Introduction

Fluvial landforms and processes have natural functions, being the geomorphologic matrix of riparian ecosystem and generating a high diversity of habitats (Erskine 1990; Leopold et al. 2012; Knighton 2014). Floodplains play an important role in economic development of settlements through accessibility, richness of resources such as water, sands, gravels and biodiversity.

Gravel mining, a widespread anthropogenic activity on almost all rivers of the world, is one of the most aggressive human actions which can cause radical changes in floodplain environment and riverbed (Gaillot and Piégay 1999; Ladson and Judd 2014). Sands and gravels are exploited in-stream (in the river channel) and off-stream (in floodplain gravel pits), by locals and on an industrial scale, in order to obtain raw material for construction and for derived materials (cement, crushed material of different size, concrete) (Collins and Dunne 1989; Kondolf 1994). Wet pit involves the excavation under the water table, or below the thalweg (Kondolf 1994; Padmalal and Maya 2014). In this case, gravel harvesting can interfere with flow velocity, flow direction and fluvial processes and lead to major changes in the flow regime (Leopold et al. 2012; Brestolani et al. 2015), subsequently in riverbed dynamic and floodplain morphology (Rinaldi et al. 2005; Calle et al. 2017). Dry pit involves the removal of soil layer, a superficial excavation of floodplain or riverbank deposits and the removal of the top layer of gravels or/and sands above the water table, which leads to the fragmentation of riparian ecosystems and to the degradation of the meadow landscape. These are true wounds in the meadows and their ecological recovery requiring time and financial effort.

Romanian rivers are also important sources for sands and gravels, and gravel mining is a very common activity in lowlands, the most spread gravel pit being along the interior rivers. There are different types of mining: in-stream by wet pit mining (using dragline or hydraulic excavators, e.g. on the Danube river to improve navigability), bar and banks skimming, mining in active channel and floodplain and low terraces mining by dry pit or wet pit. The ground studies for natural resources conservation and management, which provided the framework of the Romanian National Strategy for Territorial Development (SDTR 2014), indicate that Romania has important resources of sands and gravels, exploited in an organized way with excavation license in 203 perimeters in 32 counties (http://www.mdrap.ro/; http://sdtr.ro/). In Sibiu County, there are two mining areas: one along the Cibin River in the Orlat-Cristian area, which is the subject of this paper, and another along the Olt river in the Bradu area. The National Strategy of Mining Industry (SNIM 2010–2035) indicates that at the beginning of 2011, there were 1250 million tonnes of gravels and sands available. This large amount of such resources assures the long-term production of raw material (20–40 years), knowing that in 2011 a net production of gravels and sands was 31,326 thousand m3 (www.minind.ro/resurse_minerale).

The issue of mining impact on landforms and fluvial processes is complex due to multiple interactions between natural and anthropogenic factors. Riverbeds are dynamic landforms generated by fluvial processes (natural processes), which depend on natural control factors such as geological substrata (Walling and Collins 2005; Rovira and Kondolf 2008), climatic conditions (precipitation amount), riverbed morphology and morphometry (Walling and Collins 2005; Rădoane et al. 2008), flow and water level. Flow variability in time and space influences erosion and accumulation of sediments in riverbeds and floodplain (Hooke 2008; Leopold et al. 2012). If the liquid and solid flow remains unchanged in the long term, the riverbed is gradually adjusted (Norman et al. 1998; Zaharia et al. 2011; Calle et al. 2017) without being noticed by the common observer.

The human actions on the river channel’s and basin upstream gravel pit area change the natural flow regime. These are direct actions on the channel (dam’s construction, channelization, crossings etc.) (Kondolf 1997; Surian and Rinaldi 2003; Rădoane and Rădoane 2005, 2007; Clilverd et al. 2013; Van Looy et al. 2014; Brestolani et al. 2015; Calle et al. 2017) or indirect management actions on river basin surface (urbanization, deforestations, agricultural use) (Kondolf et al. 2002; Rinaldi et al. 2005; Rădoane and Rădoane 2007; Zaharia et al. 2011; Leopold et al. 2012) making it more difficult to assess the gravel harvesting effects. The export of sediment through gravel exploitation leads to major changes in the bedload regime. Particularly, flash floods and even a normal flow can cause rapid changes in the riverbed processes (bank erosion, deepening and channel instability) (Reid et al. 1985; Hooke 2008; Gharbi et al. 2016) or even in lowlands of floodplain when the bank levels are exceeded. In both cases, the flow and sediment load are independent variables which modify the width, depth, transversal shape of riverbed and its pattern. However, the scale of time varies (Norman et al. 1998; Kondolf et al. 2002; Knighton 2014).

The fluvial system needs different time to adapt to the new conditions of gravel harvesting. The impact is obvious in the short term in the case of land use changes, bedload and water table level, while the impact on the leakage and riverbed morphology is felt in the medium and long term (Madej and Ozaki 1996; Kondolf et al. 2002; Michalková et al. 2011). The fluvial processes and the riverbeds response to these actions are also different—changes on the river channel morphometry (depth, width), changes on channel pattern through degradation or aggradation, changes of riverbanks and of floodplain (Wishart et al. 2008; Rădoane et al. 2008; Michalková et al. 2011; Padmalal and Maya 2014).

Worldwide, the issue of gravel pit impact on riverbed dynamics and riverine ecosystems has aroused a particular interest in Australian rivers (Erskine 1990), American rivers (Norman et al. 1998; Kondolf 1994, 1997, 1998; Kondolf et al. 2002), French rivers (Gaillot and Piégay 1999) and Italian rivers (Rinaldi et al. 2005; Surian and Rinaldi 2003). The riverbed response to anthropogenic changes is also the subject of some Romanian applied geomorphology studies, where the dams impact on riverbed and slope dynamic prevail (Rădoane and Rădoane 2005; Rădoane and Rădoane 2007; Rădoane et al. 2008; Zaharia et al. 2011; Hosu and Sabo 2012). The issue on gravel exploitation from riverbeds or floodplains was rarely dealt with, almost neglected, the few Romanian approaches were focused only on sediment dynamics (Rădoane and Rădoane 2005, 2007; Zaharia et al. 2011) and spatial evolution of riverbeds (Rădoane et al. 2008; Hosu and Sabo 2012; Salit and Toroimac 2013). In this context, addressing this topic was a real challenge due to the lack of data at national level.

Moreover, there are no studies that directly address the link between gravel mining, riverbed pattern and erosion or accumulation process. The hypothesis launched in this study is that the gravel exploitation in the floodplain will radically change the geomorphologic landscape and will enhance the accumulation processes in the riverbed and its mobility. The water transfer between the river and the meadow decreases the water transport capacity and increases accumulation. The Cibin River floodplain between Orlat and Cristian settlements (Sibiu Depression) was chosen as study case, and the main pursued objectives were: (1) to highlight the impact of gravel harvesting in floodplain morphology and landscape dynamic, (2) to assess the effect of gravel mining on Cibin riverbed pattern and (3) to identify the trends of fluvial processes on different riverbank sectors. Such studies are almost non-existent thus far in Romania, being also a novelty at regional scale. This study enriches the knowledge on this subject raising the issue of gravel harvesting impact assessment on fluvial processes in the Cibin River basin for the first time and gives insights for a better understanding of net effect triggered by floodplain gravel harvesting.

Materials and methods

The study was conducted based on repeated field observation (2005–2012) on floodplain topography and measurements on riverbed channel sectors and banks to identify the impact of gravel harvesting. The plan development of the riverbed was related to the changes in the liquid and solid flow in different representative years (Walling and Collins 2005; Ta et al. 2013).

The characterization of hydrological regime was carried out on the basis of the flow data obtained from the Sibiu hydrological station (420 m), the representative station for the depressionary area. The flow regime of these years (2005, 2007 and 2009) had a great impact on riverbank morphology and fluvial processes, accelerating the lateral erosion or the accumulation (Hosu and Sabo 2012). Hydrographs of daily average flows at Sibiu station (Fig. 1a–c) were used to highlight a large variability from year to year, but also a large variation in flows during the year, with maximum values reached in different period of the year. Regressions were performed to highlight the relationship between daily average of liquid flow (Q) and bedload (R). The statistical analysis synthesized in Fig. 1 is apparently weak due to the uneven distribution of the average daily flows recorded at the Sibiu station.

Fig. 1
figure 1

Location of interest area in the Cibin River basin and leakage parameters in representative years (Sibiu hydrological station)

Full size image

GIS application was used (Rădoane et al. 2008; Salit and Toroimac 2013; Magliulo et al. 2016; Janes et al. 2017) to analyse the changes of river channel and floodplain morphology. The different time maps were overlapped as different layers to assess the dynamic of landforms and processes: the topographic map (1982, scale 1:25,000), historical maps and also the air photographs from 2005 to 2012 (for which the same scale was also used). The polygons of gravel exploitation areas from two consecutive images were overlapped providing increased gravel pit surfaces and radical changes in land use/land cover (Fig. 1).

To highlight the plan dynamics of the channel, the Cibin riverbed in the gravel pit area was divided into eight linear sectors of 500-m length upstream to downstream between Orlat and Cristian. The choice of sectors length for morphometric analysis can be justified by morphometric indicators and bed configuration, or simply by the accessibility and speed of measurements. Sinuosity was measured according to Janes et al. (2017) at 750-m intervals along the channel centreline; this length of measurements represents the average length of multiband loops across the study catchments. In the morphometric and morphodynamic studies of the riverbeds in Romania, 1-km length sectors were selected for quick and easy measurements (Rădoane et al. 2008; Costea 2010). In this research, the choice of the 500-m length of river sectors was influenced by the length of the riverbed segments with different patterns (Fig. 4).

Channel lines of three consecutive time periods were overlapped in GIS (i.e. 1982–2005–2012), providing different configurations of the river channel with sectors of erosion or accumulation between these periods (Figs. 5 and 6). Channel sinuosity coefficient and braided coefficient were calculated, and measurements have been also made on meander amplitude and length, meander radius, loop amplitude and loop angle. Measurements on maps were completed by direct observations on the field. The lateral instability (Rovira and Kondolf 2008; Michalková et al. 2011) is also revealed along the river by analysing the bank segments length with different fluvial processes. The length of eroded banks and accumulations on each sector were measured on the both banks of the river using different orthophotoplans (2005, 2012). All the measurements were also checked in the field.

Study area

The interest area is located in the central part of Romania, in the south-western extremity of the Sibiu Depression (Fig. 1) along the Cibin River, 12 km west of Sibiu. Cibin River is a right tributary of Olt River in the middle sector. The river is 82 km in length, and its catchment basin extends on 2237 km2. The upper basin occupies the northern flank of the Cindrel Mountains, and the lower basin extends on the Transylvanian Depression, in Sibiu Depression and Hârtibaciului Plateau. It flows about 50 km northward towards Sibiu City and then another 32 km southward to its confluence with the Olt River at Turnu Roşu.

The geological features of the river basin favoured in geological time, the accumulation of gravels and sands in the morphological contact area. The epimetamorphic schists, crystalline limestone and quartzite from the mountainous sector (altered and erodible) provide the river bedload. The lower river course in the Sibiu Depression is developed on Miocene, Pliocene and Quaternary sedimentary rocks (gravels, sands, clay and loam) (Sandu 1998; Costea 2010) which are not resistant to erosion and favour the riverbed mobility.

In these thick deposits, the gravel pit from Orlat was opened and expanded. The resources have been exploited since the 1960–1980. These were used as construction material for the Gura Râului dam, new industrial objectives and neighbourhoods of Sibiu, construction of homes in neighbouring villages, and as raw material in producing asphalt needed in modernization or/and rehabilitation of local and regional communication networks. Nowadays, the location of this gravel pit area in the immediate vicinity of Sibiu—Orăştie motorway construction site, a national interest objective for current national developing strategy (http://www.mdrap.ro/), has been very advantageous for the gravels and sands production and transport through the low costs of extraction and processing and reduction of transportation costs (Kondolf 1994; Norman et al. 1998).

Geomorphological features of interest area

The elevation range of the river basin (1590 m) between the source area (1950 m) and the confluence point (362 m) and steep slope from mountainous area give this river a great energy and erosion power, which is reflected in riverbed morphology, morphometry and mobility. The mountainous sector provided to the lowland area at the end of Pliocene and early Quaternary a large load of sediments and a large piedmont cone was accumulated. The Cibin River terraces and floodplain between Orlat and Cristian (study area) were added to this large accumulation piedmont, during the Quaternary (Sandu 1998), especially on the right side, giving an obvious asymmetric aspect to the valley in this sector. The total width of the floodplain in this sector varies between 1310 m downstream of Orlat and 2500 m upstream of Cristian. The width and asymmetry of the floodplain are the result of relief energy (10–5–2 m), slope decreasing (2°–0°) and migration of the Cibin channel to achieve the equilibrium profile. On the other hand, it is the effect of the pushing exerted by the Cibin and its tributaries alluvial fans, loaded as a response to slope changing.

The Cibin Piedmont, river terraces and floodplain in the morphologic contact area are important sources of gravels and sands. Downstream of Gura Râului, the Cibin River deepens in its own alluvial deposits. The piedmont deposits made by a continuous and thick layer (5–30 m) of well rolled gravels trapped in a sandy mass supply with coarse material the downstream sector of the river in Orlat—Cristian area. Also, horizons of boulders, gravels and coarse sands (2–2.5 m thick) can be identified in the Cibin terraces and floodplain, alternating with horizons of silty sand and recent alluvium. The abundance of this resource was the reason of gravel and sand pit location along the Cibin River (Sandu 1998).

Hydrological features of Cibin River

The Cibin River passes through regions with different pluviometric characteristics. The mountainous climate (with annual amount of precipitations between 700 mm at the mountain bottom and 1200 mm on the summits) is specific to the upper basin and a hilly climate in the lower basin (with annual amount of precipitation between 400 and 650 mm). In these pluvial conditions, the multiannual average flow at the Gura Râului station is 0.821 m3/s, but this increases just upstream of the Orlat settlement to 2.95 m3/s, being about 4.72 m3/s at Sibiu station and 15.5 m3/s at the confluence with the Olt River at Tălmaciu. The maximum value of the annual average flow reached 6.3 m3/s and the minimum 1.42 m3/s. Downstream of Orlat, the multiannual average flow increased because of the flow intake of the Sălişte River (basin surface of 215 km2 and river length of 26 km, morphometric elements with values close to those of the collector river basin—Cibin River upstream Gura Râului). Annual average flows vary from year to year, the module coefficients reaching 2.1 for the wet years (2005–2010) and 0.5 for the drought years (2011–2012). During the year, the highest flow occurs in spring and early summer (between May and July) and the lowest occurs in the cold period (between November and January) (Fig. 1a–c). Snow melt, spring rain and spring and summer rainstorms produce significant flash floods, which have high erosion power and bedload and can remodel the riverbed.

For identical or very close liquid flows, the bedload varies greatly from one moment to another during the year and from 1 year to another. Daily average hydrographs and correlation graphs (Fig. 1) demonstrate not only the variability of bedload at a given liquid flow, but also the solid flow differentiation in the case of flash floods on the two phases (growth and decrease), the maximum bedload being reached before the maximum liquid flow due to the increase in the water level and riverbank erosion.

For the years 2005 and 2007, the occurrence of a larger number of flash floods (6–10) led to the homogenization of the leakage data and to the obtaining of a regression line defined by a polynomial level 2 equation (Fig. 1d, e). Moreover, these two regressions are not very strong (R2 of 0.85, respectively, 0.92), particularly because the Q values are very low up to 10–15 m3/s. However, regressions improve over short periods (R2 of 0.99) in the case of flash floods of these years, the regression lines having a concave linear aspect given by points with the same value in both, the increase and the decrease phase of the flash flood (Fig. 1g).

In 2009, the low number of flash floods (3) and the high amplitude of leakage led to an overall apparently weaker regression equation (Fig. 1f). Even though the highest values of Q (e.g. 35.8 m3/s is the average value produced on 13 July 2009) are real, the equation does not describe almost all the experimental values, which are located in a cluster of values less than 15–20 m3/s. This situation is due to the maximum leakage recorded in the period between June and August and to the longer duration of the flash flood from July. The arrangement of points in a loop shape (Fig. 1f) suggests a representative flash flood with long duration and amplitude (Fig. 1h). The relationship between the two leakage parameters for both the growth and the decrease phase of the flash flood corresponds to a grade 2 polynomial equation with R2 of 0.99. However, these relations indicate the important role of torrential leakage and local riverbed morphology (pattern) and floodplain conditions in generating or storing the solid flow.

The Cibin River does not have a natural flow regime, the river being barred upstream at Gura Râului. After the construction of the dam (1973–1981), flowing regime of the Cibin River changed and the sedimentary cycle of floodplain was modified by the dam that trapped the coarse sediments and halved the suspended load. At the Sibiu station, the multiannual value of suspended load has decreased from 2.15 to 1.08 kg/s, the annual mean values vary between 0.58 and 1.55 kg/s (http://www.rowater.ro/daolt/) and are largely the result of the contribution of Sălişte River (left tributary of the Cibin River). Its basin is developed predominantly on Pannonian sedimentary deposits, and land use is predominantly agricultural (pasture and arable land) which increase the soils and rocks erodibility. In the rainy years (2005, 2007, 2009), the solid flow is higher, increasing the riverbed mobility. As shown in the Fig. 1, in 2009 at Sibiu station, the annual average of solid flow was of 0.995 kg/s, and the maximum value of solid flow was of 66.5 kg/s on 13 July, when a flash flood occurred and the liquid flow reached 43.9 m3/s, nearly 10 times higher than the annual average (http://www.rowater.ro/daolt/).

Results and discussions

Impact of gravel harvesting on floodplain

Changes in floodplain morphology and associated current processes

The comparative analysis of topographic map (1982) and recent air photograph (2005, 2012) indicates that the floodplain morphology has radically changed as the result of gravel harvesting (Fig. 2). Floodplain mining is performed by scraping the superficial layers of sands or gravels, generating dry pits (Fig. 3a, b) and by deep excavations, generating wet pits (Fig. 3c). Both types of exploitation had immediate environmental impact on floodplain geomorphologic function through alluvial deposits striping, decreasing in floodplain elevation, fluctuations in groundwater elevation and sediment load decreasing (Kondolf 1994; Norman et al. 1998; Rinaldi et al. 2005). This function consists in the stochastic role of micro-morphology for the alluvial deposits and in the contribution of fluvial processes in creating and modelling of new fluvial landforms and collateral effects in the riparian spaces.

Fig. 2
figure 2

Dynamic of Orlat gravel pit area (2005–2012)

Full size image
Fig. 3
figure 3

Aspects from Orlat gravel pit area. a Dry pit in exploitation; b abandoned dry pit with chaotic morphology and degradation of riparian habitats by installing a ruderal vegetation. c Active wet pit. Gravel removal operations below water table caused the seeping of groundwater through coarse gravels from the pit bottom. Mining activities caused radical changes in floodplain morphology, land use conversion and landscape changing as well as total degradation of riparian habitats. d The Orlat pond formed in an old gravel pit—a good example of ecological rehabilitation and integration into the landscape as aquatic habitat. Afforestation with woody and shrubs species of lake shorelines and increasing of banks sinuosity could be a better renaturation solution for this gravel pit area

Full size image

The measurements suggest that the aggregate mining activity was intensive by executing large (between 0.5 and 10 ha) and deep depressions (2–10 m) adjacent to the active river channel, most of them below the water level. It was also extensive by expanding the exploited areas (in 2012, the active pits occupied 44.35 ha and the inactive pits 79.34 ha). Wet excavations are deeper (5–10 m), below the water level, leading to the gradual flooding of these depressions through infiltration and the forming of pit ponds (Figs. 2, 3d). The surfaces of pit ponds increased from 15.46 ha in 2005 to 29.94 ha in 2012.

The altitude and floodplain morphology have changed through the occurrences of new temporary deposits of gravels, washed material and depressions having a regular geometry (Figs. 2, 3c). The geometry of pits (steep slope of walls, regular bottom and rectangular shape) favours the gravitational processes. Field observations in the studied area indicate that the heterogeneity of alluvial deposits with weak cohesiveness and a high porosity and permeability may be the main passive factor of collapses and piping. Gravels and boulders are coarse and alternate with coarse and fine sand horizons and sandy-clayey lenses. Due to the intense circulation of groundwater, piping is a process which can lead to major changes in the Cibin riverbed and floodplain morphology through triggering and/or reactivation of gravitational processes. Modelling by collapses and slides of pit walls may be also favoured by the oscillations of water level in lakes, soaking of the sands and clay horizons of alluvial deposits or fine fractions washing (piping), as shown in the previous study of Skempton and Brogan (1994). The pit walls may also collapse because of mechanical shocks from excavations or local traffic on access routes, the pressure causing the gravitational rearranging of materials, favoured by layered and heterogeneous alluvial deposits.

Impact on hydraulic exchanges between floodplain and river

Deep excavations lead to change in hydrological regime of floodplain and imply the hydraulic exchanges between riverbed, underground waters and floodplain (Clilverd et al. 2013). The high permeability of gravels and sands favours the circulation of groundwater into alluvial deposits and the creation of tunnels along the contact between different grain size horizons. In these conditions, piping could increase hydraulic conveyance and lead to hydraulic drawdown and subsequent acceleration of flow towards the pit (Ladson and Judd 2014).

The distance between the ponds and the active channel of the river varies between 15 m in the linear sector and 25–132 m in the meandering sector of the river. The riverbed is deepened, compared to the floodplain surface, by 1–2.5 m between Orlat and Cristian. Under these conditions, the surface hydrological connectivity between the active river channel and floodplain is interrupted, but intense subsurface and underground hydraulic exchanges are established between them. The bidirectional underground water circulation between river and floodplain is slow and constant (Clilverd et al. 2013). Initially, in the wet pits case, because of the lower level of the excavation, compared to the riverbed, the hydraulic transfer was carried from the river to the pits. In this case, the flow in the riverbed decreased until the water level in the lake was stabilized. This process was lengthy and influenced by seasonal variability of rainfall and flow oscillations. In the years with high rainfall (2005, 2007, 2010) and in the warm season of the year, when the average monthly flow rates were double or triple (10–13 m3/s) compared to the multiannual monthly average (4–5 m3/s) due to flash floods (see graphs from Fig. 1), the underground water transfer was carried out from the river to the excavations. This was possible due to river stage position above the floodplain water table (Burt et al. 2002; Clilverd et al. 2013) and led to a faster filling of pits. In the dry years, under base flow conditions, the water from the ponds flew into the river through groundwater circuit. In both cases, the pit ponds from the floodplain acted as stochastic structures, having, in the first situation, a protective role against flooding during high water and flash floods and, in the second situation, as a supply source in ensuring the base flow into the riverbed (Clilverd et al. 2013).

Flooding and avulsion risk

The excavation depressions are isolated from the active river by a strip dike (3–5 m width) with heights of 1.5–2 m from the floodplain surface. This led to the interruption of the floodplain hydrological function as expansion space for floods, and thus increases the flooding risk downstream of the Orlat gravel pit. Under these conditions, the maximum leakage poses the biggest problems. The Cibin active channel functions as an ideal flow channel and increases the velocity of the water. In the case of flash floods, because of the large volume of water, the levels grow and water develops frictional forces against the banks and riverbed leading to bank erosion and deepening of the riverbed (Janes et al. 2017; Calle et al. 2017).

Although woody riparian vegetation acts as a buffer between the two types of aquatic structures (river and pit ponds), there is also the risk of pond captures in case of extraordinary flash floods and flooding. Old abandoned channels located near the lakes may take over in leakage and by erosion and breaking of river banks pit capture may occur in the inflection points of the main river channel (Janes et al. 2017). This phenomenon is known as avulsion and is a very spread phenomenon for the American rivers (Norman et al. 1998; Kondolf 1998; Mossa and Marks 2011), Australian rivers (Erskine 1990; Ladson and Judd 2014) and even for European rivers (Uribelarrea et al. 2003). For the Cibin River, the risk of avulsion is hypothetical because the risk of flooding in this area is very low due to the Gura Râului upstream dam which regulates the flow. In this regard, Kondolf (1994) showed that the risk of avulsion is greatly reduced and the floodplain mining is considered geomorphologically isolated from the active channel when there is complete control through upstream reservoirs.

However, this scenario of avulsion cannot be definitively ruled out, because during previous years, in spite of the planning, the flow rate reached 30–40–50 m3/s to flash flood at the Cristian hydrologic station (e.g. 43.9 m3/s in 13.07.2009; 39.0 m3/s in 23.06.2010; 52.5 m3/s in 10.06.2012; 36.2 m3/s in 5.05.2014), values exceeding 7–12 times the multiannual average of monthly flow (5.56 m3/s for May and 4.67 m3/s for June). These flows are primarily attributable to the Sălişte stream—left tributary upstream gravel pit area and less to the Cibin River whose flow is controlled.

The maximum flow calculated for 1% probability is of 270 m3/s upstream the confluence with Sălişte River and of 350 m3/s downstream. For 10% probability, the maximum flow values decrease by half but remain high (113 m3/s upstream and 158 m3/s downstream Orlat), and the hydrological risk, however, is maintaining in this sector. Lateral erosion and overflow may be possible in these cases, and pit capture could lead to changing of fluvial processes as in the case of in-channel pit (Kondolf 1994, 1997). The regressive erosion increases upstream the pit due to the steeper gradient and riverbed incision occurs. The pit traps the sediments and the liquid flow without bedload (“hungry waters” after Kondolf 1997) acts through linear erosion downstream and leads to deepening of the river channel.

Impact on floodplain land use/land cover

The comparative analysis of historical and actual cartographic documents indicates that the floodplain land use has undergone a conversion in time. At the end of the nineteenth century, the Cibin floodplain landscape was dominated by orchards and crops on the highlands and swamps and wet meadows on the lowlands. Successive maps (1982), aerial photographs (2005–2015) and the field observations indicate that the anthropogenic changes were continuous and accelerated in recent times. In 1982, the gravel pit surface was relatively small (25 ha), the exploitation being carried out on the surface (dry pit). Its surface increased almost three times by 2005 (66.60 ha from which 23.95 ha active pits, 27.19 ha inactive pits and 15.46 ha lakes), demonstrating the acceleration of the exploitation pace and the change of technology (wet pit). In the last decade (2005–2015), mining activity in this area exploded (Fig. 2) due to much greater demand of material to build the Sibiu ring road, the motorway section of Sibiu—Orăştie, and for the rehabilitation and modernization of the streets in Sibiu city neighbourhoods.

In these conditions, gravel harvesting led to the total or partial transformation of floodplain landscape (Fig. 3a–d), and the agricultural use of land being narrowed and replaced with open pits and pit ponds. In time, some of the pit ponds entered the re-naturalizing process by installing hygrophilous vegetation on banks and aquatic vegetation into the water, thus contributing to the development of the aquatic and/or marsh habitats that are beneficial for the emergence and development of wildlife in the Cibin floodplain (Fig. 3d). However, ecological rehabilitation is a long-lasting process because of the morphological characteristics of excavations (steep walls and linear, flat bottom) which hinder the installation of hygrophilic and hydrophilic vegetation or adaptation of aquatic fauna species (Norman et al. 1998; Kondolf 1998). On the other hand, rearranging lake banks require additional costs and excavations are abandoned to the natural processes. The rehabilitation works have not always been done in the case of the dry pit either. The result was a degraded floodplain landscape with a chaotic morphology and vegetal layer dominated by ruderal species that install faster. Dry excavations can still be rehabilitated faster because the operator has the obligation to fill the hole with soil and reintroduce the surface into the ecological circuit by seeding green crops (alfalfa) or herbaceous plants.

The riverbed dynamic and pattern

The evolution and pattern of the riverbed are the result of synergistic conditions like flow and bedload dimension, riverbed slope, bedrock and local conditions (Hooke 2008; Zaharia et al. 2011; Leopold et al. 2012). Both leakage and local conditions of land use in the depression area adjacent to the Cibin River have changed over time, resulting in the modification of the riverbed characteristics. The old maps indicate three patterns of the riverbed: disentangling riverbed near Orlat, meandering riverbed for the second segment and another disentangling sector upstream Cristian. The maps comparison revealed the increase in time of meander amplitude and length. The meandered segment became shorter but wider and the overall sinuosity increased from 1.34 in 1838 to 1.43 in 1907. From the early twentieth century until 1982, the riverbed underwent a migration northward, moving meander loops to downstream and meander captures which, overall, reduced the coefficient of sinuosity to 1.37.

The map from 1982 revealed a sinuous river, with numerous abandoned meanders that are found in floodplain’s landscape through wetland surfaces. The active meanders had eroded banks, deep ruptures of 2–3 m in the concave loops and massive accumulations of sand and gravel into the convex loops. The sudden decrease of slope at the entrance of the river into Sibiu Depression decreases the water velocity and transport capacity of the river. Thus, massive accumulation processes occurred, the disentangling coefficient reaching a value of 2.32 (calculated on topographical map from 1982 for the sector downstream Orlat), enhancing the disentangling of the river course.

Since 1982 until today, the riverbed has maintained the general characteristics but there are some significant changes as a result of combined action of natural and anthropogenic factors.

On the distance of 6.5 km downstream Orlat to Cristian, the Cibin River has a sinuous character revealed by the sinuosity index of 1.171, but on shorter segments the riverbed has different pattern and mobility. The evaluation of aerial photographs (2005, 2012) reveals the alternating, different patterns along the river—two sectors of disentangling (1 and 3 in Fig. 4) and two sinuous sectors (2 and 4 in Fig. 4). The main changes revealed by measurements made on aerial photographs were: meanders radius reduction, changes in the angle between the downstream medium direction and meander trails, increase of meander amplitude, meander loops migrations northward and downstream, decrease of the braided coefficient and increase of sinuosity (Tables 1, 2).

Fig. 4
figure 4

Types of riverbed along the Cibin River in gravel pit area

Full size image
Table 1 Changes in braided sectors of Cibin River in Orlat gravel pit area.
Full size table
Table 2 Recent dynamic and morphometric parameters of meanders in the sinuous sector (2) of Cibin River in Orlat gravel pit area.
Full size table

In the sinuous sectors, the accumulation of coarse material increased in the convex bank where shore and rocky beaches are formed, and the bottom pavement was formed on the riverbed bottom close to the bank. At the reduced flow, the bedload transport is reduced and is given by wrenched materials from the concave banks or from the river bottom and the suspensions transport dominates (Walling and Collins 2005; Ta et al. 2013). These changes in shape were also triggered by an increase in the quantity of suspended alluvium and decrease of gravels and coarse sands which have been exported through excavation. The greatest mobility of the riverbed occurs during flash floods when the coarse material is redistributed in the riverbed and the erosion capacity of the river is high. In this regard, the significant changes of the channel were recently noticed in the sector 1 and 3, from multi-thread channel (2005) to single-thread channel (2012), making the river more sinuous. The multi-thread channel can be explained by the floods from 2005, 2007 and 2009 when the bedload increased and coarse sediments (gravels, boulders) forced the river to disentangling. The relation between Q and R indicates the strong torrential character of the flow in all this years, especially in 2009 (Fig. 1h), which was in generally a low flow year (annual average of 3.31 m3/s). Moreover, in all the analyzed years, the liquid flow growth is accompanied by an increase in the solid discharge, but this is significantly smaller. This is revealed by the low value of regressions coefficient (R2 > 0.76) which indicate a weak correlation between flow parameters. The lower value of R2 for 2009 indicates the importance of local factors in solid discharge control (Zaharia et al. 2011), such as the movement of alluvial deposits, bank collapse in the eroded sectors or drainage of the suspensions after gravel washing. In the last years, the river tends to return to the initial spatial configuration and becomes sinuous and even meandered on some short sectors (see the details of Fig. 4).

The analysis of fluvial processes along the river banks (Fig. 5a, b) indicates for the period 2005–2012 the alternation of eroded with accumulated segments of bank and the increasing riverbank segments length with accumulations totalling 767.85 m (of which 523.73 m on the left bank and 244.12 m to the right bank). The length of eroded sectors decreased on the left bank with 128.85 m and increased with 26.93 m on the right bank. The sequence of fluvial processes related to the riverbank indicates a general trend of accumulation, except for the 7th sector (3000–3500 m) where accumulation is missing (Fig. 5b). There is also a variation of the processes in the same sense (see Fig. 5a)—the increase in length of the eroded river banks and accumulations in the segments 2, 5 and 8 and the decrease in length in the 3rd segment (1000–1500 m). The accumulation and erosion have opposite trends in sectors 1, 4 and 6, but the accumulation is dominant. There are differences between the two banks. Erosion is dominant and increasing on the right bank and the accumulation is dominant and increasing on the left bank (Fig. 6). On both banks of the river, there is an increase in accumulation for the first 1000 m of longitudinal profile and an increase in erosion downstream this sector. This development of processes may be related to the reduction of the amount of coarse alluvium carried by the river and to the increase in the amount of suspensions by the washing of the exploited gravels (a frequent activity in the gravel pit area which increases water turbidity). In this regard, Madej and Ozaki (1996) mention the delayed response of channel (decades) to changes in coarse bedload supply in case of Redwood Creek (California). Also, the issues of channel changes that may follow after catchment modifications were argued by Kondolf et al. (2002) for Drôme River and Pine Creek channel.

Fig. 5
figure 5

Succession of fluvial processes exerted on river banks in the Orlat gravel pit area (2005–2012)

Full size image
Fig. 6
figure 6

Distribution and processes variation on the river banks (length of bank) in the gravel pit area (2005–2012)

Full size image

Conclusions

As a general conclusion, this study reveals that given the local geomorphic conditions and the flow variability the intense harvesting of gravels is the most important control factor that contribute to the changing of floodplain morphology, fluvial processes and riverbed pattern. The results in the study case area from Romania show that the gravel harvesting has a net effect on floodplain geomorphic, hydrological and ecological functions and floodplain landscape through the decreasing in elevation, chaotic morphology with large gravel deposits and rectangular pits flooded by water. Although the connectivity between the river and the meadow is interrupted by a higher buffer strip, there is a risk of avulsion of the ponds in the river meandering sectors. The gravel harvesting has led to the diversification of the gravitational processes in the floodplain: compaction, piping, collapses, landslides caused by the underground water circulation, morphology and morphometry of the excavations. The further monitoring of the gravel exploitation is required especially due to the doubling of pit surface in the last decade and the appearance of local roads with heavy traffic between the pit ponds. The analysis of different time maps demonstrates that the historical riverbed pattern is still maintained, but significant changes specific to each period occur. Discharge variations, hydraulic exchange between river and pit ponds and intensive mining are combined and interdependent and lead, in the last years, to an alternation of riverbed sectors with different patterns. Overall, the riverbed remains predominantly sinuous, even meandered, the meander loops migrating northward and downstream. The storage of water in the lakes led to the decrease of the liquid flow, the reduction of the transport capacity and the dominance of the accumulation process in the riverbed in the pit incidence area. Downstream, the solid flow drops due to gravel exploitation lead to increase of bank erosion and sinuosity and to disentangling decrease.

This study is consistent with the previous international studies and makes a modest contribution to applied geomorphology by presenting a local situation. However, the impact of gravel harvesting on floodplain morphology and riverbed dynamic remains worldwide an open issue and additional studies should be conducted through the long-term monitoring of this kind of areas. The future research should focus on the other gravel pits from Romania to identify specific aspects and to clarify the impact of gravel exploitation on morpho-hydrographical system.

The soil layer removing, fluctuation in groundwater elevation and sediment load decreasing lead to degradation of riparian habitats on short and medium term, but the gravel exploitation could also have a good significance on long term in generating resources and morpho-hydrographical conditions for the wetlands. In this regard, taking measures for the renaturation of the landscape and attenuation of the effects of anthropogenic degradation is needed. A comprehensive approach to all the factors involved in riverbed dynamic and morpho-hydrographical processes induced by gravel exploitation, calls for further interdisciplinary studies and their integration in the sustainable regional development desideratum.