Morphological Evolution of the Lower Tisza River (Hungary) in the 20th Century in Response to Human Interventions

: To understand the hydro-morphological responses of a river to engineering works is a key to future sustainable engineering projects. The Tisza River in the Carpathian Basin has been the subject of human interventions in the 19th–20th centuries, with regular monitoring since the ﬁrst intervention. Therefore, it provides a unique opportunity to understand the long-term responses of a lowland river on engineering works. Here the interventions include the construction of embanked levees, artiﬁcial meander cutoffs, construction of revetments and groynes, and construction of lock-dams, whose differing effects can be analysed in detail. The Lower Tisza River, Hungary was shortened by up to a third of its 89-km length during late-19th-century regulation works. Simultaneously, it was disconnected from its 10–20 km wide natural ﬂoodplain, which was restricted to 1–5 km. Revetments were constructed along 25% of the banks since the early 20th century. This paper evaluates the effects of these regulation works on the Lower Tisza channel based on channel-surveys (1891, 1931, 1961, 1976 and 1999). Based on the results, the channel responded to every stage of the regulation works by adjusting its cross-sectional area, thalweg depth and bankfull width. From 1891 to 1961, the cross-sectional area of the channel increased in response to the artiﬁcial meander cutoffs through rapid incision and channel widening. The construction of revetments and groynes distorted this morphological development. After their construction, the cross-sectional area of the channel reduced by up to 50% between 1961 and 1976. The channel experienced incision and narrowing within this period. Whilst the morphological response of the river to cutoffs was improved ﬂood conveyance, the revetments drastically decreased the ﬂood conveyance; therefore, the ﬂood levels increased without an increase in discharge, so the revetment constructions indirectly increased the ﬂood risk.


Introduction
Rivers as natural systems of varying forms and sizes have been responsible for shaping the Earth's landscapes [1,2]. Rivers in their natural states develop towards equilibrium. However, external forces such as climate change [3][4][5][6][7][8] or human intervention [1,5,[9][10][11][12][13] tend to create conditions of disequilibrium. The impact of climate change on rivers and river systems has been widely investigated in recent decades [8,[14][15][16][17][18][19]. Likewise, several studies have been conducted on the various human impacts on rivers [20][21][22][23][24][25][26][27][28]. The distortion of the fluvial forms in relation to the abovementioned parameters could be connected to the interplay of bed topography, flow field and sediment transport, and anthropogenic changes in channel morphology [11,[29][30][31][32]. On the Tisza, two major floods usually occur. The first flood (March-April) is caused by snow melt, and if the snowmelt is combined with rainfall extremely large floods could develop. The second flood occurs during the summer (June), induced by the early summer rainfall. As the lowland section of the river is long, and its slope is low (1.0-3.5 cm/km), this summer flood sometimes superimposes on the preceding spring flood, and thus record high floods could develop. The absolute water level change between the highest and lowest water stages is 1355 cm at Mindszent (highest stage measured in 2006: 1062 cm; lowest stage measured in 1968: −293 cm), and 1184 cm at Szeged (highest stage measured in 2006: 960 cm; lowest stage measured in 1968: −224 cm; Figure 2). The discharge also shows considerable hydrological variations. At Szeged, the maximum discharge is 4348 m 3 /s (in 1932) and the minimum value is 58 m 3 /s (in 2013; [62,64]). The yearly maximum discharge values and the corresponding water stages at Szeged (Figure 3) not clearly related, and they indicate special channel processes. In case of stages lower than the bankfull level, the discharge values increased, as at the beginning of the measurements (1920-1959) 550-600 cm stages had ca. 1750 m 3 /s discharge, but later the discharge values of the same stage interval increased, referring to incision. However, in the 21st century this process seems to be abated. On the contrary, the discharge of the overbank flood levels dropped, indicating a decrease in the flood conveyance: at the beginning of the measurements, record high floods (800-900 cm) produced a large discharge (over 3500 m 3 /s), but nowadays, even though the floods have only 2500-3700 m 3 /s discharge, the recorded flood levels increased (over 1000 cm). On the Tisza, two major floods usually occur. The first flood (March-April) is caused by snow melt, and if the snowmelt is combined with rainfall extremely large floods could develop. The second flood occurs during the summer (June), induced by the early summer rainfall. As the lowland section of the river is long, and its slope is low (1.0-3.5 cm/km), this summer flood sometimes superimposes on the preceding spring flood, and thus record high floods could develop. The absolute water level change between the highest and lowest water stages is 1355 cm at Mindszent (highest stage measured in 2006: 1062 cm; lowest stage measured in 1968: −293 cm), and 1184 cm at Szeged (highest stage measured in 2006: 960 cm; lowest stage measured in 1968: −224 cm; Figure 2). The discharge also shows considerable hydrological variations. At Szeged, the maximum discharge is 4348 m 3 /s (in 1932) and the minimum value is 58 m 3 /s (in 2013; [62,64]). The yearly maximum discharge values and the corresponding water stages at Szeged (Figure 3) not clearly related, and they indicate special channel processes. In case of stages lower than the bankfull level, the discharge values increased, as at the beginning of the measurements (1920-1959) 550-600 cm stages had ca. 1750 m 3 /s discharge, but later the discharge values of the same stage interval increased, referring to incision. However, in the 21st century this process seems to be abated. On the contrary, the discharge of the overbank flood levels dropped, indicating a decrease in the flood conveyance: at the beginning of the measurements, record high floods (800-900 cm) produced a large discharge (over 3500 m 3 /s), but nowadays, even though the floods have only 2500-3700 m 3 /s discharge, the recorded flood levels increased (over 1000 cm).

Regulation Works
The late 19th century marked the beginning of catchment-scale engineering interventions on the Tisza River. They involved the disconnection of the Tisza from its alluvial wetlands by construction of embanked levees. This flood control levee system is 2940 km long and was constructed with the aim of protecting the land from inundation, providing safe areas for agriculture, and decreasing the duration of floods [67,68]. Simultaneously, the river was shortened by 38% with 114 artificial meander cutoffs [69,70]. As a result of cutoffs, the meandering equilibrium river was transformed into an ingrowing meandering one, as its slope almost doubled. With an increased sediment load, there was accelerated deposition on the confined active floodplains [46,71].
The regulation works in the 20th century continued with the construction of revetments and groynes. Revetments were built to stop the lateral erosion of bends that migrated too close to artificial levees, whilst the groynes facilitated shipping by tightening wide sections and training sharp bends in order to improve the flood conductivity of the river channel. In total 44% of the banks of the Tisza River were stabilized, mainly between the 1930s and 1960s. However, these works started in 1886, whilst the latest revetment was built in 2016.
The second half of the 20th century was a period of dam construction on the Tisza and its tributaries. On the Tisza, three dams (which operate as locks) were built at Tiszalök (1957) and Kisköre (1973) in Hungary, and Törökbecse/Novi Becsej (1976) in Serbia [72]. The locks were built to impound the water during low stages, thus water withdrawal for irrigation could be secured, but they also produce hydropower and aid navigation. Although they are opened during floods, the retention of sediments by the dams creates a temporal sediment deficit downstream that induces incision [23,62].

Regulation Works
The late 19th century marked the beginning of catchment-scale engineering interventions on the Tisza River. They involved the disconnection of the Tisza from its alluvial wetlands by construction of embanked levees. This flood control levee system is 2940 km long and was constructed with the aim of protecting the land from inundation, providing safe areas for agriculture, and decreasing the duration of floods [67,68]. Simultaneously, the river was shortened by 38% with 114 artificial meander cutoffs [69,70]. As a result of cutoffs, the meandering equilibrium river was transformed into an ingrowing meandering one, as its slope almost doubled. With an increased sediment load, there was accelerated deposition on the confined active floodplains [46,71].
The regulation works in the 20th century continued with the construction of revetments and groynes. Revetments were built to stop the lateral erosion of bends that migrated too close to artificial levees, whilst the groynes facilitated shipping by tightening wide sections and training sharp bends in order to improve the flood conductivity of the river channel. In total 44% of the banks of the Tisza River were stabilized, mainly between the 1930s and 1960s. However, these works started in 1886, whilst the latest revetment was built in 2016.
The second half of the 20th century was a period of dam construction on the Tisza and its tributaries. On the Tisza, three dams (which operate as locks) were built at Tiszalök (1957) and Kisköre (1973) in Hungary, and Törökbecse/Novi Becsej (1976) in Serbia [72]. The locks were built to impound the water during low stages, thus water withdrawal for irrigation could be secured, but they also produce hydropower and aid navigation. Although they are opened during floods, the retention of sediments by the dams creates a temporal sediment deficit downstream that induces incision [23,62].

Regulation Works
The late 19th century marked the beginning of catchment-scale engineering interventions on the Tisza River. They involved the disconnection of the Tisza from its alluvial wetlands by construction of embanked levees. This flood control levee system is 2940 km long and was constructed with the aim of protecting the land from inundation, providing safe areas for agriculture, and decreasing the duration of floods [67,68]. Simultaneously, the river was shortened by 38% with 114 artificial meander cutoffs [69,70]. As a result of cutoffs, the meandering equilibrium river was transformed into an in-growing meandering one, as its slope almost doubled. With an increased sediment load, there was accelerated deposition on the confined active floodplains [46,71].
The regulation works in the 20th century continued with the construction of revetments and groynes. Revetments were built to stop the lateral erosion of bends that migrated too close to artificial levees, whilst the groynes facilitated shipping by tightening wide sections and training sharp bends in order to improve the flood conductivity of the river channel. In total 44% of the banks of the Tisza River were stabilized, mainly between the 1930s and 1960s. However, these works started in 1886, whilst the latest revetment was built in 2016.
The second half of the 20th century was a period of dam construction on the Tisza and its tributaries. On the Tisza, three dams (which operate as locks) were built at Tiszalök (1957) and Kisköre (1973) in Hungary, and Törökbecse/Novi Becsej (1976) in Serbia [72]. The locks were built to impound the water during low stages, thus water withdrawal for irrigation could be secured, but they also produce hydropower and aid navigation. Although they are opened during floods, the retention of sediments by the dams creates a temporal sediment deficit downstream that induces incision [23,62]. Along the studied Lower Tisza, the channel regulations started in 1855, shortening the originally 131 km long studied river segment by more than 40 km, cutting off 12 meanders [23,28]. The originally 6-8 km wide natural floodplain was reduced to 1 km on average. At Szeged, however, the floodplain is just 400 m wide, thus the artificial levees constricted the floodplain at some sections quite drastically, thereby reducing its flood conductivity and increasing the flood risk. Therefore, the height of the embanked levees increased severalfold, in keeping with the increasing height of floods. Thus, originally they were just 2.3-3.0 m high, but nowadays their height is 7.0 m, so they could provide safety despite the increasing flood levels [23,61,62,70]. In the Lower Tisza, 25% of the banks are protected by revetments, but their spatial distribution is irregular, as indicated by the different revetment density of the studied reaches (Table 1).

Materials and Methods
A dataset of the hydrological surveys of cross-sectional area profiles as well as the spatial data of cutoffs and revetments of the Lower Tisza River were provided by the Lower Tisza Water Directorate (ATIVIZIG). The Hydrological Atlas of the Lower Tisza (1974) provides maps on the channel showing the banklines (1891, 1931 and 1976), the locations of the surveyed cross sections and revetments. Channel surveys were made in 1891,1931,1961,1976 and 1999, using fixed geodetic points (known as VO points) established in 1890-91. The cross sections located at approximate 2.5 km intervals. During the surveys of cross section elevation (for the near-bank area), depth measurements (for the wetted channel) were made at 5.0 m intervals. Based on the measurements, the relative depths measured from the bankfull level were used in analysing the characteristics of the cross sections.
For a typical cross section in this present study (Figure 4), the bankfull level was determined from the lower bankline. From the relative depths obtained, the thalweg depth (deepest point measured from bankfull level) and mean depth (arithmetic mean of all depth data of a cross section) were determined. Besides these, we calculated the cross-sectional area (area of cross section up to the bankfull level), the bankfull width (width of cross section at the bankfull level) and the mean width (ratio of cross-sectional area to thalweg depth). Along the studied Lower Tisza, the channel regulations started in 1855, shortening the originally 131 km long studied river segment by more than 40 km, cutting off 12 meanders [23,28]. The originally 6-8 km wide natural floodplain was reduced to 1 km on average. At Szeged, however, the floodplain is just 400 m wide, thus the artificial levees constricted the floodplain at some sections quite drastically, thereby reducing its flood conductivity and increasing the flood risk. Therefore, the height of the embanked levees increased severalfold, in keeping with the increasing height of floods. Thus, originally they were just 2.3-3.0 m high, but nowadays their height is 7.0 m, so they could provide safety despite the increasing flood levels [23,61,62,70]. In the Lower Tisza, 25% of the banks are protected by revetments, but their spatial distribution is irregular, as indicated by the different revetment density of the studied reaches (Table 1).

Materials and Methods
A dataset of the hydrological surveys of cross-sectional area profiles as well as the spatial data of cutoffs and revetments of the Lower Tisza River were provided by the Lower Tisza Water Directorate (ATIVIZIG). The Hydrological Atlas of the Lower Tisza (1974) provides maps on the channel showing the banklines (1891, 1931 and 1976), the locations of the surveyed cross sections and revetments. Channel surveys were made in 1891,1931,1961,1976 and 1999, using fixed geodetic points (known as VO points) established in 1890-91. The cross sections located at approximate 2.5 km intervals. During the surveys of cross section elevation (for the near-bank area), depth measurements (for the wetted channel) were made at 5.0 m intervals. Based on the measurements, the relative depths measured from the bankfull level were used in analysing the characteristics of the cross sections.
For a typical cross section in this present study (Figure 4), the bankfull level was determined from the lower bankline. From the relative depths obtained, the thalweg depth (deepest point measured from bankfull level) and mean depth (arithmetic mean of all depth data of a cross section) were determined. Besides these, we calculated the cross-sectional area (area of cross section up to the bankfull level), the bankfull width (width of cross section at the bankfull level) and the mean width (ratio of cross-sectional area to thalweg depth).   The entire length of the river was dissected into bends or meanders based on Laczay's classification [73], for which the sinuosity of every segment was calculated between two inflection points (midpoint of straight sections) as the ratio of the bend length of the segment to the chord length (length of straight line joining inflection points). All segments with sinuosity values below 1.1 were classified as bends while those with sinuosity values greater than 1.1 were classified as meanders. The classification was applied because the thalweg is deeper and closer to the outer banks in case of meanders, thus the meanders and their cross sections can change more than the bends [1]. Then the cross sections were classified based on the sinuosity of the channel (location in meanders or bends) and type of human intervention directly affecting the cross-sectional profile ( Figure 5) Water 2018, 10, x FOR PEER REVIEW 6 of 19 The entire length of the river was dissected into bends or meanders based on Laczay's classification [73], for which the sinuosity of every segment was calculated between two inflection points (midpoint of straight sections) as the ratio of the bend length of the segment to the chord length (length of straight line joining inflection points). All segments with sinuosity values below 1.1 were classified as bends while those with sinuosity values greater than 1.1 were classified as meanders. The classification was applied because the thalweg is deeper and closer to the outer banks in case of meanders, thus the meanders and their cross sections can change more than the bends [1]. Then the cross sections were classified based on the sinuosity of the channel (location in meanders or bends) and type of human intervention directly affecting the cross-sectional profile ( Figure 5)

Changes in Vertical Channel Parameters
To analyse the vertical channel changes, first the cross-sectional area values of the whole studied river segment was evaluated, and then the upper, middle and lower reaches were analysed in detail ( Figure 6).

Changes in Vertical Channel Parameters
To analyse the vertical channel changes, first the cross-sectional area values of the whole studied river segment was evaluated, and then the upper, middle and lower reaches were analysed in detail ( Figure 6). The entire length of the river was dissected into bends or meanders based on Laczay's classification [73], for which the sinuosity of every segment was calculated between two inflection points (midpoint of straight sections) as the ratio of the bend length of the segment to the chord length (length of straight line joining inflection points). All segments with sinuosity values below 1.1 were classified as bends while those with sinuosity values greater than 1.1 were classified as meanders. The classification was applied because the thalweg is deeper and closer to the outer banks in case of meanders, thus the meanders and their cross sections can change more than the bends [1]. Then the cross sections were classified based on the sinuosity of the channel (location in meanders or bends) and type of human intervention directly affecting the cross-sectional profile ( Figure 5)

Changes in Vertical Channel Parameters
To analyse the vertical channel changes, first the cross-sectional area values of the whole studied river segment was evaluated, and then the upper, middle and lower reaches were analysed in detail ( Figure 6).  The mean cross-sectional area of the entire studied segment was 1530 m 2 in 1891, and then increased until 1961 (1672 m 2 ), but later it decreased until 1976 (1557 m 2 ), and finally it increased again between 1976 and 1999 (1586 m 2 ). Similar changes occurred within the upper reach and the lower reach, although the magnitudes of change varied. In the middle reach however, the trend of the changes was different: the mean cross-sectional channel area also started to increase after 1891 but it started to decline earlier , and finally in the last period  The spatial (downstream) changes in the cross-sectional area were also studied in detail ( Figure 7). Some cross-sections only slightly changed between 1891 and 1999, as the change in cross-sectional area was less than 0.5% at certain locations (VO217 and VO234), whilst others had considerable changes, e.g., in case of VO221 the changes were as high as 177%. The greatest variation in the cross-sectional area occurred within the middle reach. A general downstream trend is also visible, as in the upper reach in 1891 the cross-sectional areas were less than 1500 m 2 , whilst in the same period the lower reach always had higher values than 1500 m 2 . However, abrupt changes in cross-sectional area by neighbouring survey sites appeared in each survey period, referring to local influencing factors. The mean cross-sectional area of the entire studied segment was 1530 m 2 in 1891, and then increased until 1961 (1672 m 2 ), but later it decreased until 1976 (1557 m 2 ), and finally it increased again between 1976 and 1999 (1586 m 2 ). Similar changes occurred within the upper reach and the lower reach, although the magnitudes of change varied. In the middle reach however, the trend of the changes was different: the mean cross-sectional channel area also started to increase after 1891 but it started to decline earlier , and finally in the last period  The spatial (downstream) changes in the cross-sectional area were also studied in detail ( Figure  7). Some cross-sections only slightly changed between 1891 and 1999, as the change in cross-sectional area was less than 0.5% at certain locations (VO217 and VO234), whilst others had considerable changes, e.g., in case of VO221 the changes were as high as 177%. The greatest variation in the crosssectional area occurred within the middle reach. A general downstream trend is also visible, as in the upper reach in 1891 the cross-sectional areas were less than 1500 m 2 , whilst in the same period the lower reach always had higher values than 1500 m 2 . However, abrupt changes in cross-sectional area by neighbouring survey sites appeared in each survey period, referring to local influencing factors. The changes in cross-sectional area could be influenced by such vertical channel parameters as bankfull width and mean thalweg depth. Along the studied segment of the Tisza River, the bankfull width continuously decreased throughout all the survey years ( Figure 8). While in 1891 the mean bankfull width of the channel was 203 m, by 1999 it had reduced by 39 m, so the mean channel narrowing was 36 cm/y. Between the surveys, the bankfull width reduction was similar (0.12-0.16 m/y), though between 1961-1976 it accelerated to 1.17 m/y. At the downstream end of the middle reach (VO221) the channel became wider by 125%. The greatest narrowing characterized the upper reach (46% at VO202; and 42% at VO203) and the upstream parts of the middle reach (46% at VO214; and 43% at VO 215).
The thalweg depth also showed variation along the studied segment ( Figure 9). During the studied 108 years, the mean thalweg depth of the entire studied segment increased from 12.2 m to 14.1 m, indicating 1.9 m incision on average (1.7 cm/y). In this process, similarly to the width changes, the period of 1961-1976 was characterised by a slight aggradation of channel bed, as the mean thalweg depth decreased from 13.8 m to 13.5 m. The minimum and maximum rates of incision occurred in 1961-1976 (0.6 cm/y) and 1976-1999 (2.4 cm/y) respectively. Considering downstream changes, no obvious trend was revealed, as the average change in the mean thalweg depth was 18.7% in the upper reach, the greatest increase (19.5%) was measured in the middle reach, whilst only a 16.8% increase in thalweg depth was found in the lower reach over the study period. The greatest incision occurred at the boundaries of the reaches (63% for VO207; 52% for VO 244; and 48% for The changes in cross-sectional area could be influenced by such vertical channel parameters as bankfull width and mean thalweg depth. Along the studied segment of the Tisza River, the bankfull width continuously decreased throughout all the survey years ( Figure 8). While in 1891 the mean bankfull width of the channel was 203 m, by 1999 it had reduced by 39 m, so the mean channel narrowing was 36 cm/y. Between the surveys, the bankfull width reduction was similar (0.12-0.16 m/y), though between 1961-1976 it accelerated to 1.17 m/y. At the downstream end of the middle reach (VO221) the channel became wider by 125%. The greatest narrowing characterized the upper reach (46% at VO202; and 42% at VO203) and the upstream parts of the middle reach (46% at VO214; and 43% at VO 215).
The thalweg depth also showed variation along the studied segment ( Figure 9). During the studied 108 years, the mean thalweg depth of the entire studied segment increased from 12. upper reach, the greatest increase (19.5%) was measured in the middle reach, whilst only a 16.8% increase in thalweg depth was found in the lower reach over the study period. The greatest incision occurred at the boundaries of the reaches (63% for VO207; 52% for VO 244; and 48% for VO234). The highest change in depth (122%, by 12.5 m) for both the entire study period and between surveys occurred from 1891 to 1931 in the upper reach (VO200).

Effect of Sinuosity on Vertical Channel Parameters
In the 19th century, when the channel-survey network was established, the cross sections were set an approximately uniform distance (2.5 km) from each other. Therefore, they represent channel sections with various sinuosity, and various locations even within a meander. Therefore, to understand the role of the planform of the river in determining cross-sectional area, the cross sections were grouped based on their location: in slightly sinuous bends or in sinuous meanders ( Figure 10). In the slightly sinuous bends usually the cross-sectional area was 2-9% greater than in meanders, except in the first survey period (1891), when bends had smaller area by 6.5%. At the first half of the studied period (1891-1961) the cross-sectional area of the bends increased gradually by 16.6% (from 1491 m 2 to1739 m 2 ), while of the meanders, this increase was just 1.1% (from 1589 m 2 to 1606 m 2 ). Later, between 1961 and 1976 the cross-sectional area experienced considerable reduction within both the bends (8.1%) and in the meanders (2.7%). However, during the last period (1976-1999) the crosssectional area of bends only slightly changed (−0.7%), but the cross-sectional area of the meanders considerably increased by 6.9%, (from 1462 m 2 to 1563 m 2 ). At the end of these changes the crosssectional area conditions of the entire channel became similar, as the values of bends (1599 m 2 ) and meanders (1563 m 2 ) were initially not so close to each other.
These changes are well reflected by the yearly change in cross-sectional area. The greatest increase (5.4 m 2 /y) in cross-sectional area of bends took place between 1891 and 1931, and later the process slowed down (1931-1961: 1.1 m 2 /y) to one fifth. Later, the channel of the bends even became smaller, as its area decreased by 8.6 m 2 /y (1961-1976). The process almost terminated between 1976

Effect of Sinuosity on Vertical Channel Parameters
In the 19th century, when the channel-survey network was established, the cross sections were set an approximately uniform distance (2.5 km) from each other. Therefore, they represent channel sections with various sinuosity, and various locations even within a meander. Therefore, to understand the role of the planform of the river in determining cross-sectional area, the cross sections were grouped based on their location: in slightly sinuous bends or in sinuous meanders ( Figure 10). In the slightly sinuous bends usually the cross-sectional area was 2-9% greater than in meanders, except in the first survey period (1891), when bends had smaller area by 6.5%. At the first half of the studied period (1891-1961) the cross-sectional area of the bends increased gradually by 16.6% (from 1491 m 2 to1739 m 2 ), while of the meanders, this increase was just 1.1% (from 1589 m 2 to 1606 m 2 ). Later, between 1961 and 1976 the cross-sectional area experienced considerable reduction within both the bends (8.1%) and in the meanders (2.7%). However, during the last period (1976-1999) the crosssectional area of bends only slightly changed (−0.7%), but the cross-sectional area of the meanders considerably increased by 6.9%, (from 1462 m 2 to 1563 m 2 ). At the end of these changes the crosssectional area conditions of the entire channel became similar, as the values of bends (1599 m 2 ) and meanders (1563 m 2 ) were initially not so close to each other.
These changes are well reflected by the yearly change in cross-sectional area. The greatest increase (5.4 m 2 /y) in cross-sectional area of bends took place between 1891 and 1931, and later the process slowed down (1931-1961: 1.1 m 2 /y) to one fifth. Later, the channel of the bends even became

Effect of Sinuosity on Vertical Channel Parameters
In the 19th century, when the channel-survey network was established, the cross sections were set an approximately uniform distance (2.5 km) from each other. Therefore, they represent channel sections with various sinuosity, and various locations even within a meander. Therefore, to understand the role of the planform of the river in determining cross-sectional area, the cross sections were grouped based on their location: in slightly sinuous bends or in sinuous meanders ( Figure 10). In the slightly sinuous bends usually the cross-sectional area was 2-9% greater than in meanders, except in the first survey period (1891), when bends had smaller area by 6.5%. At the first half of the studied period (1891-1961) the cross-sectional area of the bends increased gradually by 16.6% (from 1491 m 2 to1739 m 2 ), while of the meanders, this increase was just 1.1% (from 1589 m 2 to 1606 m 2 ). Later, between 1961 and 1976 the cross-sectional area experienced considerable reduction within both the bends (8.1%) and in the meanders (2.7%). However, during the last period (1976-1999) the cross-sectional area of bends only slightly changed (−0.7%), but the cross-sectional area of the meanders considerably increased by 6.9%, (from 1462 m 2 to 1563 m 2 ). At the end of these changes the cross-sectional area conditions of the entire channel became similar, as the values of bends (1599 m 2 ) and meanders (1563 m 2 ) were initially not so close to each other. and 1999, as the cross-sectional area decrease was only 0.5 m 2 /y. The cross-sectional area in meanders also increased at the beginning of the studied period, but by much smaller rate (1891-1931: 0.6 m 2 /y).
In the meanders the deterioration of the channel area started as early as 1931 and lasted until 1976 (by 0.2 m 2 /y and 8.4 m 2 /y respectively), and recently the area increased by 4.4 m 2 /y (1976-1999). The patterns of change of the cross-sectional area for the entire studied segment and the three reaches ( Figure 11) indicated that the changes in the bends followed a similar pattern. That notwithstanding, there were differences due to the magnitudes of change. In the meanders however, the changes for each of the reaches as well as the river segment followed different patterns. To better understand the spatial characteristics of the changes above, the cross-sectional area parametric values (Figure 12) of the three reaches of the Lower Tisza River within both bends and meanders were compared. From 1891, the mean cross-sectional area of the upper reach was the least both in bends and meanders (1407 m 2 and 1291 m 2 , respectively) with the maximum for both bends and meanders in the lower reach (1676 m 2 and 1672 m 2 , respectively). Over the entire studied period, the greatest change in the mean cross-sectional area in bends occurred in the lower reach (10.9%) while in that for meanders occurred in the upper reach (18.4%). The least change in the bends over the period occurred in the upper reach (−2.7%) with the least change in meanders occurring in the middle reach (−3.7%). These changes were further highlighted by the bankfull width and the thalweg depth. Although the cross-sectional area increased over the 108-year studied period by 4.9% and 1.9% within the bends and meanders, the bankfull width decreased by 18.5% and 20.5%, respectively. Over the same period, however, the thalweg depth increased by 17.9% and 13.0% in the bends and meanders, respectively. Similar changes in bankfull width and thalweg depth took place within the reaches. In general, the greatest differences between the reaches appeared between 1961 and 1976, when as it indicated, the development pattern of bends and meanders altered. However, the difference between the three reaches became smaller by the end of the studied period, referring to more uniform channel conditions. The patterns of change of the cross-sectional area for the entire studied segment and the three reaches ( Figure 11) indicated that the changes in the bends followed a similar pattern. That notwithstanding, there were differences due to the magnitudes of change. In the meanders however, the changes for each of the reaches as well as the river segment followed different patterns. To better understand the spatial characteristics of the changes above, the cross-sectional area parametric values ( Figure 12) of the three reaches of the Lower Tisza River within both bends and meanders were compared. From 1891, the mean cross-sectional area of the upper reach was the least both in bends and meanders (1407 m 2 and 1291 m 2 , respectively) with the maximum for both bends and meanders in the lower reach (1676 m 2 and 1672 m 2 , respectively). Over the entire studied period, the greatest change in the mean cross-sectional area in bends occurred in the lower reach (10.9%) while in that for meanders occurred in the upper reach (18.4%). The least change in the bends over the period occurred in the upper reach (−2.7%) with the least change in meanders occurring in the middle reach (−3.7%). These changes were further highlighted by the bankfull width and the thalweg depth. Although the cross-sectional area increased over the 108-year studied period by 4.9% and 1.9% within the bends and meanders, the bankfull width decreased by 18.5% and 20.5%, respectively. Over the same period, however, the thalweg depth increased by 17.9% and 13.0% in the bends and meanders, respectively. Similar changes in bankfull width and thalweg depth took place within the reaches. In general, the greatest differences between the reaches appeared between 1961 and 1976, when as it indicated, the development pattern of bends and meanders altered. However, the difference between the three reaches became smaller by the end of the studied period, referring to more uniform channel conditions. and meanders, the bankfull width decreased by 18.5% and 20.5%, respectively. Over the same period, however, the thalweg depth increased by 17.9% and 13.0% in the bends and meanders, respectively. Similar changes in bankfull width and thalweg depth took place within the reaches. In general, the greatest differences between the reaches appeared between 1961 and 1976, when as it indicated, the development pattern of bends and meanders altered. However, the difference between the three reaches became smaller by the end of the studied period, referring to more uniform channel conditions.

Effect of Artificial Cutoffs on Vertical Channel Parameters
The downstream variations of cross-sectional parameters only partly could be explained by the various sinuosity conditions at the surveyed profiles, as the direct engineering works could also affect the vertical parameters of the channel. The artificial cutoffs affected the channel development directly at the 12 surveyed cross sections (see Figure 5).
The mean cross-sectional area increased by 39.4% (1891-1999) of the straightened channel sections created by artificial cutoffs. Meanwhile, the mean cross-sectional area of those sections, which were not affected directly by cutoffs slightly decreased (2.7%). The cross-sectional area of the straightened sections increased considerably soon after the cutoffs were made (1891-1931: 13.2 m 2 /y), and even the non-straightened sections had an increased cross-sectional area (0.8 m 2 /y; Figure 13). The process continued within the straightened sections later on , though its rate decreased (3.1 m 2 /y). However, the non-straightened sections had already experienced decline in their crosssectional area (0.3 m 2 /y). The deterioration of the straightened sections started after 1961, as their cross-sectional area decreased by 6.3 m 2 /y (1961-1979) and 6.5 m 2 /y . Within the nonstraightened sections, the yearly change in cross-sectional area had a different pattern, as it decreased by 8.7 m 2 /y between 1961 and 1976, but later it increased by 4.5 m 2 /y (1976-1999).
The changes within these straightened sections also indicate that the maximum and mean crosssectional area values increased just until 1961 ( Figure 14). However, the minimum cross-sectional area value (620 m 2

Effect of Artificial Cutoffs on Vertical Channel Parameters
The downstream variations of cross-sectional parameters only partly could be explained by the various sinuosity conditions at the surveyed profiles, as the direct engineering works could also affect the vertical parameters of the channel. The artificial cutoffs affected the channel development directly at the 12 surveyed cross sections (see Figure 5).
The mean cross-sectional area increased by 39.4% (1891-1999) of the straightened channel sections created by artificial cutoffs. Meanwhile, the mean cross-sectional area of those sections, which were not affected directly by cutoffs slightly decreased (2.7%). The cross-sectional area of the straightened sections increased considerably soon after the cutoffs were made (1891-1931: 13.2 m 2 /y), and even the non-straightened sections had an increased cross-sectional area (0.8 m 2 /y; Figure 13). The process continued within the straightened sections later on , though its rate decreased (3.1 m 2 /y). However, the non-straightened sections had already experienced decline in their cross-sectional area (0.3 m 2 /y). The deterioration of the straightened sections started after 1961, as their cross-sectional area decreased by 6.3 m 2 /y (1961-1979) and 6.5 m 2 /y . Within the non-straightened sections, the yearly change in cross-sectional area had a different pattern, as it decreased by 8.7 m 2 /y between 1961 and 1976, but later it increased by 4.5 m 2 /y . The changes within these straightened sections also indicate that the maximum and mean crosssectional area values increased just until 1961 ( Figure 14). However, the minimum cross-sectional area value (620 m 2 : VO221) was very low in 1891, and it increased by 166% until 1931 (1651 m 2 ). Though later on  it also decreased similarly to the maximum and mean values, though the decrease in minimum values started earlier. A comparison of the mean cross-sectional area values of the straightened sections within the reaches reflects that, there was always an increasing downstream trend, thus the lower reach had the greatest cross-sectional area in all survey years. In the straightened sections the temporal patterns of the changes in the cross-sectional area within the reaches were similar to those of entire Lower Tisza River, while those of the nonstraightened sections were different (Figure 15). In the straightened sections, the yearly change in the cross-sectional area for all the reaches decreased until 1976, after which it increased, although not up to the earlier rates. On the other hand, the non-straightened sections had varying trends in the yearly change in the cross-sectional area, as while the upper and middle reaches increased until 1931, the values of the upper reach decreased. Later there was a decreasing trend for the upper and lower reaches while the cross-sectional area of the middle reach increased. However, after 1976, the crosssectional areas of the straightened and non-straightened sections uniformly increased.
The yearly change in the bankfull width and thalweg depth highlighted their influences on the cross-sectional area ( Figure 16). Generally, there was an inverse relationship between the yearly change in bankfull width and yearly change in thalweg depth. While the rate bankfull width change decreased until 1976, later it increased. On the other hand, the yearly change in thalweg depth generally decreased until 1976, but later decreased. In spite of these general trends, between 1931 and 1961, there was an increase in the rate of bankfull width in the upper reach for the straightened sections, while an increase for the same period within the non-straightened sections of the lower reach was also detected. The changes within these straightened sections also indicate that the maximum and mean cross-sectional area values increased just until 1961 ( Figure 14). However, the minimum cross-sectional area value (620 m 2 : VO221) was very low in 1891, and it increased by 166% until 1931 (1651 m 2 ). Though later on  it also decreased similarly to the maximum and mean values, though the decrease in minimum values started earlier. A comparison of the mean cross-sectional area values of the straightened sections within the reaches reflects that, there was always an increasing downstream trend, thus the lower reach had the greatest cross-sectional area in all survey years. The changes within these straightened sections also indicate that the maximum and mean crosssectional area values increased just until 1961 ( Figure 14). However, the minimum cross-sectional area value (620 m 2 : VO221) was very low in 1891, and it increased by 166% until 1931 (1651 m 2 ). Though later on  it also decreased similarly to the maximum and mean values, though the decrease in minimum values started earlier. A comparison of the mean cross-sectional area values of the straightened sections within the reaches reflects that, there was always an increasing downstream trend, thus the lower reach had the greatest cross-sectional area in all survey years. In the straightened sections the temporal patterns of the changes in the cross-sectional area within the reaches were similar to those of entire Lower Tisza River, while those of the nonstraightened sections were different ( Figure 15). In the straightened sections, the yearly change in the cross-sectional area for all the reaches decreased until 1976, after which it increased, although not up to the earlier rates. On the other hand, the non-straightened sections had varying trends in the yearly change in the cross-sectional area, as while the upper and middle reaches increased until 1931, the values of the upper reach decreased. Later there was a decreasing trend for the upper and lower reaches while the cross-sectional area of the middle reach increased. However, after 1976, the crosssectional areas of the straightened and non-straightened sections uniformly increased.
The yearly change in the bankfull width and thalweg depth highlighted their influences on the cross-sectional area ( Figure 16). Generally, there was an inverse relationship between the yearly change in bankfull width and yearly change in thalweg depth. While the rate bankfull width change decreased until 1976, later it increased. On the other hand, the yearly change in thalweg depth generally decreased until 1976, but later decreased. In spite of these general trends, between 1931 and 1961, there was an increase in the rate of bankfull width in the upper reach for the straightened sections, while an increase for the same period within the non-straightened sections of the lower reach was also detected. The changes within these straightened sections also indicate that the maximum and mean cross-sectional area values increased just until 1961 ( Figure 14). However, the minimum cross-sectional area value (620 m 2 : VO221) was very low in 1891, and it increased by 166% until 1931 (1651 m 2 ). Though later on  it also decreased similarly to the maximum and mean values, though the decrease in minimum values started earlier. A comparison of the mean cross-sectional area values of the straightened sections within the reaches reflects that, there was always an increasing downstream trend, thus the lower reach had the greatest cross-sectional area in all survey years.
In the straightened sections the temporal patterns of the changes in the cross-sectional area within the reaches were similar to those of entire Lower Tisza River, while those of the non-straightened sections were different (Figure 15). In the straightened sections, the yearly change in the cross-sectional area for all the reaches decreased until 1976, after which it increased, although not up to the earlier rates. On the other hand, the non-straightened sections had varying trends in the yearly change in the cross-sectional area, as while the upper and middle reaches increased until 1931, the values of the upper reach decreased. Later there was a decreasing trend for the upper and lower reaches while the cross-sectional area of the middle reach increased. However, after 1976, the cross-sectional areas of the straightened and non-straightened sections uniformly increased.

Effects of the Construction of Revetments and Groynes on Vertical Channel Parameters
To evaluate the effect of revetments on the morphological evolution of the channel, the changes in cross-sectional area of non-revetmented and revetmented sections were compared ( Figure 17). The revetments and groynes were mainly constructed after the 1931 survey, thus their effect on vertical parameters should be pronounced only later on. Besides, these direct works were local (almost pointlike), affecting low number of cross sections (see Figure 5). As there are only three channel cross sections with groynes, and in the same locations there are revetments as well, in the following analysis we consider only the influence of revetments. The yearly change in the bankfull width and thalweg depth highlighted their influences on the cross-sectional area ( Figure 16). Generally, there was an inverse relationship between the yearly change in bankfull width and yearly change in thalweg depth. While the rate bankfull width change decreased until 1976, later it increased. On the other hand, the yearly change in thalweg depth generally decreased until 1976, but later decreased. In spite of these general trends, between 1931 and 1961, there was an increase in the rate of bankfull width in the upper reach for the straightened sections, while an increase for the same period within the non-straightened sections of the lower reach was also detected.

Effects of the Construction of Revetments and Groynes on Vertical Channel Parameters
To evaluate the effect of revetments on the morphological evolution of the channel, the changes in cross-sectional area of non-revetmented and revetmented sections were compared ( Figure 17). The revetments and groynes were mainly constructed after the 1931 survey, thus their effect on vertical

Effects of the Construction of Revetments and Groynes on Vertical Channel Parameters
To evaluate the effect of revetments on the morphological evolution of the channel, the changes in cross-sectional area of non-revetmented and revetmented sections were compared ( Figure 17). The revetments and groynes were mainly constructed after the 1931 survey, thus their effect on vertical parameters should be pronounced only later on. Besides, these direct works were local (almost point-like), affecting low number of cross sections (see Figure 5). As there are only three channel cross sections with groynes, and in the same locations there are revetments as well, in the following analysis we consider only the influence of revetments. The mean cross-sectional areas in non-revetmented and revetmented sections from 1891 (1484.7 m 2 and 1591.2 m 2 , respectively) changed by 7.1% and −0.5%, respectively over the 108-year period. The trends of change between surveyed years were similar, although the magnitudes of change were very different. The area of cross sections increased until 1961, decreased by 1976, and again increased by 1999. Over a period of 50 years (up to 1961), there was an increase in the mean cross-sectional area of 13.8% and 6.5% for the non-revetmented sections and the revetmented sections respectively. Between 1961 and 1976, however, the non-revetmented sections decreased by 6.6% while the revetmented sections decreased by 10.5%.
Within the reaches, the mean cross-sectional area generally followed the same pattern as the whole studied segment within the non-revetmented sections. While the rate of change in crosssectional area ( Figure 18) generally decreased until 1976 from 1891, the lower reach had an increase between 1931 and 1961. In the revetmented sections, the reaches responded differently to the construction of the revetments. The upper and middle reaches had reducing rates of change of the mean cross-sectional area until 1961, while the lower reach had an increasing rate. By 1976, the rates of change for the upper and lower reaches were very high, although all the reaches experienced a reduction in cross-sectional area. By 1999, the rates of change were all positive, resulting in increases in cross-sectional area.   (1891, 1931, 1961, 1976 and 1999); and (b) the changes in the rate of yearly mean cross-sectional area of non-revetmented and revetmented sections.
The mean cross-sectional areas in non-revetmented and revetmented sections from 1891 (1484.7 m 2 and 1591.2 m 2 , respectively) changed by 7.1% and −0.5%, respectively over the 108-year period. The trends of change between surveyed years were similar, although the magnitudes of change were very different. The area of cross sections increased until 1961, decreased by 1976, and again increased by 1999. Over a period of 50 years (up to 1961), there was an increase in the mean cross-sectional area of 13.8% and 6.5% for the non-revetmented sections and the revetmented sections respectively. Between 1961 and 1976, however, the non-revetmented sections decreased by 6.6% while the revetmented sections decreased by 10.5%.
Within the reaches, the mean cross-sectional area generally followed the same pattern as the whole studied segment within the non-revetmented sections. While the rate of change in cross-sectional area ( Figure 18) generally decreased until 1976 from 1891, the lower reach had an increase between 1931 and 1961. In the revetmented sections, the reaches responded differently to the construction of the revetments. The upper and middle reaches had reducing rates of change of the mean cross-sectional area until 1961, while the lower reach had an increasing rate. By 1976, the rates of change for the upper and lower reaches were very high, although all the reaches experienced a reduction in cross-sectional area. By 1999, the rates of change were all positive, resulting in increases in cross-sectional area.
The changes in cross-sectional area were further highlighted by the changes that occurred in the bankfull width and thalweg depth within the reaches, and the whole studied segment ( Figure 19). Generally, the changes in the bankfull width in the non-revetmented segments showed similar patterns of change for all reaches, and the range of the yearly rates was under 50 cm. Between 1961 and 1976 however, while the rate of channel narrowing within the upper and lower reaches were similar (decreases of 296 cm/y and 293 cm/y respectively), the width of the middle reach decreased only by 89 cm/y. In the revetmented sections, the reaches responded differently, and the range of channel width variations was higher than 300 cm/y. In general, channel narrowing took place, except for upper reach up to 1961, and the middle reach after 1976. between 1931 and 1961. In the revetmented sections, the reaches responded differently to the construction of the revetments. The upper and middle reaches had reducing rates of change of the mean cross-sectional area until 1961, while the lower reach had an increasing rate. By 1976, the rates of change for the upper and lower reaches were very high, although all the reaches experienced a reduction in cross-sectional area. By 1999, the rates of change were all positive, resulting in increases in cross-sectional area. The changes in cross-sectional area were further highlighted by the changes that occurred in the bankfull width and thalweg depth within the reaches, and the whole studied segment ( Figure 19). Generally, the changes in the bankfull width in the non-revetmented segments showed similar patterns of change for all reaches, and the range of the yearly rates was under 50 cm. Between 1961 and 1976 however, while the rate of channel narrowing within the upper and lower reaches were similar (decreases of 296 cm/y and 293 cm/y respectively), the width of the middle reach decreased only by 89 cm/y. In the revetmented sections, the reaches responded differently, and the range of channel width variations was higher than 300 cm/y. In general, channel narrowing took place, except for upper reach up to 1961, and the middle reach after 1976. The rate of change in the thalweg depth was similar within the reaches with a range of less than 5 cm/y in the non-revetmented sections for all periods. In the revetmented sections, although the magnitudes of change were different, the middle and lower reaches had a similar range as the nonrevetmented sections (5 cm/y). In the upper reach, however, although the pattern of change was similar, the changes were more than 35 cm/y.

Discussion
The changes in vertical channel morphology of the Lower Tisza River were driven by the sinuosity of the channel and the type and timing of engineering interventions. The repeated surveys of the channel enabled us to do a time-series analyses of the cross sections. It indicates a general incision of the whole channel, which could be related to the increased stream power as a result of the shortened channel length and increased gradient by the artificial cutoffs [74]. The incision due to similar human impacts has been observed in other rivers in the world too [20,50,60,75].
The channel developed generally according to the channel evolution model [76]. The cutoffs that began by the mid-19th century were responsible for most of the changes that occurred in the channel up to 1961, although they were more pronounced directly after the cutoffs, thus between 1891 and 1931. This meant that although the straightened sections were directly affected by the cutoffs, the non-straightened sections also developed in a similar pattern, though with different magnitudes of change in the cross-sectional area. The straightened channels, which were mainly pilot channels The rate of change in the thalweg depth was similar within the reaches with a range of less than 5 cm/y in the non-revetmented sections for all periods. In the revetmented sections, although the magnitudes of change were different, the middle and lower reaches had a similar range as the non-revetmented sections (5 cm/y). In the upper reach, however, although the pattern of change was similar, the changes were more than 35 cm/y.

Discussion
The changes in vertical channel morphology of the Lower Tisza River were driven by the sinuosity of the channel and the type and timing of engineering interventions. The repeated surveys of the channel enabled us to do a time-series analyses of the cross sections. It indicates a general incision of the whole channel, which could be related to the increased stream power as a result of the shortened channel length and increased gradient by the artificial cutoffs [74]. The incision due to similar human impacts has been observed in other rivers in the world too [20,50,60,75].
The channel developed generally according to the channel evolution model [76]. The cutoffs that began by the mid-19th century were responsible for most of the changes that occurred in the channel up to 1961, although they were more pronounced directly after the cutoffs, thus between 1891 and 1931. This meant that although the straightened sections were directly affected by the cutoffs, the non-straightened sections also developed in a similar pattern, though with different magnitudes of change in the cross-sectional area. The straightened channels, which were mainly pilot channels created from the cutoffs, experienced incision as well as channel widening as a response to the changes in flow and sediment regime. In response to the increased stream power, the entire channel generally incised. The cutoffs affected 43.5% of the bends, however their mean cross-sectional area increased by 225 m 2 (14.8%) between 1891 and 1961. In the same period cutoffs affected 7.7% of the meanders, but their cross-sectional area increased by 54 m 2 (3.5%). The temporal and spatial changes in the cross sections suggest that after the cutoffs the channel development was towards an equilibrium, as the cross-sectional area of the channel increased in various morphological settings both at bends and meanders, and in straightened and non-straightened sections too.
Between 1931 and 1961, while the straightened sections directly affected by the cutoffs continued to increase their vertical cross-sectional parameters, the directly non-affected meandering sections did not. The entire studied channel however increased at a higher rate than the straightened sections. This indicated that the cutoffs were no longer dominant in controlling the development of the channel, although the channel continued its development towards an equilibrium. This adjustment towards equilibrium was however could not be completed, as the construction of revetments and groynes impeded it, thus the morphological evolution of the river entered another phase of dis-equilibrium. After 1931, when the main period of revetment construction started, the meanders that were affected directly by revetments had started to experience a loss in cross-sectional area, although the channel was still generally increasing.
Between 1931 and 1961, the channel did not give clear morphological response on revetment constructions, probably because the elapsed 30 years was not enough for the river to adjust its morphological evolution. However, after 1961 a clear fluvial response was given, as the cross-sectional channel area decreased drastically, especially of the revetmented sections.
The cross-sectional area of the revetmented sections decreased below pre-1891 values, especially the meandering sections being worst affected. By 1976, the entire studied segment of the Lower Tisza River had started to respond to the construction of the revetments. With most of the revetments already constructed by 1976, the channel began another evolution to reach a new equilibrium. The thalweg depth increased relative to the bankfull width, referring to changing channel shape from wide to a deep and narrower channel. Between 1976 and 1999, the channel seemed to start to re-establish a new equilibrium cycle, as indicated by the increasing cross-sectional areas of non-straightened and non-revetmented segments. However, on some sections (revetmented) the process had not started yet. It is also supported by the hydrological data (discharge and stage relationship, see Figure 3), which refer to declining flood conveyance.
The changes caused by the revetments were further highlighted by the adjustment of the meandering sections. The incision after the revetment constructions tripled within meandering sections with revetments compared to those free of revetments. Within the reaches, there was intensive aggradation and reduction of bankfull width within the upper reach compared to the other reaches. This can mainly be attributed to the groynes, which are located just within the upper reaches and induced aggradation. The cross-sectional area in the upper reach therefore experienced very significant reductions. Within the lower reach, the ponding effect of the Danube on the Tisza [77] coupled with the stabilization of the channel resulted in a relatively stable reach.

Conclusions
The study highlighted the impacts of various human interventions on the morphological evolution of a typical lowland alluvial river. The temporal evolution of the river was mostly controlled by human interventions. Spatially, the various reaches responded to the changes based on the density of human interventions. The upper reach of the Lower Tisza River responded by most changes, while the middle reach had the least response. The lower reach on the other hand, was relatively stable due to its proximity to the confluence of the Tisza River with the larger Danube, which affected channel processes. The various human interventions had profound influences on the evolution of the channel. While the meanders (free of cutoffs) developed with little influence from the cutoffs, the channel's morphological evolution up to 1931 was significantly dominated by the impact of the artificial meander cutoffs. The river had achieved some form of equilibrium by 1961 based on a 100 year-readjustment period for equilibrium [23]. The construction of revetment created another period of readjustment after 1961 with the revetments influencing the morphological development.
Based on the results we suggest reconsidering the actual channel management strategy of the hydrological directorates: (1) the channel should be artificially widened to increase the flood conveyance capacity, (2) the dense vegetation on the banks should be cleared to support the natural channel widening, and (3) the establishment of new revetments should be avoided.
Although the channel is still undergoing changes based on the results, the channel's evolution supports the findings of Surian and Rinaldi [75] about the non-linear pattern of temporal change. The response of the entire studied channel refers to the fact, that adjustments to various human interventions were relatively rapid initially, after which the adjustments slowed down considerably and became asymptotic. The cumulative change based on the effect of the various human interventions is represented by a conceptual model of the evolution of the studied Lower Tisza River ( Figure 20). another period of readjustment after 1961 with the revetments influencing the morphological development.
Based on the results we suggest reconsidering the actual channel management strategy of the hydrological directorates: (1) the channel should be artificially widened to increase the flood conveyance capacity, (2) the dense vegetation on the banks should be cleared to support the natural channel widening, and (3) the establishment of new revetments should be avoided.
Although the channel is still undergoing changes based on the results, the channel's evolution supports the findings of Surian and Rinaldi [75] about the non-linear pattern of temporal change. The response of the entire studied channel refers to the fact, that adjustments to various human interventions were relatively rapid initially, after which the adjustments slowed down considerably and became asymptotic. The cumulative change based on the effect of the various human interventions is represented by a conceptual model of the evolution of the studied Lower Tisza River ( Figure 20).