the SouthCarolina coastal plain from an examination of river morphology and elevation data.Geology 21, 651–654.

Marple, R.T., Talwani, P., 2000. Evidence for a buried fault system in the Coastal Plain ofthe Carolinas and Virginia — implications for neotectonics in the southeasternUnited States. Geological Society of America Bulletin 112, 200–220.

Miall, A.D., 1977. A review of the braided river depositional environment. Earth-ScienceReviews 13, 1–62.

Nádor, A., Lantos, M., Tóth-Makk, Á., Thamó-Bozsó, E., 2003. Milankovitch-scalemultiproxy records from fluvial sediments of the last 2.6 Ma, Pannonian Basin,Hungary. Quaternary Science Reviews 22, 2157–2175.

Ouchi, S., 1985. Response of alluvial rivers to slow active tectonic movement. GeologicalSociety of America Bulletin 96, 504–515.

Petrovszki, J., Timár, G., 2010. Channel sinuosity of the Körös River system, Hungary/Romania,as possible indicator of the neotectonic activity. Geomorphology 122, 223–230.

Pišut, P., 2006. Evolution of meandering Lower Morava River (West Slovakia) duringthe first half of 20th century. Geomorphologica Slovaca 6 (1), 55–68.

Schumm, S.A., 1963. A tentative classification of alluvial river channels. U.S. GeologicalSurvey Circular V477.

Schumm, S.A. (Ed.), 1972. River Morphology. Dowden, Hutchinson & Ross, Stroudsburg,PA. 444 pp.

Schumm, S.A., 1977. The Fluvial System. John Wiley & Sons, New York. 338 pp.Schumm, S.A., 1986. Alluvial river response to active tectonics. Active Tectonics. Studies in

Geophysics. National Academy Press,Washington, D. C., pp. 80–94. ISBN 0-309-03638.Schumm, S.A., Khan, H.R., 1972. Experimental study of channel patterns. Geological

Society of America Bulletin 83, 1755–1770.Schumm, S.A., Khan, H.R., Winkley, B.R., Robbins, L.G., 1972. Variability of river patterns.

Nature 237, 75–76.Székely, B., 2009. Rediscovering the old treasures of cartography—What an almost 500-year-

old map can tell to a geoscientist. Acta Geodetica et Geophysica Hungarica 44 (1), 3–16.Timár, G., 2003. Controls on channel sinuosity changes: a case study of the Tisza River,

the Great Hungarian Plain. Quaternary Science Reviews 22, 2199–2207.Timár, G., 2004. GIS integration of the second military survey sections — a solution

valid on the territory of Slovakia and Hungary. Kartografické Listy 12, 119–126.Timár, G., Molnár, G., Székely, B., Biszak, S., Varga, J., Jankó, A., 2006. Digitized maps of

the Habsburg Empire — the map sheets of the second military survey and theirgeoreferenced version. Arcanum, Budapest. 59 pp.

van Balen, R.T., Kasse, C., Moor, J., 2008. Impact of groundwater flow on meandering:example from the Geul river, the Netherlands. Earth Surface Processes and Land-forms 33 (13), 2010–2028.

Vandenberghe, J., 1995. Timescales, climate and river development. Quaternary ScienceReviews 14, 631–638.

Varga, J., 2002. A vetületnélküli rendszerektől az UTM-ig. (From projectionless systemsto the UTM. On-line manuscript in Hungarian.)URL (accessed at 03.10.2011.):http://www.agt.bme.hu/staff_h/varga/Osszes/Dok3uj.htm.

Zámolyi, A., Székely, B., Draganits, E., Timár, G., 2010. Neotectonic control on river sinuosityat the western margin of the Little Hungarian Plain. Geomorphology 122, 231–243.